Release 0.6.2

The player was supposed to exit playback if both video and audio failed
to initialize (or if one of the streams was not selected when the other
stream failed). This didn't work; for one this check was missing from
one of the failure paths. And more importantly, both checked the
current_track array incorrectly.

Fix these issues, and move the failure handling code into a common
function.

CC: @mpv-player/stable

Conflicts:
	player/audio.c
	player/video.c

.gitignore

@@ -20,12 +20,12 @@
 /demux/ebml_types.h
 /sub/osd_font.h
 /player/lua/*.inc
-/DOCS/man/*/mpv.1
-/DOCS/man/*/mpv.aux
-/DOCS/man/*/mpv.log
-/DOCS/man/*/mpv.out
-/DOCS/man/*/mpv.pdf
-/DOCS/man/*/mpv.toc
+/DOCS/man/mpv.1
+/DOCS/man/mpv.aux
+/DOCS/man/mpv.log
+/DOCS/man/mpv.out
+/DOCS/man/mpv.pdf
+/DOCS/man/mpv.toc
 
 /waf
 /build

.travis.yml

@@ -1,5 +1,8 @@
 language: c
 
+os:
+  - linux
+  - osx
 env:
   - LIBAV=libav-stable
   - LIBAV=libav-git
@@ -7,24 +10,31 @@ env:
   - LIBAV=ffmpeg-git
 compiler:
   - clang
+
 branches:
   only:
     - master
     - ci
 
+matrix:
+  exclude:
+    - os: osx
+      env: LIBAV=libav-stable
+    - os: osx
+      env: LIBAV=libav-git
+    - os: osx
+      env: LIBAV=ffmpeg-git
+
 before_install: ./travis-deps libass-stable $LIBAV
 script:
   - ./bootstrap.py
   - ./waf configure
   - ./waf build
 
-notifications-policy: &notifications-policy
-  on_success: change
-  on_failure: always
-
 notifications:
   email: false
   irc:
     channels:
       - "irc.freenode.org#mpv-player-dev"
-    <<: *notifications-policy
+    on_success: change
+    on_failure: always

DOCS/client-api-changes.rst

@@ -0,0 +1,49 @@
+Introduction
+============
+
+This file lists all changes that can cause compatibility issues when using
+mpv through the client API (libmpv and ``client.h``). Since the client API
+interfaces to input handling (commands, properties) as well as command line
+options, this list is interesting for other uses of mpv, such as the Lua
+scripting interface, key bindings in ``input.rst``, or plain command line
+usage.
+
+Normally, changes to the C API that are incompatible to previous iterations
+receive a major version bump (i.e. the first version number is increased),
+while C API additions bump the minor version (i.e. the second number is
+increased). Changes to properties/commands/options may also lead to a minor
+version bump, in particular if they are incompatible.
+
+The version number is the same as used for MPV_CLIENT_API_VERSION (see
+``client.h`` how to convert between major/minor version numbers and the flat
+32 bit integer).
+
+Also, read the section ``Compatibility`` in ``client.h``.
+
+API changes
+===========
+
+::
+
+ --- mpv 0.6.0 is released ---
+ 1.5    - change in X11 and "--wid" behavior again. The previous change didn't
+          work as expected, and now the behavior can be explicitly controlled
+          with the "input-x11-keyboard" option. This is only a temporary
+          measure until XEmbed is implemented and confirmed working.
+ 1.4    - subtle change in X11 and "--wid" behavior
+          (this change was added to 0.5.2, and broke some things, see #1090)
+ --- mpv 0.5.0 is released ---
+ 1.3    - add MPV_MAKE_VERSION()
+ 1.2    - remove "stream-time-pos" property (no replacement)
+ 1.1    - remap dvdnav:// to dvd://
+        - add "--cache-file", "--cache-file-size"
+        - add "--colormatrix-primaries" (and property)
+        - add "primaries" sub-field to image format properties
+        - add "playback-time" property
+        - extend the "--start" option; a leading "+", which was previously
+          insignificant is now significant
+        - add "cache-free" and "cache-used" properties
+        - OSX: the "coreaudio" AO spdif code is split into a separate AO
+ --- mpv 0.4.0 is released ---
+ 1.0    - the API is declared stable
+

DOCS/client_api_examples/cocoabasic.m

@@ -0,0 +1,117 @@
+// Plays a video from the command line in a window provided by mpv.
+// You likely want to play the video in your own window instead,
+// but that's not quite ready yet.
+// You may need a basic Info.plist and MainMenu.xib to make this work.
+
+#include "../../libmpv/client.h"
+#include "shared.h"
+
+#import <Cocoa/Cocoa.h>
+
+@interface AppDelegate : NSObject <NSApplicationDelegate>
+{
+    mpv_handle *mpv;
+    dispatch_queue_t queue;
+}
+@end
+
+static void wakeup(void *);
+
+@implementation AppDelegate
+
+- (void) applicationDidFinishLaunching:(NSNotification *)notification {
+    // Read filename
+    NSArray *args = [NSProcessInfo processInfo].arguments;
+    if (args.count < 2) {
+        NSLog(@"Expected filename on command line");
+        exit(1);
+    }
+    NSString *filename = args[1];
+
+    // Deal with MPV in the background.
+    queue = dispatch_queue_create("mpv", DISPATCH_QUEUE_SERIAL);
+    dispatch_async(queue, ^{
+
+        mpv = mpv_create();
+        if (!mpv) {
+            printf("failed creating context\n");
+            exit(1);
+        }
+
+        // Maybe set some options here, like default key bindings.
+        // NOTE: Interaction with the window seems to be broken for now.
+        check_error(mpv_set_option_string(mpv, "input-default-bindings", "yes"));
+
+        check_error(mpv_initialize(mpv));
+
+        // Register to be woken up whenever mpv generates new events.
+        mpv_set_wakeup_callback(mpv, wakeup, (__bridge void *) self);
+
+        // Load the indicated file
+        const char *cmd[] = {"loadfile", filename.UTF8String, NULL};
+        check_error(mpv_command(mpv, cmd));
+    });
+}
+
+- (void) handleEvent:(mpv_event *)event
+{
+    switch (event->event_id) {
+        case MPV_EVENT_SHUTDOWN:
+            // Clean up and shut down.
+            mpv_terminate_destroy(mpv);
+            mpv = NULL;
+            dispatch_async(dispatch_get_main_queue(), ^{
+                [[NSApplication sharedApplication] terminate:nil];
+            });
+            break;
+
+        default:
+            printf("event: %s\n", mpv_event_name(event->event_id));
+    }
+}
+
+- (void) readEvents
+{
+    dispatch_async(queue, ^{
+        while (mpv) {
+            mpv_event *event = mpv_wait_event(mpv, 0);
+            if (event->event_id == MPV_EVENT_NONE)
+                break;
+            [self handleEvent:event];
+        }
+    });
+}
+
+static void wakeup(void *context) {
+    AppDelegate *a = (__bridge AppDelegate *) context;
+    [a readEvents];
+}
+
+// Ostensibly, mpv's window would be hooked up to this.
+- (BOOL) windowShouldClose:(id)sender
+{
+    [self shutdown];
+    return YES;
+}
+
+- (void) shutdown
+{
+    if (mpv) {
+        const char *args[] = {"quit", NULL};
+        mpv_command(mpv, args);
+    }
+}
+@end
+
+
+
+// Delete this if you already have a main.m.
+int main(int argc, const char * argv[]) {
+    @autoreleasepool {
+        NSApplication *app = [NSApplication new];
+        AppDelegate *delegate = [AppDelegate new];
+        app.delegate = delegate;
+        [app run];
+    }
+    return EXIT_SUCCESS;
+}

DOCS/client_api_examples/qtexample.cpp

@@ -0,0 +1,159 @@
+// License: pick one of: public domain, WTFPL, ISC, Ms-PL, AGPLv3
+
+// This example can be built with: qmake && make
+
+#include <sstream>
+
+#include <QFileDialog>
+#include <QStatusBar>
+#include <QMenuBar>
+#include <QMenu>
+#include <QGridLayout>
+#include <QApplication>
+
+#include "qtexample.h"
+
+static void wakeup(void *ctx)
+{
+    // This callback is invoked from any mpv thread (but possibly also
+    // recursively from a thread that is calling the mpv API). Just notify
+    // the Qt GUI thread to wake up (so that it can process events with
+    // mpv_wait_event()), and return as quickly as possible.
+    MainWindow *mainwindow = (MainWindow *)ctx;
+    QCoreApplication::postEvent(mainwindow, new QEvent(QEvent::User));
+}
+
+MainWindow::MainWindow(QWidget *parent) :
+    QMainWindow(parent)
+{
+    QMenu *menu = menuBar()->addMenu(tr("&File"));
+    QAction *on_open = new QAction(tr("&Open"), this);
+    on_open->setShortcuts(QKeySequence::Open);
+    on_open->setStatusTip(tr("Open a file"));
+    connect(on_open, SIGNAL(triggered()), this, SLOT(on_file_open()));
+    menu->addAction(on_open);
+
+    statusBar();
+
+    mpv = mpv_create();
+    if (!mpv)
+        throw "can't create mpv instance";
+
+    // Create a video child window. Force Qt to create a native window, and
+    // pass the window ID to the mpv wid option. This doesn't work on OSX,
+    // because Cocoa doesn't support this form of embedding.
+    mpv_container = new QWidget(this);
+    setCentralWidget(mpv_container);
+    mpv_container->setAttribute(Qt::WA_NativeWindow);
+    // If you have a HWND, use: int64_t wid = (intptr_t)hwnd;
+    int64_t wid = mpv_container->winId();
+    mpv_set_option(mpv, "wid", MPV_FORMAT_INT64, &wid);
+
+    // Enable default bindings, because we're lazy. Normally, a player using
+    // mpv as backend would implement its own key bindings.
+    mpv_set_option_string(mpv, "input-default-bindings", "yes");
+
+    // Enable keyboard input on the X11 window. For the messy details, see
+    // --input-x11-keyboard on the manpage.
+    mpv_set_option_string(mpv, "input-x11-keyboard", "yes");
+
+    // Let us receive property change events with MPV_EVENT_PROPERTY_CHANGE if
+    // this property changes.
+    mpv_observe_property(mpv, 0, "time-pos", MPV_FORMAT_DOUBLE);
+
+    // From this point on, the wakeup function will be called. The callback
+    // can come from any thread, so we use the Qt QEvent mechanism to relay
+    // the wakeup in a thread-safe way.
+    mpv_set_wakeup_callback(mpv, wakeup, this);
+
+    if (mpv_initialize(mpv) < 0)
+        throw "mpv failed to initialize";
+}
+
+void MainWindow::handle_mpv_event(mpv_event *event)
+{
+    switch (event->event_id) {
+    case MPV_EVENT_PROPERTY_CHANGE: {
+        mpv_event_property *prop = (mpv_event_property *)event->data;
+        if (strcmp(prop->name, "time-pos") == 0) {
+            if (prop->format == MPV_FORMAT_DOUBLE) {
+                double time = *(double *)prop->data;
+                std::stringstream ss;
+                ss << "At: " << time;
+                statusBar()->showMessage(QString::fromStdString(ss.str()));
+            } else if (prop->format == MPV_FORMAT_NONE) {
+                // The property is unavailable, which probably means playback
+                // was stopped.
+                statusBar()->showMessage("");
+            }
+        }
+        break;
+    }
+    case MPV_EVENT_VIDEO_RECONFIG: {
+        // Retrieve the new video size.
+        int64_t w, h;
+        if (mpv_get_property(mpv, "dwidth", MPV_FORMAT_INT64, &w) >= 0 &&
+            mpv_get_property(mpv, "dheight", MPV_FORMAT_INT64, &h) >= 0 &&
+            w > 0 && h > 0)
+        {
+            // Note that the MPV_EVENT_VIDEO_RECONFIG event doesn't necessarily
+            // imply a resize, and you should check yourself if the video
+            // dimensions really changed.
+            // mpv itself will scale/letter box the video to the container size
+            // if the video doesn't fit.
+            std::stringstream ss;
+            ss << "Reconfig: " << w << " " << h;
+            statusBar()->showMessage(QString::fromStdString(ss.str()));
+        }
+        break;
+    }
+    case MPV_EVENT_SHUTDOWN: {
+        mpv_terminate_destroy(mpv);
+        mpv = NULL;
+        break;
+    }
+    default: ;
+        // Ignore uninteresting or unknown events.
+    }
+}
+
+bool MainWindow::event(QEvent *event)
+{
+    // QEvent::User is sent by wakeup().
+    if (event->type() == QEvent::User) {
+        // Process all events, until the event queue is empty.
+        while (mpv) {
+            mpv_event *event = mpv_wait_event(mpv, 0);
+            if (event->event_id == MPV_EVENT_NONE)
+                break;
+            handle_mpv_event(event);
+        }
+        return true;
+    }
+    return QMainWindow::event(event);
+}
+
+void MainWindow::on_file_open()
+{
+    QString filename = QFileDialog::getOpenFileName(this, "Open file");
+    if (mpv) {
+        const QByteArray c_filename = filename.toUtf8();
+        const char *args[] = {"loadfile", c_filename.data(), NULL};
+        mpv_command_async(mpv, 0, args);
+    }
+}
+
+MainWindow::~MainWindow()
+{
+    if (mpv)
+        mpv_terminate_destroy(mpv);
+}
+
+int main(int argc, char *argv[])
+{
+    QApplication a(argc, argv);
+    MainWindow w;
+    w.show();
+
+    return a.exec();
+}

DOCS/client_api_examples/qtexample.h

@@ -0,0 +1,30 @@
+#ifndef QTEXAMPLE_H
+#define QTEXAMPLE_H
+
+#include <QMainWindow>
+
+#include <mpv/client.h>
+
+class MainWindow : public QMainWindow
+{
+    Q_OBJECT
+
+public:
+    explicit MainWindow(QWidget *parent = 0);
+    ~MainWindow();
+
+protected:
+    virtual bool event(QEvent *event);
+
+private slots:
+    void on_file_open();
+
+private:
+    QWidget *mpv_container;
+    mpv_handle *mpv;
+
+    void create_player();
+    void handle_mpv_event(mpv_event *event);
+};
+
+#endif // QTEXAMPLE_H

DOCS/client_api_examples/qtexample.pro

@@ -0,0 +1,12 @@
+QT       += core gui
+
+greaterThan(QT_MAJOR_VERSION, 4): QT += widgets
+
+TARGET = qtexample
+TEMPLATE = app
+
+CONFIG += link_pkgconfig
+PKGCONFIG = mpv
+
+SOURCES += qtexample.cpp
+HEADERS += qtexample.h

DOCS/client_api_examples/simple.c

@@ -21,6 +21,7 @@ int main(int argc, char *argv[])
     // Enable default key bindings, so the user can actually interact with
     // the player (and e.g. close the window).
     check_error(mpv_set_option_string(ctx, "input-default-bindings", "yes"));
+    mpv_set_option_string(ctx, "input-x11-keyboard", "yes");
     int val = 1;
     check_error(mpv_set_option(ctx, "osc", MPV_FORMAT_FLAG, &val));
 

DOCS/crosscompile-mingw.md

@@ -26,6 +26,15 @@ created scripts to help ease the process. These are the two recommended methods:
 Note that MinGW environments included in Linux distributions are often broken,
 outdated and useless, and usually don't use MinGW-w64.
 
+Additional dependencies
+-----------------------
+
+You need a pthread wrapper. It must be interoperable with native Windows
+threads. pthreads-win32 or MinGW pthreads might work.
+
+If you want ``--input-file=...`` to work, you need libwaio. It's available
+from: git://midipix.org/waio
+
 Example with MXE
 ----------------
 

DOCS/encoding.rst

@@ -25,7 +25,7 @@ from the input video. Note that not all codecs and not all formats support VFR
 encoding, and some which do have bugs when a target bitrate is specified - use
 -ofps or -oautofps to force CFR encoding in these cases.
 
-Of course, the options can be stored in a profile, like this .mpv/config
+Of course, the options can be stored in a profile, like this .config/mpv/mpv.conf
 section::
 
   [myencprofile]

DOCS/man/af.rst

@@ -72,15 +72,17 @@ Available filters are:
 
     ``tospdif=<yes|no>``
         Output raw AC-3 stream if ``no``, output to S/PDIF for
-        passthrough if ``yes`` (default).
+        pass-through if ``yes`` (default).
 
     ``bitrate=<rate>``
         The bitrate use for the AC-3 stream. Set it to 384 to get 384 kbps.
 
+        The default is 640. Some receivers might not be able to handle this.
+
         Valid values: 32, 40, 48, 56, 64, 80, 96, 112, 128,
         160, 192, 224, 256, 320, 384, 448, 512, 576, 640.
 
-        The special value ``default`` selects a default bitrate based on the
+        The special value ``auto`` selects a default bitrate based on the
         input channel number:
 
         :1ch: 96
@@ -92,7 +94,7 @@ Available filters are:
 
     ``minchn=<n>``
         If the input channel number is less than ``<minchn>``, the filter will
-        detach itself (default: 5).
+        detach itself (default: 3).
 
 ``sweep[=speed]``
     Produces a sine sweep.
@@ -101,7 +103,7 @@ Available filters are:
         Sine function delta, use very low values to hear the sweep.
 
 ``sinesuppress[=freq:decay]``
-    Remove a sine at the specified frequency. Useful to get rid of the 50/60Hz
+    Remove a sine at the specified frequency. Useful to get rid of the 50/60 Hz
     noise on low quality audio equipment. It only works on mono input.
 
     ``<freq>``
@@ -182,7 +184,7 @@ Available filters are:
 
         ``mpv --af=equalizer=11:11:10:5:0:-12:0:5:12:12 media.avi``
             Would amplify the sound in the upper and lower frequency region
-            while canceling it almost completely around 1kHz.
+            while canceling it almost completely around 1 kHz.
 
 ``channels=nch[:routes]``
     Can be used for adding, removing, routing and copying audio channels. If
@@ -263,9 +265,8 @@ Available filters are:
     Filter for internal use only. Converts between 24-bit and 32-bit sample
     formats.
 
-``convertsignendian``
-    Filter for internal use only. Converts between signed/unsigned formats
-    and formats with different endian.
+``convertsign``
+    Filter for internal use only. Converts between signed/unsigned formats.
 
 ``volume[=<volumedb>[:...]]``
     Implements software volume control. Use this filter with caution since it
@@ -277,8 +278,8 @@ Available filters are:
     once for every audio stream.
 
     ``<volumedb>``
-        Sets the desired gain in dB for all channels in the stream from -200dB
-        to +60dB, where -200dB mutes the sound completely and +60dB equals a
+        Sets the desired gain in dB for all channels in the stream from -200 dB
+        to +60 dB, where -200 dB mutes the sound completely and +60 dB equals a
         gain of 1000 (default: 0).
     ``replaygain-track``
         Adjust volume gain according to the track-gain replaygain value stored
@@ -311,7 +312,7 @@ Available filters are:
     .. admonition:: Example
 
         ``mpv --af=volume=10.1 media.avi``
-            Would amplify the sound by 10.1dB and hard-clip if the sound level
+            Would amplify the sound by 10.1 dB and hard-clip if the sound level
             is too high.
 
 ``pan=n:[<matrix>]``
@@ -368,8 +369,8 @@ Available filters are:
         to the subwoofer.
 
     ``<fc>``
-        cutoff frequency in Hz for the low-pass filter (20Hz to 300Hz)
-        (default: 60Hz) For the best result try setting the cutoff frequency
+        cutoff frequency in Hz for the low-pass filter (20 Hz to 300 Hz)
+        (default: 60 Hz) For the best result try setting the cutoff frequency
         as low as possible. This will improve the stereo or surround sound
         experience.
     ``<ch>``
@@ -381,7 +382,7 @@ Available filters are:
     .. admonition:: Example
 
         ``mpv --af=sub=100:4 --audio-channels=5 media.avi``
-            Would add a subwoofer channel with a cutoff frequency of 100Hz to
+            Would add a subwoofer channel with a cutoff frequency of 100 Hz to
             output channel 4.
 
 ``center``
@@ -409,7 +410,7 @@ Available filters are:
     .. admonition:: Example
 
         ``mpv --af=surround=15 --audio-channels=4 media.avi``
-            Would add surround sound decoding with 15ms delay for the sound to
+            Would add surround sound decoding with 15 ms delay for the sound to
             the rear speakers.
 
 ``delay[=[ch1,ch2,...]]``
@@ -437,8 +438,8 @@ Available filters are:
     .. admonition:: Example
 
         ``mpv --af=delay=[10.5,10.5,0,0,7,0] media.avi``
-            Would delay front left and right by 10.5ms, the two rear channels
-            and the subwoofer by 0ms and the center channel by 7ms.
+            Would delay front left and right by 10.5 ms, the two rear channels
+            and the subwoofer by 0 ms and the center channel by 7 ms.
 
 ``export=mmapped_file:nsamples]``
     Exports the incoming signal to other processes using memory mapping
@@ -588,7 +589,7 @@ Available filters are:
             tempo.
 
         ``mpv --af=scaletempo=stride=30:overlap=.50:search=10 media.ogg``
-            Would tweak the quality and performace parameters.
+            Would tweak the quality and performance parameters.
 
         ``mpv --af=format=float,scaletempo media.ogg``
             Would make scaletempo use float code. Maybe faster on some
@@ -600,7 +601,7 @@ Available filters are:
             1.2x.
 
 ``lavfi=graph``
-    Filter audio using ffmpeg's libavfilter.
+    Filter audio using FFmpeg's libavfilter.
 
     ``<graph>``
         Libavfilter graph. See ``lavfi`` video filter for details - the graph

DOCS/man/ao.rst

@@ -23,7 +23,7 @@ normal driver parameters.
     where PulseAudio is used. On Windows, ``--ao=wasapi`` is preferred,
     though it might cause trouble sometimes, in which case ``--ao=dsound``
     should be used. On BSD systems, ``--ao=oss`` or `--ao=sndio`` may work
-    (the latter being experimental). On OSX systems, use ``--ao=coreaudio``.
+    (the latter being experimental). On OS X systems, use ``--ao=coreaudio``.
 
 .. admonition:: Examples
 
@@ -53,6 +53,10 @@ Available audio output drivers are:
         Set the index of the mixer channel (default: 0). Consider the output of
         "``amixer scontrols``", then the index is the number that follows the
         name of the element.
+    ``non-interleaved``
+        Allow output of non-interleaved formats (if the audio decoder uses
+        this format). Currently disabled by default, because some popular
+        ALSA plugins are utterly broken with non-interleaved formats.
 
     .. note::
 
@@ -110,7 +114,19 @@ Available audio output drivers are:
         with fixed setups.
 
 ``coreaudio`` (Mac OS X only)
-    Native Mac OS X audio output driver
+    Native Mac OS X audio output driver using the AudioUnits and the CoreAudio
+    sound server.
+
+    ``device_id=<id>``
+        ID of output device to use (0 = default device)
+    ``help``
+        List all available output devices with their IDs.
+
+``coreaudio_device`` (Mac OS X only)
+    Native Mac OS X audio output driver using direct device access and
+    exclusive mode (bypasses the sound server).
+
+    Supports only compressed formats (AC3 and DTS).
 
     ``device_id=<id>``
         ID of output device to use (0 = default device)
@@ -137,6 +153,12 @@ Available audio output drivers are:
         value makes the audio stream react faster, e.g. to playback speed
         changes. Default: 250.
 
+    ``latency-hacks=<yes|no>``
+        Enable hacks to workaround PulseAudio timing bugs (default: yes). If
+        enabled, mpv will do elaborate latency calculations on its own. If
+        disabled, it will use PulseAudio automatically updated timing
+        information. Disabling this might help with e.g. networked audio.
+
 ``portaudio``
     PortAudio audio output driver. This works on all platforms, and has
     extensive MS Windows support.
@@ -159,6 +181,9 @@ Available audio output drivers are:
         Sets the device number to use. Playing a file with ``-v`` will show a
         list of available devices.
 
+    ``buffersize=<ms>``
+        DirectSound buffer size in milliseconds (default: 200).
+
 ``sdl``
     SDL 1.2+ audio output driver. Should work on any platform supported by SDL
     1.2, but may require the ``SDL_AUDIODRIVER`` environment variable to be set

DOCS/man/changes.rst

@@ -19,14 +19,14 @@ General Changes for MPlayer to mpv
   Linux 2.4 (including VIDIX)
 * Large internal cleanups
 * Removal of support for dead platforms
-* New buildsystem
-* No embedded copy of ffmpeg and other libraries
+* New build system
+* No embedded copy of FFmpeg and other libraries
 * Better pause handling (do not unpause on a command)
 * Better MKV support (such as ordered chapters)
 * vo_vdpau improvements
 * Precise seeking support
-* Native OpenGL backend for OSX
-* General OSX improvements
+* Native OpenGL backend for OS X
+* General OS X improvements
 * Improvements in audio/video sync handling
 * Cleaned up terminal output
 * Gapless audio support (``--gapless-audio``)
@@ -35,14 +35,14 @@ General Changes for MPlayer to mpv
 * OSS4 volume control
 * More correct color reproduction (color matrix generation)
 * Use libass for subtitle rendering by default (better quality)
-* Generally preferring ffmpeg/libav over internal demuxers and decoders
+* Generally preferring FFmpeg/Libav over internal demuxers and decoders
 * Improvements when playing multiple files (``--fixed-vo``)
 * Screenshot improvements (instant screenshots without 1-frame delay, allow
   taking screenshots even with hardware decoding)
 * Improved support for PulseAudio
-* Generally improved MS Windows support (dealing with unicode filenames,
+* Generally improved MS Windows support (dealing with Unicode file names,
   improved ``--vo=direct3d``, improved window handling)
-* Better OSD rendering (using libass). This has full unicode support, and
+* Better OSD rendering (using libass). This has full Unicode support, and
   languages like Arabic should be better supported.
 * Cleaned up terminal output (nicer status line, less useless noise)
 * Support for playing URLs of popular streaming sites directly
@@ -50,19 +50,19 @@ General Changes for MPlayer to mpv
 * Improved OpenGL output (``--vo=opengl-hq``)
 * Make ``--softvol`` default (**mpv** is not a mixer control panel)
 * Improved support for .cue files
-* Screenshot improvements (can save screenshots as JPG or PNG, configurable
-  filenames, support for taking screenshots with or without subtitles - the
+* Screenshot improvements (can save screenshots as JPEG or PNG, configurable
+  file names, support for taking screenshots with or without subtitles - the
   ``screenshot`` video filter is not needed anymore, and should not be put
   into the mpv config file)
 * Removal of teletext support
-* Removal of most builtin demuxers, using libavformat instead
-* Removal of builtin network support, using libavformat instead (also,
+* Removal of most built-in demuxers, using libavformat instead
+* Removal of built-in network support, using libavformat instead (also,
   support https via libavformat)
 * Replace image VOs (``--vo=jpeg`` etc.) with ``--vo=image``
 * Do not lose settings when playing a new file in the same player instance
 * New location for config files, new name for the binary.
 * Slave mode compatibility broken (see below)
-* Encoding functionality (replacement for mencoder, see ``DOCS/encoding.rst``)
+* Encoding functionality (replacement for MEncoder, see ``DOCS/encoding.rst``)
 * Remove ``--vo=gif89a``, ``--vo=md5sum``, ``--vo=yuv4mpeg``, as encoding can
   handle these use cases. For yuv4mpeg, for example, use:
   ``mpv input.mkv -o output.y4m --no-audio --oautofps --oneverdrop``.
@@ -71,10 +71,10 @@ General Changes for MPlayer to mpv
 * Wayland support
 * Support for precise scrolling which scales the parameter of commands. If the
   input doesn't support precise scrolling the scale factor stays 1.
-* OSX: Cocoa event loop is independent from mplayer's event loop, so user
+* OS X: Cocoa event loop is independent from MPlayer's event loop, so user
   actions like accessing menus and live resizing do not block the playback.
-* OSX: Apple Remote support.
-* OSX: Media Keys support.
+* OS X: Apple Remote support.
+* OS X: Media Keys support.
 * Windows: Added WASAPI audio output.
 * New OSD bar with chapter marks and not positioned in the middle of the video
   (though this can be customized with the ``--osd-bar-align-y`` option).
@@ -99,7 +99,7 @@ General Changes for MPlayer to mpv
 * Allow changing/adjusting video filters at runtime. (This is also used to make
   the ``D`` key insert vf_yadif if deinterlacing is not supported otherwise.)
 * Native VAAPI support
-* OSX: VDA support using libavcodec hwaccel API insted of FFmpeg's decoder. Up
+* OS X: VDA support using libavcodec hwaccel API instead of FFmpeg's decoder. Up
   to 2-2.5x reduction in CPU usage.
 * Make hardware decoding in general work with the ``opengl`` video output.
 * Lua scripting (see `LUA SCRIPTING`_)
@@ -308,28 +308,40 @@ input.conf and Slave Commands
 Slave mode
 ~~~~~~~~~~
 
-* Slave mode is broken. This mode is entirely insane in the ``old`` versions of
-  MPlayer. A proper slave mode application needed tons of code and hacks to get
+* Slave mode was removed. A proper slave mode application needed tons of code
+  and hacks to get
   it right. The main problem is that slave mode is a bad and incomplete
   interface, and to get around that, applications parsed output messages
   intended for users. It is hard to know which messages exactly are parsed by
   slave mode applications. This makes it virtually impossible to improve
   terminal output intended for users without possibly breaking something.
 
-  This is absolutely insane, and **mpv** will not try to keep slave mode
-  compatible. If you are a developer of a slave mode application, contact us,
-  and a new and better protocol can be developed.
+  This is absolutely insane, and since initial improvements to **mpv** quickly
+  made slave mode incompatible to most applications, it was removed as useless
+  cruft. The client API (see below) is provided instead.
 
-  ``--identify`` was already removed (``TOOLS/mpv_identify.sh`` is provided
-  instead), and ``--slave-broken`` might be removed in the future.
+  ``--identify`` was replaced by the ``TOOLS/mpv_identify.sh`` wrapper script.
+
+* For some time (until including release 0.4.x), mpv supported a
+  ``--slave-broken`` option. The following options are equivalent:
+
+  ::
+
+        --input-file=/dev/stdin --input-terminal=no
+
+
+  Assuming the system supports ``/dev/stdin``.
+
+  (The option was readded in 0.5.1 and sets exactly these options.)
+
+* A JSON RPC protocol giving access to the client API is planned, but nothing
+  has emerged yet.
 
 * **mpv** also provides a client API, which can be used to embed the player
   by loading it as shared library. (See ``libmpv/client.h`` in the sources.)
   It might also be possible to implement a custom slave mode-like protocol
   using Lua scripting.
 
-* A slave protocol is planned, but nothing has emerged yet.
-
 Policy for Removed Features
 ---------------------------
 

DOCS/man/encode.rst

@@ -38,7 +38,7 @@ You can encode files from one format/codec to another using this facility.
 
 ``--oautofps``
     Sets the output format time base to the guessed frame rate of the input
-    video (simulates mencoder behavior, useful for AVI; may cause frame drops).
+    video (simulates MEncoder behavior, useful for AVI; may cause frame drops).
     Note that not all codecs and not all formats support VFR encoding, and some
     which do have bugs when a target bitrate is specified - use ``--ofps`` or
     ``--oautofps`` to force CFR encoding in these cases.
@@ -74,7 +74,7 @@ You can encode files from one format/codec to another using this facility.
     .. admonition:: Example
 
         "``--oac=libmp3lame --oacopts=b=128000``"
-            selects 128kbps MP3 encoding.
+            selects 128 kbps MP3 encoding.
 
     Options are managed in lists. There are a few commands to manage the
     options list.

DOCS/man/input.rst

@@ -4,12 +4,19 @@ INPUT.CONF
 The input.conf file consists of a list of key bindings, for example::
 
     s screenshot      # take a screenshot with the s key
+    LEFT seek 15      # map the left-arrow key to seeking forward by 15 seconds
 
 Each line maps a key to an input command. Keys are specified with their literal
 value (upper case if combined with ``Shift``), or a name for special keys. For
 example, ``a`` maps to the ``a`` key without shift, and ``A`` maps to ``a``
 with shift.
 
+The file is located in the mpv configuration directory (normally at
+``~/.config/mpv/input.conf`` depending on platform). The default bindings are
+defined here::
+
+    https://github.com/mpv-player/mpv/blob/master/etc/input.conf
+
 A list of special keys can be obtained with
 
     ``mpv --input-keylist``
@@ -101,14 +108,14 @@ List of Input Commands
     ``revert_seek`` command itself.
 
 ``frame_step``
-    Play one frame, then pause.
+    Play one frame, then pause. Does nothing with audio-only playback.
 
 ``frame_back_step``
     Go back by one frame, then pause. Note that this can be very slow (it tries
     to be precise, not fast), and sometimes fails to behave as expected. How
     well this works depends on whether precise seeking works correctly (e.g.
     see the ``--hr-seek-demuxer-offset`` option). Video filters or other video
-    postprocessing that modifies timing of frames (e.g. deinterlacing) should
+    post-processing that modifies timing of frames (e.g. deinterlacing) should
     usually work, but might make backstepping silently behave incorrectly in
     corner cases. Using ``--hr-seek-framedrop=no`` should help, although it
     might make precise seeking slower.
@@ -153,7 +160,9 @@ List of Input Commands
         Take a single screenshot.
     <each-frame>
         Take a screenshot each frame. Issue this command again to stop taking
-        screenshots.
+        screenshots. Note that you should disable frame-dropping when using
+        this mode - or you might receive duplicate images in cases when a
+        frame was dropped.
 
 ``screenshot_to_file "<filename>" [subtitles|video|window]``
     Take a screenshot and save it to a given file. The format of the file will
@@ -193,6 +202,8 @@ List of Input Commands
         Stop playback of the current file, and play the new file immediately.
     <append>
         Append the file to the playlist.
+    <append-play>
+        Append the file, and if nothing is currently playing, start playback.
 
     The third argument is a list of options and values which should be set
     while the file is playing. It is of the form ``opt1=value1,opt2=value2,..``.
@@ -250,7 +261,7 @@ List of Input Commands
     exactly as in the ``quit`` command.
 
 ``sub_add "<file>"``
-    Load the given subtitle file. It is not selected as current subtitle after
+    Load the given subtitle file. It is selected as current subtitle after
     loading.
 
 ``sub_remove [<id>]``
@@ -258,7 +269,7 @@ List of Input Commands
     the current track. (Works on external subtitle files only.)
 
 ``sub_reload [<id>]``
-    Reload the given subtitle tracks. If the ``id`` argument is missing, remove
+    Reload the given subtitle tracks. If the ``id`` argument is missing, reload
     the current track. (Works on external subtitle files only.)
 
     This works by unloading and re-adding the subtitle track.
@@ -317,6 +328,12 @@ List of Input Commands
     Write the resume config file that the ``quit_watch_later`` command writes,
     but continue playback normally.
 
+``stop``
+    Stop playback and clear playlist. With default settings, this is
+    essentially like ``quit``. Useful for the client API: playback can be
+    stopped without terminating the player.
+
+
 Input Commands that are Possibly Subject to Change
 --------------------------------------------------
 
@@ -493,8 +510,8 @@ Input Commands that are Possibly Subject to Change
     be used in input.conf to reassign such bindings.)
 
 
-Undocumented commands: ``tv_last_channel`` (TV/DVB only), ``stop`` (questionable
-use), ``get_property`` (?), ``vo_cmdline`` (experimental).
+Undocumented commands: ``tv_last_channel`` (TV/DVB only), ``get_property`` (?),
+``vo_cmdline`` (experimental).
 
 Input Command Prefixes
 ----------------------
@@ -564,7 +581,7 @@ Property list
     See ``--osd-level``.
 
 ``osd-scale`` (RW)
-    OSD font size multiplicator, see ``--osd-scale``.
+    OSD font size multiplier, see ``--osd-scale``.
 
 ``loop`` (RW)
     See ``--loop``.
@@ -586,6 +603,18 @@ Property list
     ``${stream-end}``. For ordered chapters and such, the
     size of the currently played segment is returned.)
 
+``estimated-frame-count``
+    Total number of frames in current file.
+
+    .. note:: This is only an estimate. (It's computed from two unreliable
+              quantities: fps and stream length.)
+
+``estimated-frame-number``
+    Number of current frame in current stream.
+
+    .. note:: This is only an estimate. (It's computed from two unreliable
+              quantities: fps and possibly rounded timestamps.)
+
 ``path``
     Full path of the currently played file.
 
@@ -613,11 +642,6 @@ Property list
 ``stream-end``
     Raw end position in bytes in source stream.
 
-``stream-time-pos`` (RW)
-    Time position in source stream. This only works for DVD and Bluray. This
-    is probably never different from ``time-pos``, because ``time-pos`` is
-    forced to this value anyway.
-
 ``length``
     Length of the current file in seconds. If the length is unknown, the
     property is unavailable. Note that the file duration is not always exactly
@@ -632,8 +656,12 @@ Property list
     disabled.
 
 ``drop-frame-count``
-    Frames dropped because they arrived to late. Unavailable if video
-    is disabled
+    Frames dropped because they arrived too late. Doesn't necessarily indicate
+    actual frame-drops, just the number of times the decoder was asked to drop.
+    Unavailable if video is disabled
+
+``vo-drop-frame-count``
+    Frames dropped by VO (when using ``--framedrop=vo``).
 
 ``percent-pos`` (RW)
     Position in current file (0-100). The advantage over using this instead of
@@ -641,9 +669,6 @@ Property list
     estimating the playback position from the byte position, if the file
     duration is not known.
 
-``ratio-pos`` (RW)
-    Position in current file (0.0-1.0). higher precision that ``percent-pos``.
-
 ``time-pos`` (RW)
     Position in current file in seconds.
 
@@ -658,6 +683,9 @@ Property list
 ``playtime-remaining``
     ``time-remaining`` scaled by the the current ``speed``.
 
+``playback-time``
+    Return the playback time, which is the time difference between start PTS and current PTS.
+
 ``chapter`` (RW)
     Current chapter number. The number of the first chapter is 0.
 
@@ -759,7 +787,7 @@ Property list
     ::
 
         MPV_FORMAT_NODE_MAP
-            (key and string value for each metdata entry)
+            (key and string value for each metadata entry)
 
 ``chapter-metadata``
     Metadata of current chapter. Works similar to ``metadata`` property. It
@@ -791,7 +819,7 @@ Property list
     pauses itself due to low network cache.
 
 ``cache``
-    Network cache fill state (0-100).
+    Network cache fill state (0-100.0).
 
 ``cache-size`` (RW)
     Total network cache size in KB. This is similar to ``--cache``. This allows
@@ -803,7 +831,26 @@ Property list
     this easier, the cache resizing code will allocate the new cache while the
     old cache is still allocated.
 
-    Don't use this when playing DVD or Bluray.
+    Don't use this when playing DVD or Blu-ray.
+
+``cache-free`` (R)
+    Total free cache size in KB.
+
+``cache-used`` (R)
+    Total used cache size in KB.
+
+``cache-idle`` (R)
+    Returns ``yes`` if the cache is idle, which means the cache is filled as
+    much as possible, and is currently not reading more data.
+
+``demuxer-cache-duration``
+    Approximate duration of video buffered in the demuxer, in seconds. The
+    guess is very unreliable, and often the property will not be available
+    at all, even if data is buffered.
+
+``demuxer-cache-idle``
+    Returns ``yes`` if the demuxer is idle, which means the demuxer cache is
+    filled to the requested amount, and is currently not reading more data.
 
 ``paused-for-cache``
     Returns ``yes`` when playback is paused because of waiting for the cache.
@@ -815,6 +862,12 @@ Property list
     or enter idle mode), and in these cases the ``eof-reached`` property will
     logically be cleared immediately after it's set.
 
+``seeking``
+    Returns ``yes`` if the player is currently seeking, or otherwise trying
+    to restart playback. (It's possible that it returns ``yes`` while a file
+    is loaded, or when switching ordered chapter segments. This is because
+    the same underlying code is used for seeking and resyncing.)
+
 ``pts-association-mode`` (RW)
     See ``--pts-association-mode``.
 
@@ -978,19 +1031,24 @@ Property list
                 "par"               MPV_FORMAT_DOUBLE
                 "colormatrix"       MPV_FORMAT_STRING
                 "colorlevels"       MPV_FORMAT_STRING
+                "primaries"         MPV_FORMAT_STRING
                 "chroma-location"   MPV_FORMAT_STRING
                 "rotate"            MPV_FORMAT_INT64
 
 ``dwidth``, ``dheight``
     Video display size. This is the video size after filters and aspect scaling
     have been applied. The actual video window size can still be different
-    from this.
+    from this, e.g. if the user resized the video window manually.
+
+    These have the same values as ``video-out-params/dw`` and
+    ``video-out-params/dh``.
 
 ``video-out-params``
     Same as ``video-params``, but after video filters have been applied. If
     there are no video filters in use, this will contain the same values as
     ``video-params``. Note that this is still not necessarily what the video
-    window uses, since all real VOs do their own scaling.
+    window uses, since the user can change the window size, and all real VOs
+    do their own scaling independently from the filter chain.
 
     Has the same sub-properties as ``video-params``.
 
@@ -1002,7 +1060,7 @@ Property list
     Estimated/measured FPS of the video filter chain output. (If no filters
     are used, this corresponds to decoder output.) This uses the average of
     the 10 past frame durations to calculate the FPS. It will be inaccurate
-    if framedropping is involved (such as when framedrop is explicitly
+    if frame-dropping is involved (such as when framedrop is explicitly
     enabled, or after precise seeking). Files with imprecise timestamps (such
     as Matroska) might lead to unstable results.
 
@@ -1010,7 +1068,7 @@ Property list
     Window size multiplier. Setting this will resize the video window to the
     values contained in ``dwidth`` and ``dheight`` multiplied with the value
     set with this property. Setting ``1`` will resize to original video size
-    (or to be exactly, the size the video filters output). ``2`` will set the
+    (or to be exact, the size the video filters output). ``2`` will set the
     double size, ``0.5`` halves the size.
 
 ``video-aspect`` (RW)
@@ -1067,7 +1125,7 @@ Property list
     See ``--sub-forced-only``.
 
 ``sub-scale`` (RW)
-    Subtitle font size multiplicator.
+    Subtitle font size multiplier.
 
 ``ass-use-margins`` (RW)
     See ``--ass-use-margins``.
@@ -1079,8 +1137,8 @@ Property list
     See ``--ass-style-override``.
 
 ``stream-capture`` (RW)
-    A filename, see ``--capture``. Setting this will start capture using the
-    given filename. Setting it to an empty string will stop it.
+    A filename, see ``--stream-capture``. Setting this will start capture using
+    the given filename. Setting it to an empty string will stop it.
 
 ``tv-brightness``, ``tv-contrast``, ``tv-saturation``, ``tv-hue`` (RW)
     TV stuff.
@@ -1211,11 +1269,6 @@ Property list
                 "title" MPV_FORMAT_STRING
                 "time"  MPV_FORMAT_DOUBLE
 
-``quvi-format`` (RW)
-    See ``--quvi-format``. Cycling this property (``cycle``) will attempt to
-    cycle through known format, although currently this feature doesn't work
-    well at all.
-
 ``af`` (RW)
     See ``--af`` and the ``af`` command.
 
@@ -1241,14 +1294,36 @@ Property list
 ``seekable``
     Return whether it's generally possible to seek in the current file.
 
+``osd-sym-cc``
+    Inserts the current OSD symbol as opaque OSD control code (cc). This makes
+    sense only with the ``show_text`` command or options which set OSD messages.
+    The control code is implementation specific and is useless for anything else.
+
+``osd-ass-cc``
+    ``${osd-ass-cc/0}`` disables escaping ASS sequences of text in OSD,
+    ``${osd-ass-cc/1}`` enables it again. By default, ASS sequences are
+    escaped to avoid accidental formatting, and this property can disable
+    this behavior. Note that the properties return an opaque OSD control
+    code, which only makes sense for the ``show_text`` command or options
+    which set OSD messages.
+
+    .. admonition:: Example
+
+        --osd-status-msg='This is ${osd-ass-cc/0}{\\b1}bold text'
+
+    Any ASS override tags as understood by libass can be used.
+
+    Note that you need to escape the ``\`` character, because the string is
+    processed for C escape sequences before passing it to the OSD code.
+
 ``options/<name>`` (RW)
     Read-only access to value of option ``--<name>``. Most options can be
     changed at runtime by writing to this property. Note that many options
     require reloading the file for changes to take effect. If there is an
     equivalent property, prefer setting the property instead.
 
-``write_watch_later_config``
-    Saves current playback position.
+``property-list``
+    Return the list of top-level properties.
 
 Property Expansion
 ------------------

DOCS/man/lua.rst

@@ -1,10 +1,16 @@
 LUA SCRIPTING
 =============
 
-mpv can load Lua scripts. Scripts in ``~/.mpv/lua/`` will be loaded on program
-start, or if passed to ``--lua``. mpv provides the builtin module ``mp``, which
-provides functions to send commands to the mpv core and to retrieve information
-about playback state, user settings, file information, and so on.
+mpv can load Lua scripts. Scripts passed to the ``--lua`` option, or found in
+the ``lua`` subdirectory of the mpv configuration directory (usually
+``~/.config/mpv/lua/``) will be loaded on program start. mpv also appends the
+``lua`` subdirectory to the end of Lua's path so you can import scripts from
+there too. Since it's added to the end, don't name scripts you want to import
+the same as Lua libraries because they will be overshadowed by them.
+
+mpv provides the built-in module ``mp``, which provides functions to send
+commands to the mpv core and to retrieve information about playback state, user
+settings, file information, and so on.
 
 These scripts can be used to control mpv in a similar way to slave mode.
 Technically, the Lua code uses the client API internally.
@@ -27,12 +33,13 @@ allow users implement features which are not going to be added to the mpv core.
 Mode of operation
 -----------------
 
-Your script will be loaded by the player at program start from ``~/.mpv/lua/``,
-or ``--lua``, or in some cases, internally (like ``--osc``). Each script runs
-in its own thread. Your script is first run "as is", and once that is done,
-the event loop is entered. This event loop will dispatch events received by mpv
-and call your own event handlers which you have registered with
-``mp.register_event``, or timers added with ``mp.add_timeout`` or similar.
+Your script will be loaded by the player at program start from the ``lua``
+configuration subdirectory, from a path specified with the ``--lua`` option, or
+in some cases, internally (like ``--osc``). Each script runs in its own
+thread. Your script is first run "as is", and once that is done, the event loop
+is entered. This event loop will dispatch events received by mpv and call your
+own event handlers which you have registered with ``mp.register_event``, or
+timers added with ``mp.add_timeout`` or similar.
 
 When the player quits, all scripts will be asked to terminate. This happens via
 a ``shutdown`` event, which by default will make the event loop return. If your
@@ -47,6 +54,15 @@ The event loop will wait for events and dispatch events registered with
 ``mp.register_event``. It will also handle timers added with ``mp.add_timeout``
 and similar (by waiting with a timeout).
 
+Since mpv 0.6.0, the player will wait until the script is fully loaded before
+continuing normal operation. The player considers a script as fully loaded as
+soon as it starts waiting for mpv events (or it exits). In practice this means
+the player will more or less hang until the script returns from the main chunk
+(and ``mp_event_loop`` is called), or the script calls ``mp_event_loop`` or
+``mp.dispatch_events`` directly. This is done to make it possible for a script
+to fully setup event handlers etc. before playback actually starts. In older
+mpv versions, this happened asynchronously.
+
 mp functions
 ------------
 
@@ -74,6 +90,10 @@ The ``mp`` module is preloaded, although it can be loaded manually with
     These two commands are equivalent, except that the first version breaks
     if the filename contains spaces or certain special characters.
 
+    Note that properties are *not* expanded.  You can use either ``mp.command``,
+    the ``expand-properties`` prefix, or the ``mp.get_property`` family of
+    functions.
+
 ``mp.get_property(name [,def])``
     Return the value of the given property as string. These are the same
     properties as used in input.conf. See `Properties`_ for a list of
@@ -94,9 +114,9 @@ The ``mp`` module is preloaded, although it can be loaded manually with
     will always result in a string.
 
 ``mp.get_property_bool(name [,def])``
-    Similar to ``mp.get_property``, but return the property value as boolean.
+    Similar to ``mp.get_property``, but return the property value as Boolean.
 
-    Returns a boolean on success, or ``def, error`` on error.
+    Returns a Boolean on success, or ``def, error`` on error.
 
 ``mp.get_property_number(name [,def])``
     Similar to ``mp.get_property``, but return the property value as number.
@@ -109,7 +129,7 @@ The ``mp`` module is preloaded, although it can be loaded manually with
 
 ``mp.get_property_native(name [,def])``
     Similar to ``mp.get_property``, but return the property value using the best
-    Lua type for the property. Most time, this will return a string, boolean,
+    Lua type for the property. Most time, this will return a string, Boolean,
     or number. Some properties (for example ``chapter-list``) are returned as
     tables.
 
@@ -124,7 +144,7 @@ The ``mp`` module is preloaded, although it can be loaded manually with
 
 ``mp.set_property_bool(name, value)``
     Similar to ``mp.set_property``, but set the given property to the given
-    boolean value.
+    Boolean value.
 
 ``mp.set_property_number(name, value)``
     Similar to ``mp.set_property``, but set the given property to the given
@@ -159,7 +179,7 @@ The ``mp`` module is preloaded, although it can be loaded manually with
     (e.g. ``ctrl+a``).
 
     After calling this function, key presses will cause the function ``fn`` to
-    be called (unless the user overmapped the key with another binding).
+    be called (unless the user remapped the key with another binding).
 
     The ``name`` argument should be a short symbolic string. It allows the user
     to remap the key binding via input.conf using the ``script_message``
@@ -197,7 +217,7 @@ The ``mp`` module is preloaded, although it can be loaded manually with
 
 
     This will print the message when the key ``y`` is pressed. (``x`` will
-    still work, unless the user overmaps it.)
+    still work, unless the user remaps it.)
 
     You can also explicitly send a message to a named script only. Assume the
     above script was using the filename ``fooscript.lua``:
@@ -299,6 +319,20 @@ The ``mp`` module is preloaded, although it can be loaded manually with
             with ``add_timeout()``), this starts the timer from the beginning,
             using the initially configured timeout.
 
+        ``timeout`` (RW)
+            This field contains the current timeout period. This value is not
+            updated as time progresses. It's only used to calculate when the
+            timer should fire next when the timer expires.
+
+            If you write this, you can call ``t:kill() ; t:resume()`` to reset
+            the current timeout to the new one. (``t:stop()`` won't use the
+            new timeout.)
+
+        ``oneshot`` (RW)
+            Whether the timer is periodic (``false``) or fires just once
+            (``true``). This value is used when the timer expires (but before
+            the timer callback function fn is run).
+
 
 ``mp.get_opt(key)``
     Return a setting from the ``--lua-opts`` option. It's up to the user and
@@ -359,7 +393,7 @@ are useful only in special situations.
     control what the Lua script does (instead of being called by the default
     event loop), you can set the global variable ``mp_event_loop`` to your
     own function running the event loop. From your event loop, you should call
-    ``mp.dispatch_events()`` to unqueue and dispatch mpv events.
+    ``mp.dispatch_events()`` to dequeue and dispatch mpv events.
 
     If the ``allow_wait`` parameter is set to ``true``, the function will block
     until the next event is received or the next timer expires. Otherwise (and
@@ -468,6 +502,10 @@ This built-in module provides generic helper functions for Lua, and have
 strictly speaking nothing to do with mpv or video/audio playback. They are
 provided for convenience. Most compensate for Lua's scarce standard library.
 
+``utils.getcwd()``
+    Returns the directory that mpv was launched from. On error, ``nil, error``
+    is returned.
+
 ``utils.readdir(path [, filter])``
     Enumerate all entries at the given path on the filesystem, and return them
     as array. Each entry is a directory entry (without the path).

DOCS/man/mpv.rst

@@ -2,7 +2,7 @@ mpv
 ###
 
 ##############
-a movie player
+a media player
 ##############
 
 :Copyright: GPLv2+
@@ -19,7 +19,7 @@ SYNOPSIS
 DESCRIPTION
 ===========
 
-**mpv** is a movie player based on MPlayer and mplayer2. It supports a wide variety of video
+**mpv** is a media player based on MPlayer and mplayer2. It supports a wide variety of video
 file formats, audio and video codecs, and subtitle types. Special input URL
 types are available to read input from a variety of sources other than disk
 files. Depending on platform, a variety of different video and audio output
@@ -40,16 +40,13 @@ Keyboard Control
 ----------------
 
 LEFT and RIGHT
-    Seek backward/forward 10 seconds. Shift+arrow does a 1 second exact seek
+    Seek backward/forward 5 seconds. Shift+arrow does a 1 second exact seek
     (see ``--hr-seek``).
 
 UP and DOWN
     Seek forward/backward 1 minute. Shift+arrow does a 5 second exact seek (see
     ``--hr-seek``).
 
-PGUP and PGDWN
-    Seek forward/backward 10 minutes.
-
 [ and ]
     Decrease/increase current playback speed by 10%.
 
@@ -69,35 +66,26 @@ p / SPACE
     Pause (pressing again unpauses).
 
 \.
-    Step forward. Pressing once will pause movie, every consecutive press will
+    Step forward. Pressing once will pause, every consecutive press will
     play one frame and then go into pause mode again.
 
 ,
-    Step backward. Pressing once will pause movie, every consecutive press will
+    Step backward. Pressing once will pause, every consecutive press will
     play one frame in reverse and then go into pause mode again.
 
-q / ESC
+q
     Stop playing and quit.
 
 Q
     Like ``q``, but store the current playback position. Playing the same file
     later will resume at the old playback position if possible.
 
-U
-    Stop playing (and quit if ``--idle`` is not used).
-
-\+ and -
-    Adjust audio delay by +/- 0.1 seconds.
-
 / and *
     Decrease/increase volume.
 
 9 and 0
     Decrease/increase volume.
 
-( and )
-    Adjust audio balance in favor of left/right channel.
-
 m
     Mute sound.
 
@@ -107,19 +95,22 @@ m
 \#
     Cycle through the available audio tracks.
 
-TAB (MPEG-TS and libavformat only)
-    Cycle through the available programs.
-
 f
     Toggle fullscreen (see also ``--fs``).
 
+ESC
+    Exit fullscreen mode.
+
 T
     Toggle stay-on-top (see also ``--ontop``).
 
 w and e
     Decrease/increase pan-and-scan range.
 
-o
+o (also P)
+    Show progression bar, elapsed time and total duration on the OSD.
+
+O
     Toggle OSD states: none / seek / seek + timer / seek + timer + total time.
 
 d
@@ -132,12 +123,12 @@ v
 j and J
     Cycle through the available subtitles.
 
-F
-    Toggle displaying "forced subtitles".
-
 x and z
     Adjust subtitle delay by +/- 0.1 seconds.
 
+Ctrl + and Ctrl -
+    Adjust audio delay by +/- 0.1 seconds.
+
 V
     Toggle subtitle VSFilter aspect compatibility mode. See
     ``--ass-vsfilter-aspect-compat`` for more info.
@@ -155,23 +146,18 @@ S
 I
     Show filename on the OSD.
 
-P
-    Show progression bar, elapsed time and total duration on the OSD.
-
-! and @
+PGUP and PGDWN
     Seek to the beginning of the previous/next chapter. In most cases,
     "previous" will actually go to the beginning of the current chapter; see
     ``--chapter-seek-threshold``.
 
-D (``--vo=vdpau``, ``--vf=yadif`` only)
+Shift+PGUP and Shift+PGDWN
+    Seek backward or forward by 10 minutes. (This used to be mapped to
+    PGUP/PGDWN without Shift.)
+
+D
     Activate/deactivate deinterlacer.
 
-A
-    Cycle through the available DVD angles.
-
-c
-    Change YUV colorspace.
-
 (The following keys are valid only when using a video output that supports the
 corresponding adjustment, or the software equalizer (``--vf=eq``).)
 
@@ -187,24 +173,24 @@ corresponding adjustment, or the software equalizer (``--vf=eq``).)
 7 and 8
     Adjust saturation.
 
-(The following keys are valid only on OSX.)
+(The following keys are valid only on OS X.)
 
 command + 0
-    Resize movie window to half its original size.
+    Resize video window to half its original size.
     (On other platforms, you can bind keys to change the ``window-scale``
     property.)
 
 command + 1
-    Resize movie window to its original size.
+    Resize video window to its original size.
 
 command + 2
-    Resize movie window to double its original size.
+    Resize video window to double its original size.
 
 command + f
     Toggle fullscreen (see also ``--fs``).
 
 command + [ and command + ]
-    Set movie window alpha.
+    Set video window alpha.
 
 (The following keys are valid if you have a keyboard with multimedia keys.)
 
@@ -223,12 +209,6 @@ support.)
 h and k
     Select previous/next channel.
 
-n
-    Change norm.
-
-u
-    Change channel list.
-
 Mouse Control
 -------------
 
@@ -285,7 +265,7 @@ need to escape special characters. To work this around, the path can be
 additionally wrapped in the ``%n%string_of_length_n`` syntax (see above).
 
 Some mpv options interpret paths starting with ``~``. Currently, the prefix
-``~~/`` expands to the mpv configuration directory (usually ``~/.mpv/``).
+``~~/`` expands to the mpv configuration directory (usually ``~/.config/mpv/``).
 ``~/`` expands to the user's home directory. (The trailing ``/`` is always
 required.)
 
@@ -338,7 +318,7 @@ Location and Syntax
 You can put all of the options in configuration files which will be read every
 time mpv is run. The system-wide configuration file 'mpv.conf' is in your
 configuration directory (e.g. ``/etc/mpv`` or ``/usr/local/etc/mpv``), the
-user-specific one is ``~/.mpv/config``.
+user-specific one is ``~/.config/mpv/mpv.conf``.
 User-specific options override system-wide options and options given on the
 command line override either. The syntax of the configuration files is
 ``option=<value>``; everything after a *#* is considered a comment. Options
@@ -373,13 +353,13 @@ File-specific Configuration Files
 ---------------------------------
 
 You can also write file-specific configuration files. If you wish to have a
-configuration file for a file called 'movie.avi', create a file named
-'movie.avi.conf' with the file-specific options in it and put it in
-``~/.mpv/``. You can also put the configuration file in the same directory
+configuration file for a file called 'video.avi', create a file named
+'video.avi.conf' with the file-specific options in it and put it in
+``~/.config/mpv/``. You can also put the configuration file in the same directory
 as the file to be played, as long as you give the ``--use-filedir-conf``
 option (either on the command line or in your global config file). If a
 file-specific configuration file is found in the same directory, no
-file-specific configuration is loaded from ``~/.mpv``. In addition, the
+file-specific configuration is loaded from ``~/.config/mpv``. In addition, the
 ``--use-filedir-conf`` option enables directory-specific configuration files.
 For this, mpv first tries to load a mpv.conf from the same directory
 as the file played and then tries to load any file-specific configuration.
@@ -434,6 +414,51 @@ taking screenshots is handled by the VOs, and adding the screenshot filter will
 break hardware decoding. (The filter may still be useful for taking screenshots
 at a certain point within the video chain when using multiple video filters.)
 
+TERMINAL STATUS LINE
+====================
+
+During playback, mpv shows the playback status on the terminal. It looks like
+something like this:
+
+    ``AV: 00:03:12 / 00:24:25 (13%) A-V: -0.000``
+
+The status line can be overridden with the ``--term-status-msg`` option.
+
+The following is a list of things that can show up in the status line. Input
+properties, that can be used to get the same information manually, are also
+listed.
+
+- ``AV:`` or ``V:`` (video only) or ``A:`` (audio only)
+- The current time position in ``HH:MM:SS`` format (``playback-time`` property)
+- The total file duration (or ``00:00:00`` if unknown) (``length`` property)
+- Playback speed, e.g. `` x2.0``. Only visible if the speed is not normal. This
+  is the user-requested speed, and not the actual speed  (usually they should
+  be the same, unless playback is too slow). (``speed`` property.)
+- Playback percentage, e.g. ``(13%)``. How much of the file has been played.
+  Normally calculated out of playback position and duration, but can fallback
+  to other methods (like byte position) if these are not available.
+  (``percent-pos`` property.)
+- The audio/video sync as ``A-V:  0.000``. This is the difference between
+  audio and video time. Normally it should be 0 or close to 0. If it's growing,
+  it might indicate a playback problem. (``avsync`` property.)
+- Total A/V sync change, e.g. ``ct: -0.417``. Normally invisible. Can show up
+  if there is audio "missing", or not enough frames can be dropped. Usually
+  this will indicate a problem. (``total-avsync-change`` property.)
+- Encoding state in ``{...}``, only shown in encoding mode.
+- Decoder-dropped video frames, e.g. ``SD: 2``. Shows up only if the count is
+  not 0. Normally should never show up, unless ``--framedrop`` is set to enable
+  this mode, and your CPU is too slow. (``drop-frame-count`` property.)
+- VO-dropped video frames, e.g. ``D: 4``. Shows up only if the count is not 0.
+  Can grow if the video framerate is higher than that of the display, or if
+  video rendering is too slow. Also can be incremented on "hiccups" and when
+  the video frame couldn't be displayed on time. (``vo-drop-frame-count``
+  property.)
+- Cache state, e.g. ``Cache:  2s+134KB``. Visible if the stream cache is enabled.
+  The first value shows the amount of video buffered in the demuxer in seconds,
+  the second value shows *additional* data buffered in the stream cache in
+  kilobytes. (``demuxer-cache-duration`` and ``cached-used`` properties.)
+
+
 PROTOCOLS
 =========
 
@@ -449,8 +474,8 @@ PROTOCOLS
     Play a path from  Samba share.
 
 ``bd://[title][/device]`` ``--bluray-device=PATH``
-    Play a Blu-Ray disc. Currently, this does not accept iso files. Instead,
-    you must mount the iso file as filesystem, and point ``--bluray-device``
+    Play a Blu-Ray disc. Currently, this does not accept ISO files. Instead,
+    you must mount the ISO file as filesystem, and point ``--bluray-device``
     to the mounted directly.
 
 ``bdnav://[title][/device]``
@@ -539,15 +564,23 @@ ENVIRONMENT VARIABLES
 There are a number of environment variables that can be used to control the
 behavior of mpv.
 
-``HOME``
-    Used to determine mpv config directory: ``$HOME/.mpv``
+``HOME``, ``XDG_CONFIG_HOME``
+    Used to determine mpv config directory. If ``XDG_CONFIG_HOME`` is not set,
+    ``$HOME/.config/mpv`` is used.
+
+    ``$HOME/.mpv`` is always added to the list of config search paths with a
+    lower priority.
+
+``XDG_CONFIG_DIRS``
+    If set, XDG-style system configuration directories are used. Otherwise,
+    the UNIX convention (``PREFIX/etc/mpv/``) is used.
 
 ``TERM``
     Used to determine terminal type.
 
 ``MPV_HOME``
     Directory where mpv looks for user settings. Overrides ``HOME``, and mpv
-    will try to load the config file as ``$MPV_HOME/config``.
+    will try to load the config file as ``$MPV_HOME/mpv.conf``.
 
 ``MPV_VERBOSE`` (see also ``-v`` and ``--msg-level``)
     Set the initial verbosity level across all message modules (default: 0).
@@ -660,18 +693,23 @@ FILES
 ``/usr/local/etc/mpv/mpv.conf``
     mpv system-wide settings (depends on ``--prefix`` passed to configure)
 
-``~/.mpv/config``
+``~/.config/mpv/mpv.conf``
     mpv user settings
 
-``~/.mpv/input.conf``
-    input bindings (see ``--input-keylist`` for the full list)
+``~/.config/mpv/input.conf``
+    key bindings (see `INPUT.CONF`_ section)
 
-``~/.mpv/lua/``
+``~/.config/mpv/lua/``
     All files in this directly are loaded as if they were passed to the
     ``--lua`` option. They are loaded in alphabetical order, and sub-directories
     and files with no ``.lua`` extension are ignored. The ``--load-scripts=no``
     option disables loading these files.
 
+Note that the environment variables ``$XDG_CONFIG_HOME`` and ``$MPV_HOME`` can
+override the standard directory ``~/.config/mpv/``.
+
+Also, the old config location at ``~/.mpv/`` is still read, and if the XDG
+variant does not exist, will still be preferred.
 
 EXAMPLES OF MPV USAGE
 =====================
@@ -689,7 +727,7 @@ Play only chapters 5, 6, 7:
 Play only titles 5, 6, 7:
     ``mpv dvd://5-7``
 
-Play a multiangle DVD:
+Play a multi-angle DVD:
     ``mpv dvd://1 --dvd-angle=2``
 
 Play from a different DVD device:

DOCS/man/osc.rst

@@ -122,11 +122,11 @@ time remaining
     left-click      toggle between total and remaining time
     =============   ================================================
 
-Keybindings
-~~~~~~~~~~~
+Key Bindings
+~~~~~~~~~~~~
 
-These keybindings are active by default if nothing else is already bound to
-these keys. In case of collosision, the function needs to be bound to a
+These key bindings are active by default if nothing else is already bound to
+these keys. In case of collision, the function needs to be bound to a
 different key. See `Script Commands`_ section.
 
 =============   ================================================
@@ -161,7 +161,7 @@ To avoid collisions with other scripts, all options need to be prefixed with
 
 Example::
 
-    --lua-opts=osc-optionA=value1:osc-optionB=value2
+    --lua-opts=osc-optionA=value1,osc-optionB=value2
 
 
 Configurable Options

DOCS/man/vf.rst

@@ -47,7 +47,7 @@ Available filters are:
 
 ``crop[=w:h:x:y]``
     Crops the given part of the image and discards the rest. Useful to remove
-    black bands from widescreen movies.
+    black bands from widescreen videos.
 
     ``<w>,<h>``
         Cropped width and height, defaults to original width and height.
@@ -55,7 +55,7 @@ Available filters are:
         Position of the cropped picture, defaults to center.
 
 ``expand[=w:h:x:y:aspect:round]``
-    Expands (not scales) movie resolution to the given value and places the
+    Expands (not scales) video resolution to the given value and places the
     unscaled original at coordinates x, y.
 
     ``<w>,<h>``
@@ -93,7 +93,7 @@ Available filters are:
 
 ``scale[=w:h:param:param2:chr-drop:noup:arnd``
     Scales the image with the software scaler (slow) and performs a YUV<->RGB
-    colorspace conversion (see also ``--sws``).
+    color space conversion (see also ``--sws``).
 
     All parameters are optional.
 
@@ -120,9 +120,9 @@ Available filters are:
                 0.33:0.33 Mitchell-Netravali spline
                 1.00:0.00 cubic B-spline
 
-            --sws=7 (gaussian): sharpness (0 (soft) - 100 (sharp))
+            --sws=7 (Gaussian): sharpness (0 (soft) - 100 (sharp))
 
-            --sws=9 (lanczos):  filter length (1-10)
+            --sws=9 (Lanczos):  filter length (1-10)
 
     ``<chr-drop>``
         chroma skipping
@@ -222,8 +222,8 @@ Available filters are:
             ====================== =====================
 
 ``noformat[=fmt]``
-    Restricts the colorspace for the next filter without doing any conversion.
-    Unlike the format filter, this will allow any colorspace except the one
+    Restricts the color space for the next filter without doing any conversion.
+    Unlike the format filter, this will allow any color space except the one
     you specify.
 
     .. note:: For a list of available formats, see ``noformat=fmt=help``.
@@ -231,12 +231,23 @@ Available filters are:
     ``<fmt>``
         Format name, e.g. rgb15, bgr24, 420p, etc. (default: 420p).
 
-``pp[=[filter1[:option1[:option2...]]/[-]filter2...]]``
-    Enables the specified chain of postprocessing subfilters. Subfilters must
+``pp=[filter string]``
+    Enables the specified chain of post-processing subfilters. Subfilters must
     be separated by '/' and can be disabled by prepending a '-'. Each
     subfilter and some options have a short and a long name that can be used
-    interchangeably, i.e. ``dr``/``dering`` are the same. All subfilters share
-    common options to determine their scope:
+    interchangeably, i.e. ``dr``/``dering`` are the same.
+
+    .. note::
+
+        Unlike in MPlayer or in earlier versions, you must quote the pp string
+        if it contains ``:`` characters, e.g. ``'--vf=pp=[...]'``.
+
+    .. admonition:: Warning
+
+        This filter is most likely useless on modern HD video. It might be
+        helpful with old low-resolution files.
+
+    All subfilters share common options to determine their scope:
 
     ``a/autoq``
         Automatically switch the subfilter off if the CPU is too slow.
@@ -251,11 +262,6 @@ Available filters are:
 
         ``--vf=pp:help`` shows a list of available subfilters.
 
-    .. note::
-
-        Unlike in MPlayer or in earlier versions, you must quote the pp string
-        if it contains ``:`` characters, e.g. ``'--vf=pp=[...]'``.
-
     Available subfilters are:
 
     ``hb/hdeblock[:difference[:flatness]]``
@@ -336,7 +342,7 @@ Available filters are:
         filtering every second line with a (-1 4 2 4 -1) filter.
 
     ``l5/lowpass5``
-        Vertically applied FIR lowpass deinterlacing filter that deinterlaces
+        Vertically applied FIR low-pass deinterlacing filter that deinterlaces
         the given block by filtering all lines with a (-1 2 6 2 -1) filter.
 
     ``fq/forceQuant[:quantizer]``
@@ -356,7 +362,7 @@ Available filters are:
 
     .. note::
 
-        This filter is only available if FFmpeg/libav has been compiled
+        This filter is only available if FFmpeg/Libav has been compiled
         with libpostproc enabled.
 
     .. admonition:: Examples
@@ -443,7 +449,7 @@ Available filters are:
         temporal noise (noise pattern changes between frames)
 
     ``uniform``
-        uniform noise (gaussian otherwise)
+        uniform noise (Gaussian otherwise)
 
     ``hq``
         high quality (slightly better looking, slightly slower) - not available
@@ -503,7 +509,7 @@ Available filters are:
         :1: linear interpolation (default)
 
 ``unsharp[=lx:ly:la:cx:cy:ca]``
-    unsharp mask / gaussian blur
+    unsharp mask / Gaussian blur
 
     ``l`` is for the luma component, ``c`` for the chroma component. ``x``/``y``
     is the filter size. ``a`` is the amount.
@@ -617,7 +623,7 @@ Available filters are:
 
 ``phase[=t|b|p|a|u|T|B|A|U][:v]``
     Delay interlaced video by one field time so that the field order changes.
-    The intended use is to fix PAL movies that have been captured with the
+    The intended use is to fix PAL videos that have been captured with the
     opposite field order to the film-to-video transfer. The options are:
 
     ``t``
@@ -769,8 +775,8 @@ Available filters are:
             anaglyph red/cyan color (red filter on left eye, cyan filter on
             right eye)
         ``arcd`` or ``anaglyph_red_cyan_dubois``
-            anaglyph red/cyan color optimized with the least squares
-            projection of dubois (red filter on left eye, cyan filter on right
+            anaglyph red/cyan color optimized with the least-squares
+            projection of Dubois (red filter on left eye, cyan filter on right
             eye)
         ``agmg`` or ``anaglyph_green_magenta_gray``
             anaglyph green/magenta gray (green filter on left eye, magenta
@@ -803,7 +809,7 @@ Available filters are:
 
 ``gradfun[=strength[:radius|:size=<size>]]``
     Fix the banding artifacts that are sometimes introduced into nearly flat
-    regions by truncation to 8bit color depth. Interpolates the gradients that
+    regions by truncation to 8-bit color depth. Interpolates the gradients that
     should go where the bands are, and dithers them.
 
     ``<strength>``
@@ -969,3 +975,9 @@ Available filters are:
             Use default VDPAU scaling (default).
         1-9
             Apply high quality VDPAU scaling (needs capable hardware).
+
+``buffer=<num>``
+    Buffer ``<num>`` frames in the filter chain. This filter is probably pretty
+    useless, except for debugging. (Note that this won't help smoothing out
+    latencies with decoding, because the filter will never output a frame if
+    the buffer isn't full, except on EOF.)

DOCS/man/vo.rst

@@ -38,36 +38,36 @@ Available video output drivers are:
     .. note:: This driver is for compatibility with old systems.
 
     ``adaptor=<number>``
-        Select a specific XVideo adaptor (check xvinfo results).
+        Select a specific XVideo adapter (check xvinfo results).
     ``port=<number>``
         Select a specific XVideo port.
     ``ck=<cur|use|set>``
-        Select the source from which the colorkey is taken (default: cur).
+        Select the source from which the color key is taken (default: cur).
 
         cur
-          The default takes the colorkey currently set in Xv.
+          The default takes the color key currently set in Xv.
         use
-          Use but do not set the colorkey from mpv (use the ``--colorkey``
+          Use but do not set the color key from mpv (use the ``--colorkey``
           option to change it).
         set
-          Same as use but also sets the supplied colorkey.
+          Same as use but also sets the supplied color key.
 
     ``ck-method=<man|bg|auto>``
-        Sets the colorkey drawing method (default: man).
+        Sets the color key drawing method (default: man).
 
         man
-          Draw the colorkey manually (reduces flicker in some cases).
+          Draw the color key manually (reduces flicker in some cases).
         bg
-          Set the colorkey as window background.
+          Set the color key as window background.
         auto
-          Let Xv draw the colorkey.
+          Let Xv draw the color key.
 
     ``colorkey=<number>``
-        Changes the colorkey to an RGB value of your choice. ``0x000000`` is
+        Changes the color key to an RGB value of your choice. ``0x000000`` is
         black and ``0xffffff`` is white.
 
     ``no-colorkey``
-        Disables colorkeying.
+        Disables color-keying.
 
 ``x11`` (X11 only)
     Shared memory video output driver without hardware acceleration that works
@@ -82,7 +82,7 @@ Available video output drivers are:
     .. note::
 
         Earlier versions of mpv (and MPlayer, mplayer2) provided sub-options
-        to tune vdpau postprocessing, like ``deint``, ``sharpen``, ``denoise``,
+        to tune vdpau post-processing, like ``deint``, ``sharpen``, ``denoise``,
         ``chroma-deint``, ``pullup``, ``hqscaling``. These sub-options are
         deprecated, and you should use the ``vdpaupp`` video filter instead.
 
@@ -173,7 +173,7 @@ Available video output drivers are:
         to a YUV format before the VO. Sometimes useful to force availability
         of certain YUV-only features, like video equalizer or deinterlacing.
 
-    Using the VDPAU frame queueing functionality controlled by the queuetime
+    Using the VDPAU frame queuing functionality controlled by the queuetime
     options makes mpv's frame flip timing less sensitive to system CPU load and
     allows mpv to start decoding the next frame(s) slightly earlier, which can
     reduce jitter caused by individual slow-to-decode frames. However, the
@@ -259,13 +259,6 @@ Available video output drivers are:
 
     .. note:: This driver is for compatibility with old systems.
 
-``corevideo`` (Mac OS X 10.6 and later)
-    Mac OS X CoreVideo video output driver. Uses the CoreVideo APIs to fill
-    PixelBuffers and generate OpenGL textures from them (useful as a fallback
-    for ``opengl``).
-
-    .. note:: This driver is for compatibility with old systems.
-
 ``opengl``
     OpenGL video output driver. It supports extended scaling methods, dithering
     and color management.
@@ -282,7 +275,7 @@ Available video output drivers are:
 
     Hardware decoding over OpenGL-interop is supported to some degree. Note
     that in this mode, some corner case might not be gracefully handled, and
-    colorspace conversion and chroma upsampling is generally in the hand of
+    color space conversion and chroma upsampling is generally in the hand of
     the hardware decoder APIs.
 
     ``lscale=<filter>``
@@ -297,6 +290,9 @@ Available video output drivers are:
         ``lanczos3``
             Lanczos with radius=3.
 
+        ``lanczos``
+            Generic Lanczos scaling filter. Set radius with ``lradius``.
+
         ``spline36``
             This is the default when using ``opengl-hq``.
 
@@ -316,6 +312,14 @@ Available video output drivers are:
             Mitchell-Netravali. The ``b`` and ``c`` parameters can be set with
             ``lparam1`` and ``lparam2``. Both are set to 1/3 by default.
 
+        ``gaussian``
+            Gaussian filter with a parameter ``p`` for sharpness control.
+            ``p`` can be set to float number between 1(blurry) and 100(sharp)
+            and has a default value of about 28.8 (see ``lparam1``).
+
+            Note that for extremely small value of ``p``, a large filter radius
+            might be required to avoid unintended artifacts (see ``lradius``).
+
 
         There are some more filters. For a complete list, pass ``help`` as
         value, e.g.::
@@ -329,6 +333,16 @@ Available video output drivers are:
     ``lparam2=<value>``
         See ``lparam1``.
 
+    ``lradius=<r>``
+        Set radius for filters listed below, must be a float number between 1.0
+        and 8.0. Defaults to be 2.0 if not specified.
+
+            ``sinc``, ``lanczos``, ``blackman``, ``gaussian``
+
+        Note that depending on filter implementation details and video scaling
+        ratio, the radius that actually being used might be different
+        (most likely being increased a bit).
+
     ``scaler-resizes-only``
         Disable the scaler if the video image is not resized. In that case,
         ``bilinear`` is used instead whatever is set with ``lscale``. Bilinear
@@ -349,7 +363,7 @@ Available video output drivers are:
         green-magenta
             Convert side by side input to full-color green-magenta stereo.
         quadbuffer
-            Convert side by side input to quadbuffered stereo. Only supported
+            Convert side by side input to quad buffered stereo. Only supported
             by very few OpenGL cards.
 
     ``srgb``
@@ -430,6 +444,11 @@ Available video output drivers are:
         RGB. If chroma is not subsampled, this option is ignored, and the
         luma scaler is used instead. Setting this option is often useless.
 
+    ``cparam1``, ``cparam2``, ``cradius``
+        Set filter parameters and radius for ``cscale``.
+
+        See ``lparam1``, ``lparam2`` and ``lradius``.
+
     ``fancy-downscaling``
         When using convolution based filters, extend the filter size
         when downscaling. Trades quality for reduced downscaling performance.
@@ -439,7 +458,16 @@ Available video output drivers are:
         Borders will be distorted due to filtering.
 
     ``glfinish``
-        Call ``glFinish()`` before swapping buffers
+        Call ``glFinish()`` before and after swapping buffers (default: disabled).
+        Slower, but might help getting better results when doing framedropping.
+        The details depend entirely on the OpenGL driver.
+
+    ``waitvsync``
+        Call ``glXWaitVideoSyncSGI`` after each buffer swap (default: disabled).
+        This may or may not help with video timing accuracy and frame drop. It's
+        possible that this makes video output slower, or has no effect at all.
+
+        X11 only.
 
     ``sw``
         Continue even if a software renderer is detected.
@@ -452,7 +480,7 @@ Available video output drivers are:
         auto
             auto-select (default)
         cocoa
-            Cocoa/OSX
+            Cocoa/OS X
         win
             Win32/WGL
         x11
@@ -479,7 +507,7 @@ Available video output drivers are:
 
     ``gamma=<0.0..10.0>``
         Set a gamma value. If gamma is adjusted in other ways (like with
-        the ``--gamma`` option or keybindings and the ``gamma`` property), the
+        the ``--gamma`` option or key bindings and the ``gamma`` property), the
         value is multiplied with the other gamma value.
 
         Setting this value to 1.0 can be used to always enable gamma control.
@@ -487,7 +515,7 @@ Available video output drivers are:
 
     ``icc-profile=<file>``
         Load an ICC profile and use it to transform linear RGB to screen output.
-        Needs LittleCMS2 support compiled in. This option overrides the ``srgb``
+        Needs LittleCMS 2 support compiled in. This option overrides the ``srgb``
         property, as using both is somewhat redundant. It also enables linear
         light scaling.
 
@@ -500,12 +528,12 @@ Available video output drivers are:
 
     ``icc-cache=<file>``
         Store and load the 3D LUT created from the ICC profile in this file.
-        This can be used to speed up loading, since LittleCMS2 can take a while
+        This can be used to speed up loading, since LittleCMS 2 can take a while
         to create the 3D LUT. Note that this file contains an uncompressed LUT.
         Its size depends on the ``3dlut-size``, and can be very big.
 
     ``icc-intent=<value>``
-        Specifies the ICC Intent used for transformations between colorspaces.
+        Specifies the ICC Intent used for transformations between color spaces.
         This affects the rendering when using ``icc-profile`` or ``srgb`` and
         also affects the way DCP XYZ content gets converted to RGB.
 
@@ -524,7 +552,7 @@ Available video output drivers are:
         for mastering instead of the true curve. Most notably, anything in the
         Apple ecosystem uses this approximation - including all programs
         compatible with it. It's a sound idea to try enabling this flag first
-        when watching movies and shows to see if things look better that way.
+        when watching videos and shows to see if things look better that way.
 
         This only affects the output when using either ``icc-profile`` or ``srgb``.
 
@@ -562,7 +590,7 @@ Available video output drivers are:
 
     This is equivalent to::
 
-        --vo=opengl:lscale=spline36:dither-depth=auto:fbo-format=rgb16
+        --vo=opengl:lscale=spline36:dither-depth=auto:fbo-format=rgba16
 
     Note that some cheaper LCDs do dithering that gravely interferes with
     ``opengl``'s dithering. Disabling dithering with ``dither-depth=no`` helps.
@@ -571,7 +599,7 @@ Available video output drivers are:
     can achieve better quality or performance by changing the ``fbo-format``
     suboption to ``rgb16f``, ``rgb32f`` or ``rgb``. Known problems include
     Mesa/Intel not accepting ``rgb16``, Mesa sometimes not being compiled with
-    float texture support, and some OSX setups being very slow with ``rgb16``
+    float texture support, and some OS X setups being very slow with ``rgb16``
     but fast with ``rgb32f``.
 
 ``opengl-old``
@@ -706,7 +734,7 @@ Available video output drivers are:
         2
             Convert side by side input to full-color green-magenta stereo.
         3
-            Convert side by side input to quadbuffered stereo. Only supported
+            Convert side by side input to quad buffered stereo. Only supported
             by very few OpenGL cards.
 
     The following options are only useful if writing your own fragment programs.
@@ -748,7 +776,7 @@ Available video output drivers are:
         auto
             auto-select (default)
         cocoa
-            Cocoa/OSX
+            Cocoa/OS X
         win
             Win32/WGL
         x11

DOCS/release-policy.md

@@ -24,11 +24,14 @@ Release procedure
 
 - Create and/or update the ``VERSION`` file.
 
+- Update ``DOCS/client-api-changes.rst`` (on major releases).
+
 - Create tag v0.X.Y.
 
 - Push branch and tag to GitHub.
 
-- Create a new GitHub release using the content of ``RELEASE_NOTES``.
+- Create a new GitHub release using the content of ``RELEASE_NOTES`` related to
+  the new version.
 
 Release notes template
 ----------------------
@@ -36,6 +39,9 @@ Release notes template
 Here is a template that can be used for writing the ``RELEASE_NOTES`` file.
 
 ```markdown
+Release 0.X.Y
+=============
+
 Changes
 -------
 
@@ -55,3 +61,11 @@ This listing is not complete. There are many more bug fixes and changes. The
 complete change log can be viewed by running ``git log <start>..<end>`` in
 the git repository.
 ```
+
+Note that the "Release 0.X.Y" title should be removed when creating a new GitHub
+release.
+
+When creating a new point release its changes should be added on top of the
+``RELEASE_NOTES`` file (with the appropriate title) so that all the changes in
+the current 0.X branch will be included. This way the ``RELEASE_NOTES`` file
+can be used by distributors as changelog for point releases too.

DOCS/tech-overview.txt

@@ -1,4 +1,4 @@
-This file intends to give a big picture of how mpv is structured.
+This file intends to give a big picture overview of how mpv is structured.
 
 player/*.c:
     Essentially makes up the player applications, including the main() function
@@ -17,12 +17,12 @@ player/*.c:
         * check help options etc. (call handle_help_options()), possibly exit
         * call play_files() function that works down the playlist:
             * run idle loop (idle_loop()), until there are files in the
-                playlist or an exit command was given (slave mode only)
+                playlist or an exit command was given (only if --idle it set)
             * actually load and play a file in play_current_file():
                 * run all the dozens of functions to load the file and
                   initialize playback
                 * run a small loop that does normal playback, until the file is
-                  done or a slave command terminates playback
+                  done or a command terminates playback
                   (on each iteration, run_playloop() is called, which is rather
                    big and complicated - it decodes some audio and video on
                    each frame, waits for input, etc.)
@@ -31,15 +31,28 @@ player/*.c:
             * loop, or exit if no next file or quit is requested
               (see enum stop_play_reason)
         * call exit_player_with_rc()
+    * run_playloop():
+        * calls fill_audio_out_buffers()
+            This checks whether new audio needs to be decoded, and pushes it
+            to the AO.
+        * calls write_video()
+            Decode new video, and push it to the VO.
+        * determines whether playback of the current file has ended
+        * determines when to start playback after seeks
+        * and calls a whole lot of other stuff
+            (Really, this function does everything.)
 
     Things worth saying about the playback core:
-    - the currently played tracks are in sh_video and sh_audio
+    - most state is in MPContext (core.h), which is not available to the
+      subsystems
+    - the currently played tracks are in mpctx->current_tracks, and decoder
+      state in d_video/d_audio/d_sub
     - the timeline stuff is used only with MKV ordered chapters (and some other
       minor features: cue, edl)
-    - most state is in MPContext (mp_core.h), which is not available to the
-      subsystems
     - the other subsystems rarely call back into the frontend, and the frontend
       polls them instead (probably a good thing)
+    - one exceptions are wakeup callbacks, which notify a "higher" component
+      of a changed situation in a subsystem
 
     I like to call the player/*.c files the "frontend".
 
@@ -82,7 +95,7 @@ talloc.h & talloc.c:
              (See ta/README for details.)
 
 player/command.c:
-    This contains the implementation for slave commands and properties.
+    This contains the implementation for client API commands and properties.
     Properties are essentially dynamic variables changed by certain commands.
     This is basically responsible for all user commands, like initiating
     seeking, switching tracks, etc. It calls into other player/*.c files,
@@ -92,9 +105,10 @@ player/core.h:
     Data structures and function prototypes for most of player/*.c. They are
     usually not accessed by other parts of mpv for the sake of modularization.
 
-    Note that there are lots of global variables floating around everywhere
-    else. This is an ongoing transition, and eventually there should be no
-    global variables anymore.
+player/client.c:
+    This implements the client API (libmpv/client.h). For the most part, this
+    just calls into other parts of the player. This also manages a ringbuffer
+    of events from player to clients.
 
 options/options.h, options/options.c
     options.h contains the global option struct MPOpts. The option declarations
@@ -102,7 +116,7 @@ options/options.h, options/options.c
     options.c. Most default values for options and MPOpts are in
     mp_default_opts at the end of options.c.
 
-    MPOpts is unfortunarely quite monolithic, and virtually accessed by
+    MPOpts is unfortunately quite monolithic, and virtually accessed by
     everything.But some components (like video outputs and video filters) have
     their own sub-option tables separate from MPOpts.
 
@@ -110,19 +124,18 @@ options/options.h, options/options.c
     parser-mpcmd.c, and uses the option table in options.c.
 
 input/input.c:
-    This translates keyboard input comming from libvo and other sources (such
-    as remote control devices like Apple IR or slave mode commands) to the
+    This translates keyboard input comming from VOs and other sources (such
+    as remote control devices like Apple IR or client API commands) to the
     key bindings listed in the user's (or the builtin) input.conf and turns
     them into items of type struct mp_cmd. These commands are queued, and read
     by playloop.c. They get pushed with run_command() to command.c.
 
-    Note that keyboard input and slave mode input are essentially the same
-    things. Just looking at input.conf should make this clear. (The other
-    direction of slave mode communication, mpv to application, consists of
-    random mp_msg() calls all over the code in all parts of the player.)
+    Note that keyboard input and commands used by the client API are the same.
+    The client API only uses the command parser though, and has its own queue
+    of input commands somewhere else.
 
 common/msg.h:
-    All terminal output should go through mp_msg().
+    All terminal output must go through mp_msg().
 
 stream/*:
     File input is implemented here. stream.h/.c provides a simple stream based
@@ -181,7 +194,7 @@ video/out/:
     vo_opengl can pick a windowing system at runtime, e.g. the same binary can
     provide both X11 and Cocoa support on OSX.
 
-    VOs can be reconfigured at runtime. A config() call can change the video
+    VOs can be reconfigured at runtime. A vo_config() call can change the video
     resolution and format, without destroying the window.
 
     vo_vdpau and vo_opengl should be taken as reference.
@@ -203,9 +216,12 @@ audio/filter/:
 audio/out/:
     Audio outputs.
 
-    Unlike VOs, AOs can't be reconfigured on a format change. Without
-    --gapless-audio, even playing a new file will close and re-open the audio
-    device.
+    Unlike VOs, AOs can't be reconfigured on a format change. On audio format
+    changes, the AO will simply be closed and re-opened.
+
+    There are wrappers to support for two types of audio APIs: push.c and
+    pull.c. ao.c calls into one of these. They contain generic code to deal
+    with the data flow these APIs impose.
 
     Note that mpv synchronizes the video to the audio. That's the reason
     why buggy audio drivers can have a bad influence on playback quality.
@@ -213,17 +229,19 @@ audio/out/:
 sub/:
     Contains subtitle and OSD rendering.
 
-    sub.c/.h is actually the OSD code. It queries dec_sub.c to retrieve
+    osd.c/.h is actually the OSD code. It queries dec_sub.c to retrieve
     decoded/rendered subtitles. osd_libass.c is the actual implementation of
     the OSD text renderer (which uses libass, and takes care of all the tricky
     fontconfig/freetype API usage and text layouting).
 
-    Subtitle loading is now in demux/ instead. demux_libass.c wraps loading
-    .ass subtitles via libass. demux_lavf.c loads most subtitle types via
-    FFmpeg. demux_subreader.c is the old MPlayer code. It's used as last
-    fallback, or to handle some text subtitle types on Libav. (It also can
-    load UTF-16 encoded subtitles without requiring the use of -sub-codepage.)
-    demux_subreader.c should eventually go away (maybe).
+    The VOs call osd.c to render OSD and subtitle (via e.g. osd_draw()). osd.c
+    in turn asks dec_sub.c for subtitle overlay bitmaps, which relays the
+    request to one of the sd_*.c subtitle decoders/renderers.
+
+    Subtitle loading is in demux/. The MPlayer subreader.c is mostly gone - parts
+    of it survive in demux_subreader.c. It's used as last fallback, or to handle
+    some text subtitle types on Libav. It should go away eventually. Normally,
+    subtitles are loaded via demux_lavf.c.
 
     The subtitles are passed to dec_sub.c and the subtitle decoders in sd_*.c
     as they are demuxed. All text subtitles are rendered by sd_ass.c. If text

README.md

@@ -136,4 +136,4 @@ To contact the `mpv` team in private write to `mpv-team@googlegroups.com`.
 [mpv-users]: https://groups.google.com/forum/?hl=en#!forum/mpv-users
 [mpv-devel]: https://groups.google.com/forum/?hl=en#!forum/mpv-devel
 [ffmpeg_vs_libav]: https://github.com/mpv-player/mpv/wiki/FFmpeg-versus-Libav
-[release-policy]: https://github.com/mpv-player/mpv/blob/master/release-policy.md
+[release-policy]: https://github.com/mpv-player/mpv/blob/master/DOCS/release-policy.md

RELEASE_NOTES

@@ -0,0 +1,126 @@
+Release 0.6.2
+=============
+
+Changes
+-------
+
+- The --ass-style-override option now requires a newer version of libass.
+
+Bug fixes
+---------
+
+- Correctly signal EOF when using --frames.
+- Strictly endorce the limit imposed by --frames (i.e. avoid producing an extra
+  frame at the end).
+- Fix \x-style escaping when parsing commands.
+- Correctly handle ^h key on the terminal (this should fix the when inserted
+  fromBackspace key on new versions of xterm).
+- Don't put the player to sleep after audio reinit.
+- Properly exit if both audio and video fail to initialize.
+
+This listing is not complete. There are many more bug fixes and changes. The
+complete change log can be viewed by running ``git log a6eb071..0a40fb0`` in
+the git repository.
+
+Release 0.6.1
+=============
+
+Changes
+-------
+
+- Don't show message on "osd" command by default (this essentially reverts a
+  change introduced in the last release).
+- Add an alternative input configuration that reverts the key binding changes
+  introduced by previous releases (see etc/restore-old-bindings.conf in the
+  sources).
+- Automatically install example input configuration files.
+
+Bug fixes
+---------
+
+- Make stream language work with Bluray as well (the previous release
+  introduced this for DVDs, but it didn't work for Blurays).
+- Don't try to create surfaces of size 0 in the VDPAU vo.
+- Don't print audio/video init failure message twice.
+- Fix playback on the root window on X11 (i.e. when using --wid=0).
+- Fix the autodetection of the input format in the vf_stereo3d filter when mpv
+  is built against libav (this makes --video-stereo-mode work with libav as
+  well).
+- Fix compilation with PulseAudio 1.0.
+- Fix compilation with clang.
+- Try harder to decode cover art picture only once.
+- Fix cursor autohide (--cursor-autohide) with audio-only files on file change
+  (e.g. when skipping to the next playlist file).
+
+This listing is not complete. There are many more bug fixes and changes. The
+complete change log can be viewed by running ``git log 061f509..2ffdc0d`` in
+the git repository.
+
+Release 0.6.0
+=============
+
+Changes
+-------
+
+- Map ESC to exiting fullscreen.
+- Run the vo inside its own thread.
+- Enable the demuxer thread by default.
+- Prefer the opengl vo over the vdpau one.
+- Complete profile names in the zsh completion script.
+- Raise dsound ao buffer size to 200ms and make it configurable.
+- Bind ctrl+c to quit, so that it works both in the terminal and the player
+  window.
+- Automatically load playlists without needing --load-unsafe-playlists or
+  --playlist. Unsafe URLs are still rejected (e.g. remote playlists can't load
+  local files).
+- Remove quvi-format property (which never really worked).
+- Wait until all Lua scripts are loaded on start.
+- Use separate threads for input sources (terminal, lirc, joystick) instead of
+  using a central I/O loop.
+- Complete values for options that take file names in the zsh completion script.
+- Use libwaio for pipe input on Windows (if you want --input-file to work on
+  Windows you need libwaio as an additional dependency).
+- Remove the --volstep option (users should change the argument to the "add"
+  function in input.conf instead).
+- Always show the OSD message when changing volume, gamma, panscan, etc along
+  with the OSD bar.
+- When changing a property always show it on the OSD (except for "pause" and
+  "fullscreen").
+- Prefer the libavcodec decoder over libmpg123 for mp3 decoding (ffmpeg/libav
+  support for gapless audio works better for mpv's use-case).
+- Allow passing number of loops to --loop-file.
+- Change --keep-open option semantics to not terminate the player when playing
+  or seeking past the end of the file and there's no next file to be played.
+- Disable the sndio ao by default and prefer other outputs even if it's enabled.
+
+New features
+------------
+
+- Add new framedrop mode that limits the frequency of new video frames displayed
+  based on the display framerate (this is enabled by default, the old
+  decoder-based mode can be enabled with --framedrop=decoder).
+- Add --demuxer-readahead-secs option to control the readahead buffer using a
+  duration value.
+- Add estimated-frame-count and estimated-frame-number properties.
+- Add --cache-secs option which overrides the value of --demuxer-readahead-secs
+  if the stream cache is active.
+- Add demuxer-cache-duration and demuxer-cache-idle properties.
+- Make --cache-file create a temporary file if the string "TMP" is passed
+  instead of a file name.
+- Add support for Matroska 3D (if a file is marked as 3D using the StereoMode
+  element, an automatic conversion filter is added).
+- Add --osd-playing-msg option to show a message on the OSD when the playback
+  starts.
+- Add --media-title option to force the content of the media-title property.
+- Add --osd-msg1, --osd-msg2 and --osd-msg3 options to override the OSD message
+  for the corresponding OSD level (before only level 3 could be overridden with
+  --osd-status-message).
+- Add support for digital passthrough in the pulse ao.
+- Add support for opening DVD and Bluray directories directly.
+- Add --input-x11-keyboard option to control whether keyboard input should be
+  enabled on the X11 window ot not (this is disabled by default when using the
+  libmpv library).
+
+This listing is not complete. There are many more bug fixes and changes. The
+complete change log can be viewed by running ``git log a1be3cf..61ebe61`` in
+the git repository.

TOOLS/dylib-unhell.py

@@ -15,10 +15,10 @@ def is_user_lib(objfile, libname):
     return not sys_re.match(libname) and \
            not usr_re.match(libname) and \
            not exe_re.match(libname) and \
-           not "libobjc" in libname and \
-           not "libSystem" in libname and \
-           not "libc" in libname and \
-           not "libgcc" in libname and \
+           not "libobjc." in libname and \
+           not "libSystem." in libname and \
+           not "libc." in libname and \
+           not "libgcc." in libname and \
            not os.path.basename(objfile) in libname
 
 def otool(objfile):

TOOLS/gen-x11-icon.sh

@@ -5,8 +5,8 @@
 
 conv() {
     echo
-    identify -format "icon: %w %h" $1
-    convert $1 -depth 8 rgba:-
+    identify -format "icon: %w %h" "$1"
+    convert "$1" -depth 8 rgba:-
 }
 
 (echo "# File generated by gen-x11-icon.sh" ;

TOOLS/idet.sh

@@ -1,23 +1,15 @@
 #!/bin/sh
 
-case "$0" in
-    */*)
-        MYDIR=${0%/*}
-        ;;
-    *)
-        MYDIR=.
-        ;;
-esac
+: "${MPV:=mpv}"
+: "${ILDETECT_MPV:=$MPV}"
+: "${ILDETECT_MPVFLAGS:=--start=35% --length=35}"
+: "${ILDETECT_DRY_RUN:=}"
+: "${ILDETECT_QUIET:=}"
+: "${ILDETECT_RUN_INTERLACED_ONLY:=}"
+: "${ILDETECT_FORCE_RUN:=}"
 
-: ${MPV:=mpv}
-: ${ILDETECT_MPV:=$MPV}
-: ${ILDETECT_MPV:=$MPV}
-: ${ILDETECT_MPVFLAGS:=--start=35% --length=35}
-: ${ILDETECT_DRY_RUN:=}
-: ${ILDETECT_QUIET:=}
-: ${ILDETECT_RUN_INTERLACED_ONLY:=}
-: ${ILDETECT_FORCE_RUN:=}
-: ${MAKE:=make}
+# This script uses ffmpeg's "idet" filter for interlace detection. In the
+# long run this should replace ildetect.sh+ildetect.so.
 
 # exit status:
 # 0 progressive
@@ -30,126 +22,135 @@ esac
 testfun()
 {
     $ILDETECT_MPV "$@" \
-        --vf=lavfi="[idet]" --msg-level ffmpeg=v \
+        --vf-add=lavfi="[idet]" --msg-level ffmpeg=v \
         --o= --vo=null --no-audio --untimed \
         $ILDETECT_MPVFLAGS \
         | { if [ -n "$ILDETECT_QUIET" ]; then cat; else tee /dev/stderr; fi } \
-        | grep "Parsed_idet_0: Multi frame detection: " | tail -n 1
+        | grep "Parsed_idet_0: Multi frame detection: "
 }
 
 judge()
 {
-    out=`testfun "$@"`
-
-    tff=${out##* TFF:}; tff=${tff%% *}
-    bff=${out##* BFF:}; bff=${bff%% *}
-    progressive=${out##* Progressive:}; progressive=${progressive%% *}
-    undetermined=${out##* Undetermined:}; undetermined=${undetermined%% *}
-
-    case "$tff$bff$progressive$undetermined" in
-        *[!0-9]*)
-            echo >&2 "ERROR: Unrecognized idet output: $out"
-            exit 16
-            ;;
-    esac
+    tff=0
+    bff=0
+    progressive=0
+    undetermined=0
+    while IFS= read -r out; do
+        tff1=${out##* TFF:}; tff1=${tff1%% *}
+        bff1=${out##* BFF:}; bff1=${bff1%% *}
+        progressive1=${out##* Progressive:}; progressive1=${progressive1%% *}
+        undetermined1=${out##* Undetermined:}; undetermined1=${undetermined1%% *}
+        case "$tff1$bff1$progressive1$undetermined1" in
+            *[!0-9]*)
+                printf >&2 'ERROR: Unrecognized idet output: %s\n' "$out"
+                exit 16
+                ;;
+        esac
+        tff=$((tff + tff1))
+        bff=$((bff + bff1))
+        progressive=$((progressive + progressive1))
+        undetermined=$((undetermined + undetermined1))
+    done <<EOF
+$(testfun "$@")
+EOF
 
     interlaced=$((bff + tff))
     determined=$((interlaced + progressive))
 
-    if [ $undetermined -gt $determined ] || [ $determined -lt 250 ]; then
+    if [ "$undetermined" -gt "$determined" ] || [ "$determined" -lt 250 ]; then
         echo >&2 "ERROR: Less than 50% or 250 frames are determined."
         [ -n "$ILDETECT_FORCE_RUN" ] || exit 8
         echo >&2 "Assuming interlacing."
-        if [ $tff -gt $((bff * 10)) ]; then
-            verdict=interlaced-tff
-        elif [ $bff -gt $((tff * 10)) ]; then
-            verdict=interlaced-bff
+        if [ "$tff" -gt $((bff * 10)) ]; then
+            verdict="interlaced-tff"
+        elif [ "$bff" -gt $((tff * 10)) ]; then
+            verdict="interlaced-bff"
         else
-            verdict=interlaced
+            verdict="interlaced"
         fi
-    elif [ $((interlaced * 20)) -gt $progressive ]; then
+    elif [ $((interlaced * 20)) -gt "$progressive" ]; then
         # At least 5% of the frames are interlaced!
-        if [ $tff -gt $((bff * 10)) ]; then
-            verdict=interlaced-tff
-        elif [ $bff -gt $((tff * 10)) ]; then
-            verdict=interlaced-bff
+        if [ "$tff" -gt $((bff * 10)) ]; then
+            verdict="interlaced-tff"
+        elif [ "$bff" -gt $((tff * 10)) ]; then
+            verdict="interlaced-bff"
         else
             echo >&2 "ERROR: Content is interlaced, but can't determine field order."
             [ -n "$ILDETECT_FORCE_RUN" ] || exit 8
             echo >&2 "Assuming interlacing with default field order."
-            verdict=interlaced
+            verdict="interlaced"
         fi
     else
-        # Likely progrssive.
-        verdict=progressive
+        # Likely progressive
+        verdict="progressive"
     fi
 
-    echo "$verdict"
+    printf '%s\n' "$verdict"
 }
 
-judge "$@"
+judge "$@" --vf-clr
 case "$verdict" in
     progressive)
         [ -n "$ILDETECT_DRY_RUN" ] || \
             [ -n "$ILDETECT_RUN_INTERLACED_ONLY" ] || \
             $ILDETECT_MPV "$@"
         r=$?
-        [ $r -eq 0 ] || exit $(($r | 16))
+        [ $r -eq 0 ] || exit $((r | 16))
         exit 0
         ;;
     interlaced-tff)
-        judge "$@" --vf-pre=pullup --field-dominance=top
+        judge "$@" --vf-clr --vf-pre=pullup --field-dominance=top
         case "$verdict" in
             progressive)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=pullup --field-dominance=top
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 1
                 ;;
             *)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=yadif --field-dominance=top
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 2
                 ;;
         esac
         ;;
     interlaced-bff)
-        judge "$@" --vf-pre=pullup --field-dominance=bottom
+        judge "$@" --vf-clr --vf-pre=pullup --field-dominance=bottom
         case "$verdict" in
             progressive)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=pullup --field-dominance=bottom
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 1
                 ;;
             *)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=yadif --field-dominance=bottom
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 2
                 ;;
         esac
         ;;
     interlaced)
-        judge "$@" --vf-pre=pullup
+        judge "$@" --vf-clr --vf-pre=pullup
         case "$verdict" in
             progressive)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=pullup
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 1
                 ;;
             *)
                 [ -n "$ILDETECT_DRY_RUN" ] || \
                     $ILDETECT_MPV "$@" --vf-pre=yadif
                 r=$?
-                [ $r -eq 0 ] || exit $(($r | 16))
+                [ $r -eq 0 ] || exit $((r | 16))
                 exit 2
                 ;;
         esac

TOOLS/lib/Parse/Matroska/Definitions.pm

@@ -265,6 +265,7 @@ sub define_matroska {
                         elem('DisplayUnit',   '54b2', 'uint'),
                         elem('FrameRate',   '2383e3', 'float'),
                         elem('ColourSpace', '2eb524', 'binary'),
+                        elem('StereoMode',    '53b8', 'uint'),
                     }),
                     elem('Audio', 'e1', {
                         elem('SamplingFrequency',         'b5', 'float'),

TOOLS/lua/observe-all.lua

@@ -0,0 +1,16 @@
+-- Test script for property change notification mechanism.
+
+function format_property_val(v)
+    if type(v) == "table" then
+        return mp.format_table(v) -- undocumented function; might be removed
+    else
+        return tostring(v)
+    end
+end
+
+for i,name in ipairs(mp.get_property_native("property-list")) do
+    mp.observe_property(name, "native", function(name, val)
+        print("property '" .. name .. "' changed to '" ..
+              format_property_val(val) .. "'")
+    end)
+end

TOOLS/lua/status-line.lua

@@ -1,4 +1,4 @@
--- Rebuild the status line as a lua script
+-- Rebuild the terminal status line as a lua script
 -- Be aware that this will require more cpu power!
 
 -- Add a string to the status line

TOOLS/mpv_identify.sh

@@ -17,157 +17,134 @@
 # When multiple files were specified, their info will be put into FOO_* for the
 # first file, FOO_1_* for the second file, FOO_2_* for the third file, etc.
 
-case "$0" in
-    mpv_identify.sh|*/mpv_identify.sh)
-        # we are NOT being sourced
-        case "$1" in
-            '')
+__midentify__main() {
+
+    case "$0" in
+        mpv_identify.sh|*/mpv_identify.sh)
+            # we are NOT being sourced
+            [ -n "$1" ] && set -- '' "$@"
+            ;;
+    esac
+
+    if [ "$#" -lt 2 ]; then
+        cat >&2 <<EOF
+Usage 1 (for humans only): $0 filename.mkv
+will print all property values.
+Note that this output really shouldn't be parsed, as the
+format is subject to change.
+
+Usage 2 (for use by scripts): see top of this file
+
+NOTE: for mkv with ordered chapters, this may
+not always identify the specified file, but the
+file providing the first chapter. Specify
+--no-ordered-chapters to prevent this.
+EOF
+        return 2
+    fi
+
+    local LF="
+"
+
+    local nextprefix="$1"
+    shift
+
+    if [ -n "$nextprefix" ]; then
+        # in case of error, we always want this unset
+        unset "${nextprefix}path"
+    fi
+
+    local allprops="
+        filename
+        path
+        stream-start
+        stream-end
+        stream-length
+
+        demuxer
+
+        length
+        chapters
+        editions
+        titles
+
+        audio
+        audio-bitrate
+        audio-codec
+        audio-format
+        channels
+        samplerate
+
+        video
+        angle
+        video-bitrate
+        video-codec
+        video-format
+        video-aspect
+        fps
+        width
+        height
+        dwidth
+        dheight
+
+        sub
+    "
+    # TODO add metadata support once mpv can do it
+
+    local propstr="X-MIDENTIFY-START:$LF"
+    local key
+    for key in $allprops; do
+        propstr="${propstr}X-MIDENTIFY: $key \${=$key}$LF"
+        key="$(printf '%s\n' "$key" | tr - _)"
+        unset "$nextprefix$key"
+    done
+
+    local fileindex=0
+    local prefix=
+    local line
+    while IFS= read -r line; do
+        case "$line" in
+            X-MIDENTIFY-START:)
+                if [ -n "$nextprefix" ]; then
+                    prefix="$nextprefix"
+                    if [ "$fileindex" -gt 0 ]; then
+                        nextprefix="${prefix%${fileindex}_}"
+                    fi
+                    fileindex="$((fileindex+1))"
+                    nextprefix="${nextprefix}${fileindex}_"
+                    for key in $allprops; do
+                        key="$(printf '%s\n' "$key" | tr - _)"
+                        unset "$nextprefix$key"
+                    done
+                else
+                    if [ "$fileindex" -gt 0 ]; then
+                        printf '\n'
+                    fi
+                    fileindex="$((fileindex+1))"
+                fi
                 ;;
-            *)
-                set -- '' "$@"
+            X-MIDENTIFY:\ *)
+                local key="${line#X-MIDENTIFY: }"
+                local value="${key#* }"
+                key="${key%% *}"
+                key="$(printf '%s\n' "$key" | tr - _)"
+                if [ -n "$nextprefix" ]; then
+                    if [ -z "$prefix" ]; then
+                        echo >&2 "Got X-MIDENTIFY: without X-MIDENTIFY-START:"
+                    elif [ -n "$value" ]; then
+                        eval "$prefix$key"='"$value"'
+                    fi
+                else
+                    if [ -n "$value" ]; then
+                        printf '%s=%s\n' "$key" "$value"
+                    fi
+                fi
                 ;;
         esac
-        ;;
-esac
+    done <<EOF
+$(${MPV:-mpv} --term-playing-msg="$propstr" --vo=null --ao=null \
+              --frames=1 --quiet --no-cache --no-config -- "$@")
+EOF
+}
 
-if [ $# -lt 2 ]; then
-    echo >&2 "Usage 1 (for humans only): $0 filename.mkv"
-    echo >&2 "will print all property values."
-    echo >&2 "Note that this output really shouldn't be parsed, as the"
-    echo >&2 "format is subject to change."
-    echo >&2
-    echo >&2 "Usage 2 (for use by scripts): see top of this file"
-    echo >&2
-    echo >&2 "NOTE: for mkv with ordered chapters, this may"
-    echo >&2 "not always identify the specified file, but the"
-    echo >&2 "file providing the first chapter. Specify"
-    echo >&2 "--no-ordered-chapters to prevent this."
-    exit 1
-fi
-
-if [ -z "$MPV" ]; then
-    MPV="mpv"
-fi
-
-__midentify__LF="
-"
-
-__midentify__nextprefix=$1
-shift
-
-if [ -n "$__midentify__nextprefix" ]; then
-    # in case of error, we always want this unset
-    eval unset $__midentify__nextprefix'path'
-fi
-
-__midentify__allprops="
-    filename
-    path
-    stream-start
-    stream-end
-    stream-length
-
-    demuxer
-
-    length
-    chapters
-    editions
-    titles
-
-    audio
-    audio-bitrate
-    audio-codec
-    audio-format
-    channels
-    samplerate
-
-    video
-    angle
-    video-bitrate
-    video-codec
-    video-format
-    video-aspect
-    fps
-    width
-    height
-    dwidth
-    dheight
-
-    sub
-"
-# TODO add metadata support once mpv can do it
-
-__midentify__propstr="X-MIDENTIFY-START:$__midentify__LF"
-for __midentify__key in $__midentify__allprops; do
-    __midentify__propstr=$__midentify__propstr"X-MIDENTIFY: $__midentify__key \${=$__midentify__key}$__midentify__LF"
-    __midentify__key=`echo "$__midentify__key" | tr - _`
-    eval unset $__midentify__nextprefix$__midentify__key
-done
-
-__midentify__output=`$MPV --term-playing-msg="$__midentify__propstr" --vo=null --ao=null --frames=1 --quiet --no-cache --no-config "$@"`
-__midentify__fileindex=0
-__midentify__prefix=
-while :; do
-    case "$__midentify__output" in
-        '')
-            break
-            ;;
-        *$__midentify__LF*)
-            __midentify__line=${__midentify__output%%$__midentify__LF*}
-            __midentify__output=${__midentify__output#*$__midentify__LF}
-            ;;
-        *)
-            __midentify__line=$__midentify__output
-            __midentify__output=
-            ;;
-    esac
-    case "$__midentify__line" in
-        X-MIDENTIFY-START:)
-            if [ -n "$__midentify__nextprefix" ]; then
-                __midentify__prefix=$__midentify__nextprefix
-                if [ $__midentify__fileindex -gt 0 ]; then
-                    __midentify__nextprefix=${__midentify__prefix%$__midentify__fileindex\_}
-                fi
-                __midentify__fileindex=$(($__midentify__fileindex+1))
-                __midentify__nextprefix=$__midentify__nextprefix$__midentify__fileindex\_
-                for __midentify__key in $__midentify__allprops; do
-                    __midentify__key=`echo "$__midentify__key" | tr - _`
-                    eval unset $__midentify__nextprefix$__midentify__key
-                done
-            else
-                if [ $__midentify__fileindex -gt 0 ]; then
-                    echo
-                fi
-                __midentify__fileindex=$(($__midentify__fileindex+1))
-            fi
-            ;;
-        X-MIDENTIFY:\ *)
-            __midentify__key=${__midentify__line#X-MIDENTIFY:\ }
-            __midentify__value=${__midentify__key#* }
-            __midentify__key=${__midentify__key%% *}
-            __midentify__key=`echo "$__midentify__key" | tr - _`
-            if [ -n "$__midentify__nextprefix" ]; then
-                if [ -z "$__midentify__prefix" ]; then
-                    echo >&2 "Got X-MIDENTIFY: without X-MIDENTIFY-START:"
-                elif [ -n "$__midentify__value" ]; then
-                    eval $__midentify__prefix$__midentify__key=\$__midentify__value
-                fi
-            else
-                if [ -n "$__midentify__value" ]; then
-                    echo "$__midentify__key=$__midentify__value"
-                fi
-            fi
-            ;;
-    esac
-done
-
-unset __midentify__fileindex
-unset __midentify__allprops
-unset __midentify__key
-unset __midentify__LF
-unset __midentify__line
-unset __midentify__output
-unset __midentify__nextprefix
-unset __midentify__prefix
-unset __midentify__propstr
-unset __midentify__value
+__midentify__main "$@"

TOOLS/osxbundle/mpv.app/Contents/Info.plist

@@ -624,6 +624,90 @@
           </array>
         </dict>
       </dict>
+      <dict>
+        <key>UTTypeConformsTo</key>
+        <array>
+          <string>public.movie</string>
+        </array>
+        <key>UTTypeDescription</key>
+        <string>AVC raw stream</string>
+        <key>UTTypeIconFile</key>
+        <string>document.icns</string>
+        <key>UTTypeIdentifier</key>
+        <string>io.mpv.h264</string>
+        <key>UTTypeReferenceURL</key>
+        <string>http://www.itu.int/rec/T-REC-H.264</string>
+        <key>UTTypeTagSpecification</key>
+        <dict>
+          <key>public.filename-extension</key>
+          <array>
+            <string>264</string>
+          </array>
+        </dict>
+      </dict>
+      <dict>
+        <key>UTTypeConformsTo</key>
+        <array>
+          <string>public.movie</string>
+        </array>
+        <key>UTTypeDescription</key>
+        <string>HEVC raw stream</string>
+        <key>UTTypeIconFile</key>
+        <string>document.icns</string>
+        <key>UTTypeIdentifier</key>
+        <string>io.mpv.hevc</string>
+        <key>UTTypeReferenceURL</key>
+        <string>http://hevc.info</string>
+        <key>UTTypeTagSpecification</key>
+        <dict>
+          <key>public.filename-extension</key>
+          <array>
+            <string>hevc</string>
+          </array>
+        </dict>
+      </dict>
+      <dict>
+        <key>UTTypeConformsTo</key>
+        <array>
+          <string>public.movie</string>
+        </array>
+        <key>UTTypeDescription</key>
+        <string>YUV stream</string>
+        <key>UTTypeIconFile</key>
+        <string>document.icns</string>
+        <key>UTTypeIdentifier</key>
+        <string>io.mpv.yuv</string>
+        <key>UTTypeReferenceURL</key>
+        <string>http://en.wikipedia.org/wiki/YUV</string>
+        <key>UTTypeTagSpecification</key>
+        <dict>
+          <key>public.filename-extension</key>
+          <array>
+            <string>yuv</string>
+          </array>
+        </dict>
+      </dict>
+      <dict>
+        <key>UTTypeConformsTo</key>
+        <array>
+          <string>public.movie</string>
+        </array>
+        <key>UTTypeDescription</key>
+        <string>YUV4MPEG2 stream</string>
+        <key>UTTypeIconFile</key>
+        <string>document.icns</string>
+        <key>UTTypeIdentifier</key>
+        <string>io.mpv.y4m</string>
+        <key>UTTypeReferenceURL</key>
+        <string>http://wiki.multimedia.cx/index.php?title=YUV4MPEG2</string>
+        <key>UTTypeTagSpecification</key>
+        <dict>
+          <key>public.filename-extension</key>
+          <array>
+            <string>y4m</string>
+          </array>
+        </dict>
+      </dict>
       <dict>
         <key>UTTypeConformsTo</key>
         <array>

TOOLS/stats-conv.py

@@ -18,10 +18,11 @@ e.g.:
 
 Currently, the following event types are supported:
 
-    'start' <name>          start of the named event
-    'end' <name>            end of the named event
-    'value' <float> <name>  a normal value (as opposed to event)
-    <event>                 singular event
+    'start' <name>              start of the named event
+    'end' <name>                end of the named event
+    'value' <float> <name>      a normal value (as opposed to event)
+    'event-timed' <ts> <name>   singular event at the given timestamp
+    <name>                      singular event
 
 """
 
@@ -29,19 +30,32 @@ class G:
     events = {}
     sevents = []  # events, deterministically sorted
     start = None
+    # http://matplotlib.org/api/markers_api.html#module-matplotlib.markers
+    markers = ["o", "8", "s", "p", "*", "h", "+", "x", "D"]
+
+def find_marker():
+    if len(G.markers) == 0:
+        return "o"
+    m = G.markers[0]
+    G.markers = G.markers[1:]
+    return m
 
 class Event:
     pass
 
-def get_event(event):
+def get_event(event, evtype):
     if event not in G.events:
         e = Event()
         G.events[event] = e
         e.name = event
         e.vals = []
-        e.type = "unknown"
+        e.type = evtype
+        e.marker = "o"
+        e.numid = len(G.events)
         G.sevents = list(G.events.values())
         G.sevents.sort(key=lambda x: x.name)
+        if e.type == "event-signal":
+            e.marker = find_marker()
     return G.events[event]
 
 for line in [line.split("#")[0].strip() for line in open(filename, "r")]:
@@ -54,30 +68,29 @@ for line in [line.split("#")[0].strip() for line in open(filename, "r")]:
         G.start = ts
     ts = ts - G.start
     if event.startswith("start "):
-        e = get_event(event[6:])
-        e.type = "event"
+        e = get_event(event[6:], "event")
         e.vals.append((ts, 0))
         e.vals.append((ts, 1))
     elif event.startswith("end "):
-        e = get_event(event[4:])
-        e.type = "event"
+        e = get_event(event[4:], "event")
         e.vals.append((ts, 1))
         e.vals.append((ts, 0))
     elif event.startswith("value "):
         _, val, name = event.split(" ", 2)
         val = float(val)
-        e = get_event(name)
-        e.type = "value"
+        e = get_event(name, "value")
         e.vals.append((ts, val))
+    elif event.startswith("event-timed "):
+        _, val, name = event.split(" ", 2)
+        val = int(val) / 1000 - G.start
+        e = get_event(name, "event-signal")
+        e.vals.append((val, 1))
     else:
-        e = get_event(event)
-        e.type = "event-signal"
+        e = get_event(event, "event-signal")
         e.vals.append((ts, 1))
 
-ao_events = ["ao fill", "audio wait"]
 for e in G.sevents:
-    if e.name in ao_events:
-        e.vals = [(x, y * 0.5) for (x, y) in e.vals]
+    e.vals = [(x, y * e.numid / len(G.events)) for (x, y) in e.vals]
 
 plot.hold(True)
 mainpl = plot.subplot(2, 1, 1)
@@ -89,7 +102,7 @@ for e in G.sevents:
         plot.subplot(2, 1, 1)
     pl, = plot.plot([x for x,y in e.vals], [y for x,y in e.vals], label=e.name)
     if e.type == "event-signal":
-        plot.setp(pl, marker = "o", linestyle = "None")
+        plot.setp(pl, marker = e.marker, linestyle = "None")
     legend.append(pl)
 plot.subplot(2, 1, 1)
 plot.legend(legend, [pl.get_label() for pl in legend])

TOOLS/umpv

@@ -23,10 +23,11 @@ Note that you can control the mpv instance by writing to the command fifo:
 
     echo "cycle fullscreen" > /tmp/umpv-fifo-*
 
-Note: you can supply custom options with the UMPV_OPTIONS environment variable.
-      The environment variable will be split on whitespace, and each item is
-      passed as option to mpv _if_ the script starts mpv. If mpv is not started
-      by the script (i.e. mpv is already running), the options will be ignored.
+Note: you can supply custom mpv path and options with the MPV environment
+      variable. The environment variable will be split on whitespace, and the
+      first item is used as path to mpv binary and the rest is passed as options
+      _if_ the script starts mpv. If mpv is not started by the script (i.e. mpv
+      is already running), this will be ignored.
 
 Warning:
 
@@ -45,21 +46,17 @@ import fcntl
 import stat
 import string
 
-if len(sys.argv) < 1:
-    sys.exit(1)
 files = sys.argv[1:]
 
-# this is about the same method mpv uses to decide this
+# this is the same method mpv uses to decide this
 def is_url(filename):
-    parts = filename.split(":", 1)
+    parts = filename.split("://", 1)
     if len(parts) < 2:
         return False
     # protocol prefix has no special characters => it's an URL
+    allowed_symbols = string.ascii_letters + string.digits + '_'
     prefix = parts[0]
-    for c in prefix:
-        if string.ascii_letters.find(c) < 0:
-            return False
-    return True
+    return all(map(lambda c: c in allowed_symbols, prefix))
 
 # make them absolute; also makes them safe against interpretation as options
 def make_abs(filename):
@@ -68,9 +65,6 @@ def make_abs(filename):
     return filename
 files = [make_abs(f) for f in files]
 
-opts_env = os.getenv("UMPV_OPTIONS")
-opts_env = opts_env.split() if opts_env else []
-
 FIFO = "/tmp/umpv-fifo-" + os.getenv("USER")
 
 fifo_fd = -1
@@ -89,7 +83,8 @@ if fifo_fd >= 0:
     if (((st.st_mode & (stat.S_IRWXG | stat.S_IRWXO)) != 0) or
         (not stat.S_ISFIFO(st.st_mode)) or
         (st.st_uid != os.getuid())):
-        print("error: command FIFO is not a FIFO or has bogus permissions")
+        sys.stderr.write("error: command FIFO is not a FIFO or has bogus "
+                         "permissions\n")
         sys.exit(1)
 
 if fifo_fd >= 0:
@@ -112,11 +107,11 @@ else:
             pass
         else:
             raise e
-    os.mkfifo(FIFO, int("0600", 8))
+    os.mkfifo(FIFO, 0o600)
 
-    opts = ["mpv", "--no-terminal", "--force-window", "--input-file=" + FIFO]
-    opts.extend(opts_env)
-    opts.append("--")
+    opts = (os.getenv("MPV") or "mpv").split()
+    opts.extend(["--no-terminal", "--force-window", "--input-file=" + FIFO,
+                 "--"])
     opts.extend(files)
 
     subprocess.check_call(opts)

TOOLS/vf_dlopen/framestep.c

@@ -90,8 +90,7 @@ int vf_dlopen_getcontext(struct vf_dlopen_context *ctx, int argc, const char **a
     if (argc != 1 && argc != 2)
         return -1;
 
-    framestep_data_t *framestep = malloc(sizeof(framestep_data_t));
-    memset(framestep, 0, sizeof(*framestep));
+    framestep_data_t *framestep = calloc(1,sizeof(framestep_data_t));
 
     framestep->step = atoi(argv[0]);
     framestep->pos = (argc >= 2) ? atoi(argv[1]) : 0;

TOOLS/vf_dlopen/ildetect.c

@@ -267,8 +267,7 @@ int vf_dlopen_getcontext(struct vf_dlopen_context *ctx, int argc, const char **a
     (void) argc;
     (void) argv;
 
-    ildetect_data_t *il = malloc(sizeof(ildetect_data_t));
-    memset(il, 0, sizeof(*il));
+    ildetect_data_t *il = calloc(1,sizeof(ildetect_data_t));
 
 #define A(i,d) ((argc>(i) && *argv[i]) ? atof(argv[i]) : (d))
     il->method = A(0, 0);

TOOLS/vf_dlopen/rectangle.c

@@ -348,8 +348,7 @@ int vf_dlopen_getcontext(struct vf_dlopen_context *ctx, int argc, const char **a
         { "yuv420p", "yuv420p" },
         { NULL, NULL }
     };
-    privdata *priv = malloc(sizeof(privdata));
-    memset(priv, 0, sizeof(*priv));
+    privdata *priv = calloc(1,sizeof(privdata));
     priv->step = 8;
     if(argc >= 1)
         priv->w = atoi(argv[0]);

TOOLS/vf_dlopen/telecine.c

@@ -237,8 +237,7 @@ int vf_dlopen_getcontext(struct vf_dlopen_context *ctx, int argc, const char **a
     if (!a0[0] || a0[1] || !a1[0] || argc > 2)
         return -1;
 
-    tc_data_t *tc = malloc(sizeof(tc_data_t));
-    memset(tc, 0, sizeof(*tc));
+    tc_data_t *tc = calloc(1,sizeof(tc_data_t));
 
     if (a0[0] == 't')
         tc->firstfield = 0;

TOOLS/vf_dlopen/tile.c

@@ -150,8 +150,7 @@ int vf_dlopen_getcontext(struct vf_dlopen_context *ctx, int argc, const char **a
     if (argc != 2)
         return -1;
 
-    tile_data_t *tile = malloc(sizeof(tile_data_t));
-    memset(tile, 0, sizeof(*tile));
+    tile_data_t *tile = calloc(1,sizeof(tile_data_t));
 
     tile->cols = atoi(argv[0]);
     tile->rows = atoi(argv[1]);

TOOLS/youtube-dl_mpv.sh

@@ -10,14 +10,13 @@ video_url="$(youtube-dl \
     --user-agent="$user_agent" \
     --cookies="$cookies_file" \
     --get-url \
-    "$1")"
+    "$@")"
 
-shift
-
-mpv \
+set -f
+${MPV:-mpv} \
     --cookies \
     --cookies-file="$cookies_file" \
     --user-agent="$user_agent" \
-    "$@" -- $video_url
+    -- $video_url
 
 rm -rf "$cookies_dir"

TOOLS/zsh.pl

@@ -7,7 +7,7 @@ use warnings;
 
 my $mpv = $ARGV[0] || 'mpv';
 
-my @opts = parse_opts("$mpv --list-options", '^ (\-\-[^\s\*]*)\*?\s*(.*)');
+my @opts = parse_opts("$mpv --list-options", '^ (\-\-[^\s\*]*)\*?\s*(.*)', 1);
 
 my @ao = parse_opts("$mpv --ao=help", '^  ([^\s\:]*)\s*: (.*)');
 my @vo = parse_opts("$mpv --vo=help", '^  ([^\s\:]*)\s*: (.*)');
@@ -15,9 +15,11 @@ my @vo = parse_opts("$mpv --vo=help", '^  ([^\s\:]*)\s*: (.*)');
 my @af = parse_opts("$mpv --af=help", '^  ([^\s\:]*)\s*: (.*)');
 my @vf = parse_opts("$mpv --vf=help", '^  ([^\s\:]*)\s*: (.*)');
 
-my ($opts_str, $ao_str, $vo_str, $af_str, $vf_str);
+my @protos = parse_opts("$mpv --list-protocols", '^ ([^\s]*)');
 
-$opts_str .= qq{  '$_': \\\n} foreach (@opts);
+my ($opts_str, $ao_str, $vo_str, $af_str, $vf_str, $protos_str);
+
+$opts_str .= qq{  '$_' \\\n} foreach (@opts);
 chomp $opts_str;
 
 $ao_str .= qq{      '$_' \\\n} foreach (@ao);
@@ -32,14 +34,52 @@ chomp $af_str;
 $vf_str .= qq{      '$_' \\\n} foreach (@vf);
 chomp $vf_str;
 
+$protos_str .= qq{$_ } foreach (@protos);
+chomp $protos_str;
+
+my $profile_comp = <<'EOS';
+      local -a profiles
+      local current
+      for current in "${(@f)$($words[1] --profile=help)}"; do
+        current=${current//\*/\\\*}
+        current=${current//\:/\\\:}
+        current=${current//\[/\\\[}
+        current=${current//\]/\\\]}
+        if [[ $current =~ $'\t'([^$'\t']*)$'\t'(.*) ]]; then
+          if [[ -n $match[2] ]]; then
+            current="$match[1][$match[2]]"
+          else
+            current="$match[1]"
+          fi
+          profiles=($profiles $current)
+        fi
+      done
+      if [[ $state == profile ]]; then
+        # For --show-profile, only one allowed
+        if (( ${#profiles} > 0 )); then
+          _values 'profile' $profiles && rc=0
+        fi
+      else
+        # For --profile, multiple allowed
+        profiles=($profiles 'help[list profiles]')
+        _values -s , 'profile(s)' $profiles && rc=0
+      fi
+EOS
+chomp $profile_comp;
+
 my $tmpl = <<"EOS";
 #compdef mpv
 
 # mpv zsh completion
 
-_x_arguments -C -s \\
+local curcontext="\$curcontext" state state_descr line
+typeset -A opt_args
+
+local rc=1
+
+_arguments -C -S \\
 $opts_str
-  '*:files:->mfiles'
+  '*:files:->mfiles' && rc=0
 
 case \$state in
   ao)
@@ -48,7 +88,7 @@ case \$state in
 $ao_str
     )
 
-    _describe -t values 'audio outputs' values
+    _describe -t values 'audio outputs' values && rc=0
   ;;
 
   vo)
@@ -57,7 +97,7 @@ $ao_str
 $vo_str
     )
 
-    _describe -t values 'video outputs' values
+    _describe -t values 'video outputs' values && rc=0
   ;;
 
   af)
@@ -66,7 +106,7 @@ $vo_str
 $af_str
     )
 
-    _describe -t values 'audio filters' values
+    _describe -t values 'audio filters' values && rc=0
   ;;
 
   vf)
@@ -75,55 +115,88 @@ $af_str
 $vf_str
     )
 
-    _describe -t values 'video filters' values
+    _describe -t values 'video filters' values && rc=0
+  ;;
+
+  profile|profiles)
+$profile_comp
+  ;;
+
+  files)
+    compset -P '*,'
+    compset -S ',*'
+    _files -r ',/ \\t\\n\\-' && rc=0
   ;;
 
   mfiles)
+    local expl
     _tags files urls
     while _tags; do
       _requested files expl 'media file' _files -g \\
-         "*.(#i)(asf|asx|avi|flac|flv|m1v|m2p|m2v|m4v|mjpg|mka|mkv|mov|mp3|mp4|mpe|mpeg|mpg|ogg|ogm|ogv|qt|rm|ts|vob|wav|webm|wma|wmv)(-.)" && ret=0
+         "*.(#i)(asf|asx|avi|flac|flv|m1v|m2p|m2v|m4v|mjpg|mka|mkv|mov|mp3|mp4|mpe|mpeg|mpg|ogg|ogm|ogv|qt|rm|ts|vob|wav|webm|wma|wmv)(-.)" && rc=0
       if _requested urls; then
         while _next_label urls expl URL; do
-          _urls "\$expl[@]" && ret=0
-          compadd -S '' "\$expl[@]" \{dvd,vcd,cdda,cddb,tv\}:// && ret=0
+          _urls "\$expl[@]" && rc=0
+          compadd -S '' "\$expl[@]" $protos_str && rc=0
         done
       fi
-      (( ret )) || return 0
+      (( rc )) || return 0
     done
   ;;
 esac
+
+return rc
 EOS
 
 print $tmpl;
 
 sub parse_opts {
-	my ($cmd, $regex) = @_;
+    my ($cmd, $regex, $parsing_main_options) = @_;
 
-	my @list;
-	my @lines = split /\n/, `$cmd`;
+    my @list;
+    my @lines = split /\n/, `$cmd`;
 
-	foreach my $line (@lines) {
-		if ($line !~ /^$regex/) {
-			next;
-		}
+    foreach my $line (@lines) {
+        if ($line !~ /^$regex/) {
+            next;
+        }
 
-		my $entry = "$1:";
+        my $entry = $1;
 
-		if (defined $2) {
-			my $desc = $2;
-			$desc =~ s/\:/\\:/;
+        if ($parsing_main_options) {
+            $entry .= '=-';
+        }
 
-			$entry .= $desc;
-		}
+        if (defined $2) {
+            my $desc = $2;
+            $desc =~ s/\:/\\:/g;
 
-		$entry .= ':->ao' if ($1 eq '--ao');
-		$entry .= ':->vo' if ($1 eq '--vo');
-		$entry .= ':->af' if ($1 eq '--af');
-		$entry .= ':->vf' if ($1 eq '--vf');
+            $entry .= ':' . $desc;
+        }
 
-		push @list, $entry if ($line =~ /^$regex/)
-	}
+        if ($parsing_main_options) {
+            $entry .= ':';
 
-	return @list;
+            $entry .= '->ao' if ($1 eq '--ao');
+            $entry .= '->vo' if ($1 eq '--vo');
+            $entry .= '->af' if ($1 eq '--af');
+            $entry .= '->vf' if ($1 eq '--vf');
+            $entry .= '->profiles' if ($1 eq '--profile');
+            $entry .= '->profile' if ($1 eq '--show-profile');
+            $entry .= '->files' if ($line =~ /\[file\]/);
+        }
+
+        push @list, $entry
+    }
+
+    if ($parsing_main_options) {
+        @list = sort {
+            $a =~ /(.*?)\:/; my $ma = $1;
+            $b =~ /(.*?)\:/; my $mb = $1;
+
+            length($mb) <=> length($ma)
+        } @list;
+    }
+
+    return @list;
 }

VERSION

@@ -0,0 +1 @@
+0.6.2

audio/chmap.h

@@ -20,7 +20,7 @@
 
 #include <inttypes.h>
 #include <stdbool.h>
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 
 #define MP_NUM_CHANNELS 8
 

audio/decode/ad.h

@@ -36,7 +36,7 @@ struct ad_functions {
     int (*init)(struct dec_audio *da, const char *decoder);
     void (*uninit)(struct dec_audio *da);
     int (*control)(struct dec_audio *da, int cmd, void *arg);
-    int (*decode_audio)(struct dec_audio *da, struct mp_audio *buffer, int maxlen);
+    int (*decode_packet)(struct dec_audio *da);
 };
 
 enum ad_ctrl {

audio/decode/ad_lavc.c

@@ -33,13 +33,10 @@
 #include "common/codecs.h"
 #include "common/msg.h"
 #include "options/options.h"
-#include "common/av_opts.h"
 
 #include "ad.h"
 #include "audio/fmt-conversion.h"
 
-#include "compat/libav.h"
-
 struct priv {
     AVCodecContext *avctx;
     AVFrame *avframe;
@@ -49,14 +46,13 @@ struct priv {
 };
 
 static void uninit(struct dec_audio *da);
-static int decode_new_packet(struct dec_audio *da);
 
 #define OPT_BASE_STRUCT struct ad_lavc_params
 struct ad_lavc_params {
     float ac3drc;
     int downmix;
     int threads;
-    char *avopt;
+    char **avopts;
 };
 
 const struct m_sub_options ad_lavc_conf = {
@@ -64,7 +60,7 @@ const struct m_sub_options ad_lavc_conf = {
         OPT_FLOATRANGE("ac3drc", ac3drc, 0, 0, 2),
         OPT_FLAG("downmix", downmix, 0),
         OPT_INTRANGE("threads", threads, 0, 1, 16),
-        OPT_STRING("o", avopt, 0),
+        OPT_KEYVALUELIST("o", avopts, 0),
         {0}
     },
     .size = sizeof(struct ad_lavc_params),
@@ -75,79 +71,10 @@ const struct m_sub_options ad_lavc_conf = {
     },
 };
 
-struct pcm_map
-{
-    int tag;
-    const char *codecs[6]; // {any, 1byte, 2bytes, 3bytes, 4bytes, 8bytes}
-};
-
-// NOTE: these are needed to make rawaudio with demux_mkv work.
-static const struct pcm_map tag_map[] = {
-    // Microsoft PCM
-    {0x0,           {NULL, "pcm_u8", "pcm_s16le", "pcm_s24le", "pcm_s32le"}},
-    {0x1,           {NULL, "pcm_u8", "pcm_s16le", "pcm_s24le", "pcm_s32le"}},
-    // MS PCM, Extended
-    {0xfffe,        {NULL, "pcm_u8", "pcm_s16le", "pcm_s24le", "pcm_s32le"}},
-    // IEEE float
-    {0x3,           {"pcm_f32le", [5] = "pcm_f64le"}},
-    // 'raw '
-    {0x20776172,    {"pcm_s16be", [1] = "pcm_u8"}},
-    // 'twos', used by demux_mkv.c internally
-    {MKTAG('t', 'w', 'o', 's'),
-                    {NULL, "pcm_s8", "pcm_s16be", "pcm_s24be", "pcm_s32be"}},
-    {-1},
-};
-
-// For demux_rawaudio.c; needed because ffmpeg doesn't have these sample
-// formats natively.
-static const struct pcm_map af_map[] = {
-    {AF_FORMAT_U8,              {"pcm_u8"}},
-    {AF_FORMAT_S8,              {"pcm_u8"}},
-    {AF_FORMAT_U16_LE,          {"pcm_u16le"}},
-    {AF_FORMAT_U16_BE,          {"pcm_u16be"}},
-    {AF_FORMAT_S16_LE,          {"pcm_s16le"}},
-    {AF_FORMAT_S16_BE,          {"pcm_s16be"}},
-    {AF_FORMAT_U24_LE,          {"pcm_u24le"}},
-    {AF_FORMAT_U24_BE,          {"pcm_u24be"}},
-    {AF_FORMAT_S24_LE,          {"pcm_s24le"}},
-    {AF_FORMAT_S24_BE,          {"pcm_s24be"}},
-    {AF_FORMAT_U32_LE,          {"pcm_u32le"}},
-    {AF_FORMAT_U32_BE,          {"pcm_u32be"}},
-    {AF_FORMAT_S32_LE,          {"pcm_s32le"}},
-    {AF_FORMAT_S32_BE,          {"pcm_s32be"}},
-    {AF_FORMAT_FLOAT_LE,        {"pcm_f32le"}},
-    {AF_FORMAT_FLOAT_BE,        {"pcm_f32be"}},
-    {AF_FORMAT_DOUBLE_LE,       {"pcm_f64le"}},
-    {AF_FORMAT_DOUBLE_BE,       {"pcm_f64be"}},
-    {-1},
-};
-
-static const char *find_pcm_decoder(const struct pcm_map *map, int format,
-                                    int bits_per_sample)
-{
-    int bytes = (bits_per_sample + 7) / 8;
-    if (bytes == 8)
-        bytes = 5; // 64 bit entry
-    for (int n = 0; map[n].tag != -1; n++) {
-        const struct pcm_map *entry = &map[n];
-        if (entry->tag == format) {
-            const char *dec = NULL;
-            if (bytes >= 1 && bytes <= 5)
-                dec = entry->codecs[bytes];
-            if (!dec)
-                dec = entry->codecs[0];
-            if (dec)
-                return dec;
-        }
-    }
-    return NULL;
-}
-
-static int setup_format(struct dec_audio *da)
+static void set_data_from_avframe(struct dec_audio *da)
 {
     struct priv *priv = da->priv;
     AVCodecContext *lavc_context = priv->avctx;
-    struct sh_audio *sh_audio = da->header->audio;
 
     // Note: invalid parameters are rejected by dec_audio.c
 
@@ -157,12 +84,6 @@ static int setup_format(struct dec_audio *da)
         MP_FATAL(da, "unsupported lavc format %s", av_get_sample_fmt_name(fmt));
 
     da->decoded.rate = lavc_context->sample_rate;
-    if (!da->decoded.rate && sh_audio->wf) {
-        // If not set, try container samplerate.
-        // (Maybe this can't happen, and it's an artifact from the past.)
-        da->decoded.rate = sh_audio->wf->nSamplesPerSec;
-        MP_WARN(da, "using container rate.\n");
-    }
 
     struct mp_chmap lavc_chmap;
     mp_chmap_from_lavc(&lavc_chmap, lavc_context->channel_layout);
@@ -170,24 +91,15 @@ static int setup_format(struct dec_audio *da)
     if (lavc_chmap.num != lavc_context->channels)
         mp_chmap_from_channels(&lavc_chmap, lavc_context->channels);
     if (priv->force_channel_map) {
+        struct sh_audio *sh_audio = da->header->audio;
         if (lavc_chmap.num == sh_audio->channels.num)
             lavc_chmap = sh_audio->channels;
     }
     mp_audio_set_channels(&da->decoded, &lavc_chmap);
 
-    return 0;
-}
-
-static void set_from_wf(AVCodecContext *avctx, MP_WAVEFORMATEX *wf)
-{
-    avctx->channels = wf->nChannels;
-    avctx->sample_rate = wf->nSamplesPerSec;
-    avctx->bit_rate = wf->nAvgBytesPerSec * 8;
-    avctx->block_align = wf->nBlockAlign;
-    avctx->bits_per_coded_sample = wf->wBitsPerSample;
-
-    if (wf->cbSize > 0)
-        mp_lavc_set_extradata(avctx, wf + 1, wf->cbSize);
+    da->decoded.samples = priv->avframe->nb_samples;
+    for (int n = 0; n < da->decoded.num_planes; n++)
+        da->decoded.planes[n] = priv->avframe->data[n];
 }
 
 static int init(struct dec_audio *da, const char *decoder)
@@ -202,13 +114,7 @@ static int init(struct dec_audio *da, const char *decoder)
     struct priv *ctx = talloc_zero(NULL, struct priv);
     da->priv = ctx;
 
-    if (sh_audio->wf && strcmp(decoder, "pcm") == 0) {
-        decoder = find_pcm_decoder(tag_map, sh->format,
-                                   sh_audio->wf->wBitsPerSample);
-    } else if (sh_audio->wf && strcmp(decoder, "mp-pcm") == 0) {
-        decoder = find_pcm_decoder(af_map, sh->format, 0);
-        ctx->force_channel_map = true;
-    }
+    ctx->force_channel_map = sh_audio->force_channels;
 
     lavc_codec = avcodec_find_decoder_by_name(decoder);
     if (!lavc_codec) {
@@ -233,23 +139,18 @@ static int init(struct dec_audio *da, const char *decoder)
     av_opt_set_double(lavc_context, "drc_scale", opts->ac3drc,
                       AV_OPT_SEARCH_CHILDREN);
 
-    if (opts->avopt) {
-        if (parse_avopts(lavc_context, opts->avopt) < 0) {
-            MP_ERR(da, "setting AVOptions '%s' failed.\n", opts->avopt);
-            uninit(da);
-            return 0;
-        }
-    }
+    mp_set_avopts(da->log, lavc_context, opts->avopts);
 
     lavc_context->codec_tag = sh->format;
     lavc_context->sample_rate = sh_audio->samplerate;
     lavc_context->bit_rate = sh_audio->bitrate;
-    lavc_context->channel_layout = mp_chmap_to_lavc(&sh_audio->channels);
+    lavc_context->block_align = sh_audio->block_align;
+    lavc_context->bits_per_coded_sample = sh_audio->bits_per_coded_sample;
+    lavc_context->channels = sh_audio->channels.num;
+    if (!mp_chmap_is_unknown(&sh_audio->channels))
+        lavc_context->channel_layout = mp_chmap_to_lavc(&sh_audio->channels);
 
-    if (sh_audio->wf)
-        set_from_wf(lavc_context, sh_audio->wf);
-
-    // demux_mkv, demux_mpg
+    // demux_mkv
     if (sh_audio->codecdata_len && sh_audio->codecdata &&
             !lavc_context->extradata) {
         mp_lavc_set_extradata(lavc_context, sh_audio->codecdata,
@@ -267,27 +168,9 @@ static int init(struct dec_audio *da, const char *decoder)
         uninit(da);
         return 0;
     }
-    MP_VERBOSE(da, "INFO: libavcodec \"%s\" init OK!\n",
-           lavc_codec->name);
-
-    // Decode at least 1 sample:  (to get header filled)
-    for (int tries = 1; ; tries++) {
-        int x = decode_new_packet(da);
-        if (x >= 0 && ctx->frame.samples > 0) {
-            MP_VERBOSE(da, "Initial decode succeeded after %d packets.\n", tries);
-            break;
-        }
-        if (tries >= 50) {
-            MP_ERR(da, "initial decode failed\n");
-            uninit(da);
-            return 0;
-        }
-    }
 
     if (lavc_context->bit_rate != 0)
         da->bitrate = lavc_context->bit_rate;
-    else if (sh_audio->wf && sh_audio->wf->nAvgBytesPerSec)
-        da->bitrate = sh_audio->wf->nAvgBytesPerSec * 8;
 
     return 1;
 }
@@ -314,7 +197,7 @@ static int control(struct dec_audio *da, int cmd, void *arg)
     switch (cmd) {
     case ADCTRL_RESET:
         avcodec_flush_buffers(ctx->avctx);
-        ctx->frame.samples = 0;
+        mp_audio_set_null_data(&da->decoded);
         talloc_free(ctx->packet);
         ctx->packet = NULL;
         return CONTROL_TRUE;
@@ -322,16 +205,18 @@ static int control(struct dec_audio *da, int cmd, void *arg)
     return CONTROL_UNKNOWN;
 }
 
-static int decode_new_packet(struct dec_audio *da)
+static int decode_packet(struct dec_audio *da)
 {
     struct priv *priv = da->priv;
     AVCodecContext *avctx = priv->avctx;
 
-    priv->frame.samples = 0;
+    mp_audio_set_null_data(&da->decoded);
 
     struct demux_packet *mpkt = priv->packet;
-    if (!mpkt)
-        mpkt = demux_read_packet(da->header);
+    if (!mpkt) {
+        if (demux_read_packet_async(da->header, &mpkt) == 0)
+            return AD_WAIT;
+    }
 
     priv->packet = talloc_steal(priv, mpkt);
 
@@ -364,22 +249,16 @@ static int decode_new_packet(struct dec_audio *da)
         }
         // LATM may need many packets to find mux info
         if (ret == AVERROR(EAGAIN))
-            return 0;
+            return AD_OK;
     }
     if (ret < 0) {
-        MP_VERBOSE(da, "lavc_audio: error\n");
-        return -1;
+        MP_ERR(da, "Error decoding audio.\n");
+        return AD_ERR;
     }
     if (!got_frame)
-        return mpkt ? 0 : -1; // -1: eof
+        return mpkt ? AD_OK : AD_EOF;
 
-    if (setup_format(da) < 0)
-        return -1;
-
-    priv->frame.samples = priv->avframe->nb_samples;
-    mp_audio_copy_config(&priv->frame, &da->decoded);
-    for (int n = 0; n < priv->frame.num_planes; n++)
-        priv->frame.planes[n] = priv->avframe->data[n];
+    set_data_from_avframe(da);
 
     double out_pts = mp_pts_from_av(priv->avframe->pkt_pts, NULL);
     if (out_pts != MP_NOPTS_VALUE) {
@@ -387,35 +266,13 @@ static int decode_new_packet(struct dec_audio *da)
         da->pts_offset = 0;
     }
 
-    MP_DBG(da, "Decoded %d -> %d samples\n", in_len,
-           priv->frame.samples);
-    return 0;
-}
-
-static int decode_audio(struct dec_audio *da, struct mp_audio *buffer, int maxlen)
-{
-    struct priv *priv = da->priv;
-
-    if (!priv->frame.samples) {
-        if (decode_new_packet(da) < 0)
-            return -1;
-    }
-
-    if (!mp_audio_config_equals(buffer, &priv->frame))
-        return 0;
-
-    buffer->samples = MPMIN(priv->frame.samples, maxlen);
-    mp_audio_copy(buffer, 0, &priv->frame, 0, buffer->samples);
-    mp_audio_skip_samples(&priv->frame, buffer->samples);
-    da->pts_offset += buffer->samples;
+    MP_DBG(da, "Decoded %d -> %d samples\n", in_len, da->decoded.samples);
     return 0;
 }
 
 static void add_decoders(struct mp_decoder_list *list)
 {
     mp_add_lavc_decoders(list, AVMEDIA_TYPE_AUDIO);
-    mp_add_decoder(list, "lavc", "pcm", "pcm", "Raw PCM");
-    mp_add_decoder(list, "lavc", "mp-pcm", "mp-pcm", "Raw PCM");
 }
 
 const struct ad_functions ad_lavc = {
@@ -424,5 +281,5 @@ const struct ad_functions ad_lavc = {
     .init = init,
     .uninit = uninit,
     .control = control,
-    .decode_audio = decode_audio,
+    .decode_packet = decode_packet,
 };

audio/decode/ad_mpg123.c

@@ -35,9 +35,7 @@
 
 struct ad_mpg123_context {
     mpg123_handle *handle;
-    bool new_format;
     int sample_size;
-    bool need_data;
     /* Running mean for bit rate, stream length estimation. */
     float mean_rate;
     unsigned int mean_count;
@@ -58,7 +56,7 @@ static void uninit(struct dec_audio *da)
 
 /* This initializes libmpg123 and prepares the handle, including funky
  * parameters. */
-static int preinit(struct dec_audio *da)
+static int init(struct dec_audio *da, const char *decoder)
 {
     int err;
     struct ad_mpg123_context *con;
@@ -111,15 +109,18 @@ static int preinit(struct dec_audio *da)
      * We need at least 1152 samples. dec_audio.c normally guarantees this. */
     mpg123_param(con->handle, MPG123_REMOVE_FLAGS, MPG123_AUTO_RESAMPLE, 0.);
 
+    err = mpg123_open_feed(con->handle);
+    if (err != MPG123_OK)
+        goto bad_end;
+
     return 1;
 
   bad_end:
-    if (!con->handle)
-        MP_ERR(da, "mpg123 preinit error: %s\n",
-               mpg123_plain_strerror(err));
-    else
-        MP_ERR(da, "mpg123 preinit error: %s\n",
-               mpg123_strerror(con->handle));
+    if (!con->handle) {
+        MP_ERR(da, "mpg123 preinit error: %s\n", mpg123_plain_strerror(err));
+    } else {
+        MP_ERR(da, "mpg123 preinit error: %s\n", mpg123_strerror(con->handle));
+    }
 
     uninit(da);
     return 0;
@@ -140,111 +141,6 @@ static int mpg123_format_to_af(int mpg123_encoding)
     return 0;
 }
 
-/* libmpg123 has a new format ready; query and store, return return value
-   of mpg123_getformat() */
-static int set_format(struct dec_audio *da)
-{
-    struct ad_mpg123_context *con = da->priv;
-    int ret;
-    long rate;
-    int channels;
-    int encoding;
-    ret = mpg123_getformat(con->handle, &rate, &channels, &encoding);
-    if (ret == MPG123_OK) {
-        mp_audio_set_num_channels(&da->decoded, channels);
-        da->decoded.rate = rate;
-        int af = mpg123_format_to_af(encoding);
-        if (!af) {
-            /* This means we got a funny custom build of libmpg123 that only supports an unknown format. */
-            MP_ERR(da, "Bad encoding from mpg123: %i.\n", encoding);
-            return MPG123_ERR;
-        }
-        mp_audio_set_format(&da->decoded, af);
-        con->sample_size = channels * af_fmt2bps(af);
-        con->new_format = 0;
-    }
-    return ret;
-}
-
-static int feed_new_packet(struct dec_audio *da)
-{
-    struct ad_mpg123_context *con = da->priv;
-    int ret;
-
-    struct demux_packet *pkt = demux_read_packet(da->header);
-    if (!pkt)
-        return -1; /* EOF. */
-
-    /* Next bytes from that presentation time. */
-    if (pkt->pts != MP_NOPTS_VALUE) {
-        da->pts        = pkt->pts;
-        da->pts_offset = 0;
-    }
-
-    /* Have to use mpg123_feed() to avoid decoding here. */
-    ret = mpg123_feed(con->handle, pkt->buffer, pkt->len);
-    talloc_free(pkt);
-
-    if (ret == MPG123_ERR)
-        return -1;
-
-    if (ret == MPG123_NEW_FORMAT)
-        con->new_format = 1;
-
-    return 0;
-}
-
-/* Now we really start accessing some data and determining file format.
- * Format now is allowed to change on-the-fly. Here is the only point
- * that has MPlayer react to errors. We have to pray that exceptional
- * erros in other places simply cannot occur. */
-static int init(struct dec_audio *da, const char *decoder)
-{
-    if (!preinit(da))
-        return 0;
-
-    struct ad_mpg123_context *con = da->priv;
-    int ret;
-
-    ret = mpg123_open_feed(con->handle);
-    if (ret != MPG123_OK)
-        goto fail;
-
-    for (int n = 0; ; n++) {
-        if (feed_new_packet(da) < 0) {
-            ret = MPG123_NEED_MORE;
-            goto fail;
-        }
-        size_t got_now = 0;
-        ret = mpg123_decode_frame(con->handle, NULL, NULL, &got_now);
-        if (ret == MPG123_OK || ret == MPG123_NEW_FORMAT) {
-            ret = set_format(da);
-            if (ret == MPG123_OK)
-                break;
-        }
-        if (ret != MPG123_NEED_MORE)
-            goto fail;
-        // max. 16 retries (randomly chosen number)
-        if (n > 16) {
-            ret = MPG123_NEED_MORE;
-            goto fail;
-        }
-    }
-
-    return 1;
-
-fail:
-    if (ret == MPG123_NEED_MORE) {
-        MP_ERR(da, "Could not find mp3 stream.\n");
-    } else {
-        MP_ERR(da, "mpg123 init error: %s\n",
-               mpg123_strerror(con->handle));
-    }
-
-    uninit(da);
-    return 0;
-}
-
 /* Compute bitrate from frame size. */
 static int compute_bitrate(struct mpg123_frameinfo *i)
 {
@@ -290,57 +186,86 @@ static void update_info(struct dec_audio *da)
     }
 }
 
-static int decode_audio(struct dec_audio *da, struct mp_audio *buffer, int maxlen)
+/* libmpg123 has a new format ready; query and store, return return value
+   of mpg123_getformat() */
+static int set_format(struct dec_audio *da)
+{
+    struct ad_mpg123_context *con = da->priv;
+    int ret;
+    long rate;
+    int channels;
+    int encoding;
+    ret = mpg123_getformat(con->handle, &rate, &channels, &encoding);
+    if (ret == MPG123_OK) {
+        mp_audio_set_num_channels(&da->decoded, channels);
+        da->decoded.rate = rate;
+        int af = mpg123_format_to_af(encoding);
+        if (!af) {
+            /* This means we got a funny custom build of libmpg123 that only supports an unknown format. */
+            MP_ERR(da, "Bad encoding from mpg123: %i.\n", encoding);
+            return MPG123_ERR;
+        }
+        mp_audio_set_format(&da->decoded, af);
+        con->sample_size = channels * af_fmt2bps(af);
+    }
+    return ret;
+}
+
+static int decode_packet(struct dec_audio *da)
 {
     struct ad_mpg123_context *con = da->priv;
-    void *buf = buffer->planes[0];
     int ret;
 
-    if (con->new_format) {
-        ret = set_format(da);
-        if (ret == MPG123_OK) {
-            return 0; // let caller handle format change
-        } else if (ret == MPG123_NEED_MORE) {
-            con->need_data = true;
-        } else {
-            goto mpg123_fail;
-        }
+    mp_audio_set_null_data(&da->decoded);
+
+    struct demux_packet *pkt;
+    if (demux_read_packet_async(da->header, &pkt) == 0)
+        return AD_WAIT;
+    if (!pkt)
+        return AD_EOF;
+
+    /* Next bytes from that presentation time. */
+    if (pkt->pts != MP_NOPTS_VALUE) {
+        da->pts        = pkt->pts;
+        da->pts_offset = 0;
     }
 
-    if (con->need_data) {
-        if (feed_new_packet(da) < 0)
-            return -1;
-    }
+    /* Have to use mpg123_feed() to avoid decoding here. */
+    ret = mpg123_feed(con->handle, pkt->buffer, pkt->len);
+    talloc_free(pkt);
 
-    if (!mp_audio_config_equals(&da->decoded, buffer))
-        return 0;
-
-    size_t got_now = 0;
-    ret = mpg123_replace_buffer(con->handle, buf, maxlen * con->sample_size);
     if (ret != MPG123_OK)
         goto mpg123_fail;
 
-    ret = mpg123_decode_frame(con->handle, NULL, NULL, &got_now);
+    unsigned char *audio = NULL;
+    size_t bytes = 0;
+    ret = mpg123_decode_frame(con->handle, NULL, &audio, &bytes);
 
-    int got_samples = got_now / con->sample_size;
-    buffer->samples += got_samples;
-    da->pts_offset += got_samples;
+    if (ret == MPG123_NEED_MORE)
+        return 0;
 
-    if (ret == MPG123_NEW_FORMAT) {
-        con->new_format = true;
-    } else if (ret == MPG123_NEED_MORE) {
-        con->need_data = true;
-    } else if (ret != MPG123_OK && ret != MPG123_DONE) {
+    if (ret != MPG123_OK && ret != MPG123_DONE && ret != MPG123_NEW_FORMAT)
         goto mpg123_fail;
+
+    ret = set_format(da);
+    if (ret != MPG123_OK)
+        goto mpg123_fail;
+
+    if (con->sample_size < 1) {
+        MP_ERR(da, "no sample size\n");
+        return AD_ERR;
     }
 
+    int got_samples = bytes / con->sample_size;
+    da->decoded.planes[0] = audio;
+    da->decoded.samples = got_samples;
+
     update_info(da);
     return 0;
 
 mpg123_fail:
-    MP_ERR(da, "mpg123 decoding error: %s\n",
-           mpg123_strerror(con->handle));
-    return -1;
+    MP_ERR(da, "mpg123 decoding error: %s\n", mpg123_strerror(con->handle));
+    return AD_ERR;
 }
 
 static int control(struct dec_audio *da, int cmd, void *arg)
@@ -349,6 +274,7 @@ static int control(struct dec_audio *da, int cmd, void *arg)
 
     switch (cmd) {
     case ADCTRL_RESET:
+        mp_audio_set_null_data(&da->decoded);
         mpg123_close(con->handle);
 
         if (mpg123_open_feed(con->handle) != MPG123_OK) {
@@ -373,5 +299,5 @@ const struct ad_functions ad_mpg123 = {
     .init = init,
     .uninit = uninit,
     .control = control,
-    .decode_audio = decode_audio,
+    .decode_packet = decode_packet,
 };

audio/decode/ad_spdif.c

@@ -38,8 +38,7 @@ struct spdifContext {
     AVFormatContext *lavf_ctx;
     int              iec61937_packet_size;
     int              out_buffer_len;
-    int              out_buffer_size;
-    uint8_t         *out_buffer;
+    uint8_t          out_buffer[OUTBUF_SIZE];
     bool             need_close;
 };
 
@@ -47,7 +46,7 @@ static int write_packet(void *p, uint8_t *buf, int buf_size)
 {
     struct spdifContext *ctx = p;
 
-    int buffer_left = ctx->out_buffer_size - ctx->out_buffer_len;
+    int buffer_left = OUTBUF_SIZE - ctx->out_buffer_len;
     if (buf_size > buffer_left) {
         MP_ERR(ctx, "spdif packet too large.\n");
         buf_size = buffer_left;
@@ -118,13 +117,13 @@ static int init(struct dec_audio *da, const char *decoder)
     switch (stream->codec->codec_id) {
     case AV_CODEC_ID_AAC:
         spdif_ctx->iec61937_packet_size = 16384;
-        sample_format                   = AF_FORMAT_IEC61937_LE;
+        sample_format                   = AF_FORMAT_S_AAC;
         samplerate                      = 48000;
         num_channels                    = 2;
         break;
     case AV_CODEC_ID_AC3:
         spdif_ctx->iec61937_packet_size = 6144;
-        sample_format                   = AF_FORMAT_AC3_LE;
+        sample_format                   = AF_FORMAT_S_AC3;
         samplerate                      = 48000;
         num_channels                    = 2;
         break;
@@ -132,31 +131,31 @@ static int init(struct dec_audio *da, const char *decoder)
         if (da->opts->dtshd) {
             av_dict_set(&format_opts, "dtshd_rate", "768000", 0); // 4*192000
             spdif_ctx->iec61937_packet_size = 32768;
-            sample_format                   = AF_FORMAT_IEC61937_LE;
+            sample_format                   = AF_FORMAT_S_DTSHD;
             samplerate                      = 192000;
             num_channels                    = 2*4;
         } else {
             spdif_ctx->iec61937_packet_size = 32768;
-            sample_format                   = AF_FORMAT_AC3_LE;
+            sample_format                   = AF_FORMAT_S_DTS;
             samplerate                      = 48000;
             num_channels                    = 2;
         }
         break;
     case AV_CODEC_ID_EAC3:
         spdif_ctx->iec61937_packet_size = 24576;
-        sample_format                   = AF_FORMAT_IEC61937_LE;
+        sample_format                   = AF_FORMAT_S_EAC3;
         samplerate                      = 192000;
         num_channels                    = 2;
         break;
     case AV_CODEC_ID_MP3:
         spdif_ctx->iec61937_packet_size = 4608;
-        sample_format                   = AF_FORMAT_MPEG2;
+        sample_format                   = AF_FORMAT_S_MP3;
         samplerate                      = 48000;
         num_channels                    = 2;
         break;
     case AV_CODEC_ID_TRUEHD:
         spdif_ctx->iec61937_packet_size = 61440;
-        sample_format                   = AF_FORMAT_IEC61937_LE;
+        sample_format                   = AF_FORMAT_S_TRUEHD;
         samplerate                      = 192000;
         num_channels                    = 8;
         break;
@@ -183,24 +182,21 @@ fail:
     return 0;
 }
 
-static int decode_audio(struct dec_audio *da, struct mp_audio *buffer, int maxlen)
+static int decode_packet(struct dec_audio *da)
 {
     struct spdifContext *spdif_ctx = da->priv;
     AVFormatContext     *lavf_ctx  = spdif_ctx->lavf_ctx;
 
-    int sstride = 2 * da->decoded.channels.num;
-    assert(sstride == buffer->sstride);
-
-    if (maxlen * sstride < spdif_ctx->iec61937_packet_size)
-        return 0;
+    mp_audio_set_null_data(&da->decoded);
 
     spdif_ctx->out_buffer_len  = 0;
-    spdif_ctx->out_buffer_size = maxlen * sstride;
-    spdif_ctx->out_buffer      = buffer->planes[0];
 
-    struct demux_packet *mpkt = demux_read_packet(da->header);
+    struct demux_packet *mpkt;
+    if (demux_read_packet_async(da->header, &mpkt) == 0)
+        return AD_WAIT;
+
     if (!mpkt)
-        return -1;
+        return AD_EOF;
 
     AVPacket pkt;
     mp_set_av_packet(&pkt, mpkt, NULL);
@@ -211,12 +207,13 @@ static int decode_audio(struct dec_audio *da, struct mp_audio *buffer, int maxle
         da->pts_offset = 0;
     }
     int ret = av_write_frame(lavf_ctx, &pkt);
-    avio_flush(lavf_ctx->pb);
-    buffer->samples = spdif_ctx->out_buffer_len / sstride;
-    da->pts_offset += buffer->samples;
     talloc_free(mpkt);
+    avio_flush(lavf_ctx->pb);
     if (ret < 0)
-        return -1;
+        return AD_ERR;
+
+    da->decoded.planes[0] = spdif_ctx->out_buffer;
+    da->decoded.samples = spdif_ctx->out_buffer_len / da->decoded.sstride;
 
     return 0;
 }
@@ -253,5 +250,5 @@ const struct ad_functions ad_spdif = {
     .init = init,
     .uninit = uninit,
     .control = control,
-    .decode_audio = decode_audio,
+    .decode_packet = decode_packet,
 };

audio/decode/dec_audio.c

@@ -28,7 +28,7 @@
 #include "config.h"
 #include "common/codecs.h"
 #include "common/msg.h"
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 
 #include "stream/stream.h"
 #include "demux/demux.h"
@@ -48,21 +48,14 @@ extern const struct ad_functions ad_lavc;
 extern const struct ad_functions ad_spdif;
 
 static const struct ad_functions * const ad_drivers[] = {
+    &ad_lavc,
 #if HAVE_MPG123
     &ad_mpg123,
 #endif
-    &ad_lavc,
     &ad_spdif,
     NULL
 };
 
-// ad_mpg123 needs to be able to decode 1152 samples at once
-// ad_spdif needs up to 8192
-#define DECODE_MAX_UNIT MPMAX(8192, 1152)
-
-// At least 8192 samples, plus hack for ad_mpg123 and ad_spdif
-#define DECODE_BUFFER_SAMPLES (8192 + DECODE_MAX_UNIT)
-
 // Drop audio buffer and reinit it (after format change)
 // Returns whether the format was valid at all.
 static bool reinit_audio_buffer(struct dec_audio *da)
@@ -73,7 +66,6 @@ static bool reinit_audio_buffer(struct dec_audio *da)
         return false;
     }
     mp_audio_buffer_reinit(da->decode_buffer, &da->decoded);
-    mp_audio_buffer_preallocate_min(da->decode_buffer, DECODE_BUFFER_SAMPLES);
     return true;
 }
 
@@ -98,11 +90,6 @@ static int init_audio_codec(struct dec_audio *d_audio, const char *decoder)
     }
 
     d_audio->decode_buffer = mp_audio_buffer_create(NULL);
-    if (!reinit_audio_buffer(d_audio)) {
-        uninit_decoder(d_audio);
-        return 0;
-    }
-
     return 1;
 }
 
@@ -164,9 +151,6 @@ int audio_init_best_codec(struct dec_audio *d_audio, char *audio_decoders)
             talloc_asprintf(d_audio, "%s [%s:%s]", decoder->desc, decoder->family,
                             decoder->decoder);
         MP_VERBOSE(d_audio, "Selected audio codec: %s\n", d_audio->decoder_desc);
-        MP_VERBOSE(d_audio, "AUDIO: %d Hz, %d ch, %s\n",
-                   d_audio->decoded.rate, d_audio->decoded.channels.num,
-                   af_fmt_to_str(d_audio->decoded.format));
     } else {
         MP_ERR(d_audio, "Failed to initialize an audio decoder for codec '%s'.\n",
                d_audio->header->codec ? d_audio->header->codec : "<unknown>");
@@ -209,8 +193,6 @@ int audio_init_filters(struct dec_audio *d_audio, int in_samplerate,
     mp_audio_set_channels(&afs->output, out_channels);
     mp_audio_set_format(&afs->output, *out_format);
 
-    afs->metadata = d_audio->metadata;
-
     afs->replaygain_data = d_audio->replaygain_data;
 
     char *s_from = mp_audio_config_to_str(&afs->input);
@@ -233,61 +215,82 @@ int audio_init_filters(struct dec_audio *d_audio, int in_samplerate,
     return 1;
 }
 
+/* Decode packets until we know the audio format. Then reinit the buffer.
+ * Returns AD_OK on success, negative AD_* code otherwise.
+ * Also returns AD_OK if already initialized (and does nothing).
+ */
+int initial_audio_decode(struct dec_audio *da)
+{
+    while (!mp_audio_config_valid(&da->decoded)) {
+        if (da->decoded.samples > 0)
+            return AD_ERR; // invalid format, rather than uninitialized
+        int ret = da->ad_driver->decode_packet(da);
+        if (ret < 0)
+            return ret;
+    }
+    if (mp_audio_buffer_samples(da->decode_buffer) > 0) // avoid accidental flush
+        return AD_OK;
+    return reinit_audio_buffer(da) ? AD_OK : AD_ERR;
+}
+
 // Filter len bytes of input, put result into outbuf.
 static int filter_n_bytes(struct dec_audio *da, struct mp_audio_buffer *outbuf,
                           int len)
 {
+    bool format_change = false;
     int error = 0;
 
-    struct mp_audio config;
-    mp_audio_buffer_get_format(da->decode_buffer, &config);
+    assert(len > 0); // would break EOF logic below
 
     while (mp_audio_buffer_samples(da->decode_buffer) < len) {
-        int maxlen = mp_audio_buffer_get_write_available(da->decode_buffer);
-        if (maxlen < DECODE_MAX_UNIT)
-            break;
-        struct mp_audio buffer;
-        mp_audio_buffer_get_write_buffer(da->decode_buffer, maxlen, &buffer);
-        buffer.samples = 0;
-        error = da->ad_driver->decode_audio(da, &buffer, maxlen);
-        if (error < 0)
-            break;
-        // Commit the data just read as valid data
-        mp_audio_buffer_finish_write(da->decode_buffer, buffer.samples);
-        // Format change
-        if (!mp_audio_config_equals(&da->decoded, &config)) {
-            // If there are still samples left in the buffer, let them drain
-            // first, and don't signal a format change to the caller yet.
-            if (mp_audio_buffer_samples(da->decode_buffer) > 0)
-                break;
-            error = -2;
+        // Check for a format change
+        struct mp_audio config;
+        mp_audio_buffer_get_format(da->decode_buffer, &config);
+        format_change = !mp_audio_config_equals(&da->decoded, &config);
+        if (format_change) {
+            error = AD_EOF; // drain remaining data left in the current buffer
             break;
         }
+        if (da->decoded.samples > 0) {
+            int copy = MPMIN(da->decoded.samples, len);
+            struct mp_audio append = da->decoded;
+            append.samples = copy;
+            mp_audio_buffer_append(da->decode_buffer, &append);
+            mp_audio_skip_samples(&da->decoded, copy);
+            da->pts_offset += copy;
+            continue;
+        }
+        error = da->ad_driver->decode_packet(da);
+        if (error < 0)
+            break;
     }
 
+    if (error == AD_WAIT)
+        return error;
+
     // Filter
     struct mp_audio filter_data;
     mp_audio_buffer_peek(da->decode_buffer, &filter_data);
     filter_data.rate = da->afilter->input.rate; // due to playback speed change
     len = MPMIN(filter_data.samples, len);
     filter_data.samples = len;
-    bool eof = filter_data.samples == 0 && error < 0;
+    bool eof = error == AD_EOF && filter_data.samples == 0;
 
     if (af_filter(da->afilter, &filter_data, eof ? AF_FILTER_FLAG_EOF : 0) < 0)
-        return -1;
+        return AD_ERR;
 
     mp_audio_buffer_append(outbuf, &filter_data);
-    if (eof && filter_data.samples > 0)
+    if (error == AD_EOF && filter_data.samples > 0)
         error = 0; // don't end playback yet
 
     // remove processed data from decoder buffer:
     mp_audio_buffer_skip(da->decode_buffer, len);
 
-    // Assume the filter chain is drained from old data at this point.
-    // (If not, the remaining old data is discarded.)
-    if (error == -2) {
+    // if format was changed, and all data was drained, execute the format change
+    if (format_change && eof) {
+        error = AD_NEW_FMT;
         if (!reinit_audio_buffer(da))
-            error = -1; // switch to invalid format
+            error = AD_ERR; // switch to invalid format
     }
 
     return error;
@@ -295,17 +298,17 @@ static int filter_n_bytes(struct dec_audio *da, struct mp_audio_buffer *outbuf,
 
 /* Try to get at least minsamples decoded+filtered samples in outbuf
  * (total length including possible existing data).
- * Return 0 on success, -1 on error/EOF (not distinguidaed).
+ * Return 0 on success, or negative AD_* error code.
  * In the former case outbuf has at least minsamples buffered on return.
  * In case of EOF/error it might or might not be. */
 int audio_decode(struct dec_audio *d_audio, struct mp_audio_buffer *outbuf,
                  int minsamples)
 {
+    if (!d_audio->afilter)
+        return AD_ERR;
+
     // Indicates that a filter seems to be buffering large amounts of data
     int huge_filter_buffer = 0;
-    // Decoded audio must be cut at boundaries of this many samples
-    // (Note: the reason for this is unknown, possibly a refactoring artifact)
-    int unitsize = 16;
 
     /* Filter output size will be about filter_multiplier times input size.
      * If some filter buffers audio in big blocks this might only hold
@@ -322,8 +325,8 @@ int audio_decode(struct dec_audio *d_audio, struct mp_audio_buffer *outbuf,
         prev_buffered = buffered;
 
         int decsamples = (minsamples - buffered) / filter_multiplier;
-        // + some extra for possible filter buffering
-        decsamples += unitsize << 5;
+        // + some extra for possible filter buffering, and avoid 0
+        decsamples += 512;
 
         if (huge_filter_buffer) {
             /* Some filter must be doing significant buffering if the estimated

audio/decode/dec_audio.h

@@ -36,12 +36,10 @@ struct dec_audio {
     struct mp_audio_buffer *decode_buffer;
     struct af_stream *afilter;
     char *decoder_desc;
-    struct mp_tags *metadata;
     struct replaygain_data *replaygain_data;
+    int init_retries;
     // set by decoder
-    struct mp_audio decoded;    // format of decoded audio (no data, temporarily
-                                // different from decode_buffer during format
-                                // changes)
+    struct mp_audio decoded;    // decoded audio set by last decode_packet() call
     int bitrate;                // input bitrate, can change with VBR sources
     // last known pts value in output from decoder
     double pts;
@@ -51,10 +49,19 @@ struct dec_audio {
     void *priv;
 };
 
+enum {
+    AD_OK = 0,
+    AD_ERR = -1,
+    AD_EOF = -2,
+    AD_NEW_FMT = -3,
+    AD_WAIT = -4,
+};
+
 struct mp_decoder_list *audio_decoder_list(void);
 int audio_init_best_codec(struct dec_audio *d_audio, char *audio_decoders);
 int audio_decode(struct dec_audio *d_audio, struct mp_audio_buffer *outbuf,
                  int minsamples);
+int initial_audio_decode(struct dec_audio *d_audio);
 void audio_reset_decoding(struct dec_audio *d_audio);
 void audio_uninit(struct dec_audio *d_audio);
 

audio/filter/af.c

@@ -204,7 +204,6 @@ static struct af_instance *af_create(struct af_stream *s, char *name,
         .mul = 1,
         .data = talloc_zero(af, struct mp_audio),
         .log = mp_log_new(af, s->log, name),
-        .metadata = s->metadata,
         .replaygain_data = s->replaygain_data,
     };
     struct m_config *config = m_config_from_obj_desc(af, s->log, &desc);
@@ -311,22 +310,22 @@ static void af_print_filter_chain(struct af_stream *s, struct af_instance *at,
 
     struct af_instance *af = s->first;
     while (af) {
-        MP_MSG(s, msg_level, "  [%s] ", af->info->name);
+        char b[128] = {0};
+        mp_snprintf_cat(b, sizeof(b), "  [%s] ", af->info->name);
         if (af->data) {
             char *info = mp_audio_config_to_str(af->data);
-            MP_MSG(s, msg_level, "%s", info);
+            mp_snprintf_cat(b, sizeof(b), "%s", info);
             talloc_free(info);
         }
         if (af == at)
-            MP_MSG(s, msg_level, " <-");
-        MP_MSG(s, msg_level, "\n");
+            mp_snprintf_cat(b, sizeof(b), " <-");
+        MP_MSG(s, msg_level, "%s\n", b);
 
         af = af->next;
     }
 
-    MP_MSG(s, msg_level, "  [ao] ");
     char *info = mp_audio_config_to_str(&s->output);
-    MP_MSG(s, msg_level, "%s\n", info);
+    MP_MSG(s, msg_level, "  [ao] %s\n", info);
     talloc_free(info);
 }
 

audio/filter/af.h

@@ -59,7 +59,6 @@ struct af_info {
 struct af_instance {
     const struct af_info *info;
     struct mp_log *log;
-    struct mp_tags *metadata;
     struct replaygain_data *replaygain_data;
     int (*control)(struct af_instance *af, int cmd, void *arg);
     void (*uninit)(struct af_instance *af);
@@ -92,7 +91,6 @@ struct af_stream {
 
     struct mp_log *log;
     struct MPOpts *opts;
-    struct mp_tags *metadata;
     struct replaygain_data *replaygain_data;
 };
 

audio/filter/af_bs2b.c

@@ -48,20 +48,12 @@ static int filter_##name(struct af_instance *af, struct mp_audio *data, int f) \
 }
 
 FILTER(f, float)
-FILTER(fbe, float)
-FILTER(fle, float)
-FILTER(s32be, int32_t)
-FILTER(u32be, uint32_t)
-FILTER(s32le, int32_t)
-FILTER(u32le, uint32_t)
-FILTER(s24be, bs2b_int24_t)
-FILTER(u24be, bs2b_uint24_t)
-FILTER(s24le, bs2b_int24_t)
-FILTER(u24le, bs2b_uint24_t)
-FILTER(s16be, int16_t)
-FILTER(u16be, uint16_t)
-FILTER(s16le, int16_t)
-FILTER(u16le, uint16_t)
+FILTER(s32, int32_t)
+FILTER(u32, uint32_t)
+FILTER(s24, bs2b_int24_t)
+FILTER(u24, bs2b_uint24_t)
+FILTER(s16, int16_t)
+FILTER(u16, uint16_t)
 FILTER(s8, int8_t)
 FILTER(u8, uint8_t)
 
@@ -85,47 +77,26 @@ static int control(struct af_instance *af, int cmd, void *arg)
         /* check for formats supported by libbs2b
            and assign corresponding handlers */
         switch (format) {
-            case AF_FORMAT_FLOAT_BE:
-                af->filter = filter_fbe;
+            case AF_FORMAT_FLOAT:
+                af->filter = filter_f;
                 break;
-            case AF_FORMAT_FLOAT_LE:
-                af->filter = filter_fle;
+            case AF_FORMAT_S32:
+                af->filter = filter_s32;
                 break;
-            case AF_FORMAT_S32_BE:
-                af->filter = filter_s32be;
+            case AF_FORMAT_U32:
+                af->filter = filter_u32;
                 break;
-            case AF_FORMAT_U32_BE:
-                af->filter = filter_u32be;
+            case AF_FORMAT_S24:
+                af->filter = filter_s24;
                 break;
-            case AF_FORMAT_S32_LE:
-                af->filter = filter_s32le;
+            case AF_FORMAT_U24:
+                af->filter = filter_u24;
                 break;
-            case AF_FORMAT_U32_LE:
-                af->filter = filter_u32le;
+            case AF_FORMAT_S16:
+                af->filter = filter_s16;
                 break;
-            case AF_FORMAT_S24_BE:
-                af->filter = filter_s24be;
-                break;
-            case AF_FORMAT_U24_BE:
-                af->filter = filter_u24be;
-                break;
-            case AF_FORMAT_S24_LE:
-                af->filter = filter_s24le;
-                break;
-            case AF_FORMAT_U24_LE:
-                af->filter = filter_u24le;
-                break;
-            case AF_FORMAT_S16_BE:
-                af->filter = filter_s16be;
-                break;
-            case AF_FORMAT_U16_BE:
-                af->filter = filter_u16be;
-                break;
-            case AF_FORMAT_S16_LE:
-                af->filter = filter_s16le;
-                break;
-            case AF_FORMAT_U16_LE:
-                af->filter = filter_u16le;
+            case AF_FORMAT_U16:
+                af->filter = filter_u16;
                 break;
             case AF_FORMAT_S8:
                 af->filter = filter_s8;

audio/filter/af_convert24.c

@@ -20,6 +20,7 @@
 
 #include "audio/format.h"
 #include "af.h"
+#include "osdep/endian.h"
 
 static bool test_conversion(int src_format, int dst_format)
 {

audio/filter/af_convertsignendian.c

@@ -20,21 +20,16 @@
 
 #include "af.h"
 #include "audio/format.h"
-#include "compat/mpbswap.h"
+#include "osdep/endian.h"
 
 static bool test_conversion(int src_format, int dst_format)
 {
     if ((src_format & AF_FORMAT_PLANAR) ||
         (dst_format & AF_FORMAT_PLANAR))
         return false;
-    int src_noend = src_format & ~AF_FORMAT_END_MASK;
-    int dst_noend = dst_format & ~AF_FORMAT_END_MASK;
-    // We can swap endian for all formats, but sign only for integer formats.
-    if (src_noend == dst_noend)
-        return true;
-    if (((src_noend & ~AF_FORMAT_SIGN_MASK) ==
-         (dst_noend & ~AF_FORMAT_SIGN_MASK)) &&
-        ((src_noend & AF_FORMAT_POINT_MASK) == AF_FORMAT_I))
+    if (((src_format & ~AF_FORMAT_SIGN_MASK) ==
+         (dst_format & ~AF_FORMAT_SIGN_MASK)) &&
+        ((src_format & AF_FORMAT_TYPE_MASK) == AF_FORMAT_I))
         return true;
     return false;
 }
@@ -63,34 +58,11 @@ static int control(struct af_instance *af, int cmd, void *arg)
     return AF_UNKNOWN;
 }
 
-static void endian(void *data, int len, int bps)
-{
-    switch (bps) {
-    case 2:
-        for (int i = 0; i < len; i++) {
-            ((uint16_t*)data)[i] = bswap_16(((uint16_t *)data)[i]);
-        }
-        break;
-    case 3:
-        for(int i = 0; i < len; i++) {
-            uint8_t s = ((uint8_t *)data)[3 * i];
-            ((uint8_t *)data)[3 * i] = ((uint8_t *)data)[3 * i + 2];
-            ((uint8_t *)data)[3 * i + 2] = s;
-        }
-        break;
-    case 4:
-        for(int i = 0; i < len; i++) {
-            ((uint32_t*)data)[i] = bswap_32(((uint32_t *)data)[i]);
-        }
-        break;
-    }
-}
-
-static void si2us(void *data, int len, int bps, bool le)
+static void si2us(void *data, int len, int bps)
 {
     ptrdiff_t i = -(len * bps);
     uint8_t *p = &((uint8_t *)data)[len * bps];
-    if (le && bps > 1)
+    if (BYTE_ORDER == LITTLE_ENDIAN && bps > 1)
         p += bps - 1;
     if (len <= 0)
         return;
@@ -105,12 +77,8 @@ static int filter(struct af_instance *af, struct mp_audio *data, int flags)
     int outfmt = af->data->format;
     size_t len = data->samples * data->nch;
 
-    if ((infmt & AF_FORMAT_END_MASK) != (outfmt & AF_FORMAT_END_MASK))
-        endian(data->planes[0], len, data->bps);
-
     if ((infmt & AF_FORMAT_SIGN_MASK) != (outfmt & AF_FORMAT_SIGN_MASK))
-        si2us(data->planes[0], len, data->bps,
-              (outfmt & AF_FORMAT_END_MASK) == AF_FORMAT_LE);
+        si2us(data->planes[0], len, data->bps);
 
     mp_audio_set_format(data, outfmt);
     return 0;
@@ -124,8 +92,8 @@ static int af_open(struct af_instance *af)
 }
 
 const struct af_info af_info_convertsignendian = {
-    .info = "Convert between sample format sign/endian",
-    .name = "convertsignendian",
+    .info = "Convert between sample format sign",
+    .name = "convertsign",
     .open = af_open,
     .test_conversion = test_conversion,
 };

audio/filter/af_hrtf.c

@@ -284,6 +284,27 @@ static inline void update_ch(af_hrtf_t *s, short *in, const int k)
     s->ba_r[k] = in[4] + in[1] + in[3];
 }
 
+static void clear_coeff(af_hrtf_t *s, float *c)
+{
+    memset(c, 0, s->dlbuflen * sizeof(float));
+}
+
+static void reset(af_hrtf_t *s)
+{
+    clear_coeff(s, s->lf);
+    clear_coeff(s, s->rf);
+    clear_coeff(s, s->lr);
+    clear_coeff(s, s->rr);
+    clear_coeff(s, s->cf);
+    clear_coeff(s, s->cr);
+    clear_coeff(s, s->ba_l);
+    clear_coeff(s, s->ba_r);
+    clear_coeff(s, s->fwrbuf_l);
+    clear_coeff(s, s->fwrbuf_r);
+    clear_coeff(s, s->fwrbuf_lr);
+    clear_coeff(s, s->fwrbuf_rr);
+}
+
 /* Initialization and runtime control */
 static int control(struct af_instance *af, int cmd, void* arg)
 {
@@ -292,30 +313,27 @@ static int control(struct af_instance *af, int cmd, void* arg)
 
     switch(cmd) {
     case AF_CONTROL_REINIT:
-        af->data->rate   = ((struct mp_audio*)arg)->rate;
-        if(af->data->rate != 48000) {
-            // automatic samplerate adjustment in the filter chain
-            // is not yet supported.
-            MP_ERR(af, "ERROR: Sampling rate is not 48000 Hz (%d)!\n",
-                   af->data->rate);
-            return AF_ERROR;
-        }
+        reset(s);
+        af->data->rate = 48000;
         mp_audio_set_channels_old(af->data, ((struct mp_audio*)arg)->nch);
-            if(af->data->nch == 2) {
-               /* 2 channel input */
-               if(s->decode_mode != HRTF_MIX_MATRIX2CH) {
-                  /* Default behavior is stereo mixing. */
-                  s->decode_mode = HRTF_MIX_STEREO;
-               }
+        if(af->data->nch == 2) {
+            /* 2 channel input */
+            if(s->decode_mode != HRTF_MIX_MATRIX2CH) {
+                /* Default behavior is stereo mixing. */
+                s->decode_mode = HRTF_MIX_STEREO;
             }
-            else if (af->data->nch < 5)
-              mp_audio_set_channels_old(af->data, 5);
+        } else if (af->data->nch < 5) {
+            mp_audio_set_channels_old(af->data, 5);
+        }
         mp_audio_set_format(af->data, AF_FORMAT_S16);
         test_output_res = af_test_output(af, (struct mp_audio*)arg);
         // after testing input set the real output format
         mp_audio_set_num_channels(af->data, 2);
         s->print_flag = 1;
         return test_output_res;
+    case AF_CONTROL_RESET:
+        reset(s);
+        return AF_OK;
     }
 
     return AF_UNKNOWN;

audio/filter/af_lavcac3enc.c

@@ -30,6 +30,7 @@
 #include <libavutil/audioconvert.h>
 #include <libavutil/intreadwrite.h>
 #include <libavutil/common.h>
+#include <libavutil/bswap.h>
 #include <libavutil/mem.h>
 
 #include "common/common.h"
@@ -87,7 +88,7 @@ static int control(struct af_instance *af, int cmd, void *arg)
         if (in->nch > AC3_MAX_CHANNELS)
             mp_audio_set_num_channels(in, AC3_MAX_CHANNELS);
 
-        mp_audio_set_format(af->data, AF_FORMAT_AC3_BE);
+        mp_audio_set_format(af->data, AF_FORMAT_S_AC3);
         mp_audio_set_num_channels(af->data, 2);
 
         if (!mp_audio_config_equals(in, &orig_in))
@@ -150,6 +151,12 @@ static void uninit(struct af_instance* af)
     }
 }
 
+static void swap_16(uint16_t *ptr, size_t size)
+{
+    for (size_t n = 0; n < size; n++)
+        ptr[n] = av_bswap16(ptr[n]);
+}
+
 // Filter data through filter
 static int filter(struct af_instance* af, struct mp_audio* audio, int flags)
 {
@@ -223,11 +230,11 @@ static int filter(struct af_instance* af, struct mp_audio* audio, int flags)
             frame_size = AC3_FRAME_SIZE * 2 * 2;
             header_len = 8;
 
-            AV_WB16(hdr,     0xF872);   // iec 61937 syncword 1
-            AV_WB16(hdr + 2, 0x4E1F);   // iec 61937 syncword 2
-            hdr[4] = bsmod;             // bsmod
-            hdr[5] = 0x01;              // data-type ac3
-            AV_WB16(hdr + 6, len << 3); // number of bits in payload
+            AV_WL16(hdr,     0xF872);   // iec 61937 syncword 1
+            AV_WL16(hdr + 2, 0x4E1F);   // iec 61937 syncword 2
+            hdr[5] = bsmod;             // bsmod
+            hdr[4] = 0x01;              // data-type ac3
+            AV_WL16(hdr + 6, len << 3); // number of bits in payload
         }
 
         size_t max_size = (max_out_samples - out->samples) * out->sstride;
@@ -239,6 +246,7 @@ static int filter(struct af_instance* af, struct mp_audio* audio, int flags)
         memcpy(buf + header_len, s->pkt.data, s->pkt.size);
         memset(buf + header_len + s->pkt.size, 0,
                frame_size - (header_len + s->pkt.size));
+        swap_16((uint16_t *)(buf + header_len), s->pkt.size / 2);
         out->samples += frame_size / out->sstride;
     }
 
@@ -312,12 +320,13 @@ const struct af_info af_info_lavcac3enc = {
     .priv_size = sizeof(struct af_ac3enc_s),
     .priv_defaults = &(const struct af_ac3enc_s){
         .cfg_add_iec61937_header = 1,
-        .cfg_min_channel_num = 5,
+        .cfg_bit_rate = 640,
+        .cfg_min_channel_num = 3,
     },
     .options = (const struct m_option[]) {
         OPT_FLAG("tospdif", cfg_add_iec61937_header, 0),
         OPT_CHOICE_OR_INT("bitrate", cfg_bit_rate, 0, 32, 640,
-                          ({"default", 0})),
+                          ({"auto", 0}, {"default", 0})),
         OPT_INTRANGE("minch", cfg_min_channel_num, 0, 2, 6),
         {0}
     },

audio/filter/af_lavfi.c

@@ -38,8 +38,9 @@
 #include "audio/fmt-conversion.h"
 #include "af.h"
 
+#include "common/av_common.h"
+
 #include "options/m_option.h"
-#include "common/av_opts.h"
 
 // FFmpeg and Libav have slightly different APIs, just enough to cause us
 // unnecessary pain. <Expletive deleted.>
@@ -62,7 +63,7 @@ struct priv {
 
     // options
     char *cfg_graph;
-    char *cfg_avopts;
+    char **cfg_avopts;
 };
 
 static void destroy_graph(struct af_instance *af)
@@ -90,11 +91,8 @@ static bool recreate_graph(struct af_instance *af, struct mp_audio *config)
     if (!graph)
         goto error;
 
-    if (parse_avopts(graph, p->cfg_avopts) < 0) {
-        MP_FATAL(af, "lavfi: could not set opts: '%s'\n",
-               p->cfg_avopts);
+    if (mp_set_avopts(af->log, graph, p->cfg_avopts) < 0)
         goto error;
-    }
 
     AVFilterInOut *outputs = avfilter_inout_alloc();
     AVFilterInOut *inputs  = avfilter_inout_alloc();
@@ -306,7 +304,7 @@ const struct af_info af_info_lavfi = {
     .priv_size = sizeof(struct priv),
     .options = (const struct m_option[]) {
         OPT_STRING("graph", cfg_graph, 0),
-        OPT_STRING("o", cfg_avopts, 0),
+        OPT_KEYVALUELIST("o", cfg_avopts, 0),
         {0}
     },
 };

audio/filter/af_lavrresample.c

@@ -52,9 +52,9 @@
 #error "config.h broken or no resampler found"
 #endif
 
+#include "common/av_common.h"
 #include "common/msg.h"
 #include "options/m_option.h"
-#include "common/av_opts.h"
 #include "audio/filter/af.h"
 #include "audio/fmt-conversion.h"
 
@@ -74,7 +74,7 @@ struct af_resample_opts {
 
 struct af_resample {
     int allow_detach;
-    char *avopts;
+    char **avopts;
     struct AVAudioResampleContext *avrctx;
     struct AVAudioResampleContext *avrctx_out; // for output channel reordering
     struct af_resample_opts ctx;   // opts in the context
@@ -164,10 +164,8 @@ static int configure_lavrr(struct af_instance *af, struct mp_audio *in,
 
     av_opt_set_double(s->avrctx, "cutoff",          s->ctx.cutoff, 0);
 
-    if (parse_avopts(s->avrctx, s->avopts) < 0) {
-        MP_FATAL(af, "af_lavrresample: could not set opts: '%s'\n", s->avopts);
+    if (mp_set_avopts(af->log, s->avrctx, s->avopts) < 0)
         return AF_ERROR;
-    }
 
     struct mp_chmap map_in = in->channels;
     struct mp_chmap map_out = out->channels;
@@ -216,8 +214,7 @@ static int configure_lavrr(struct af_instance *af, struct mp_audio *in,
     if (avresample_open(s->avrctx) < 0 ||
         avresample_open(s->avrctx_out) < 0)
     {
-        MP_ERR(af, "Cannot open "
-                "Libavresample Context. \n");
+        MP_ERR(af, "Cannot open Libavresample Context. \n");
         return AF_ERROR;
     }
     return AF_OK;
@@ -341,7 +338,7 @@ static int filter(struct af_instance *af, struct mp_audio *data, int flags)
     if (needs_reorder(s->reorder_out, out->nch)) {
         if (af_fmt_is_planar(out->format)) {
             reorder_planes(data, s->reorder_out);
-        } else {
+        } else if (out->samples) {
             int out_size = out->samples * out->sstride;
             if (talloc_get_size(s->reorder_buffer) < out_size)
                 s->reorder_buffer = talloc_realloc_size(s, s->reorder_buffer, out_size);
@@ -349,7 +346,8 @@ static int filter(struct af_instance *af, struct mp_audio *data, int flags)
             int out_samples = avresample_convert(s->avrctx_out,
                     (uint8_t **) data->planes, out_size, out->samples,
                     (uint8_t **) out->planes, out_size, out->samples);
-            assert(out_samples == data->samples);
+            if (out_samples < 0)
+                MP_ERR(af, "Reordering failed.\n");
         }
     }
 
@@ -373,8 +371,7 @@ static int af_open(struct af_instance *af)
     if (s->avrctx && s->avrctx_out) {
         return AF_OK;
     } else {
-        MP_ERR(af, "Cannot initialize "
-               "Libavresample Context. \n");
+        MP_ERR(af, "Cannot initialize Libavresample Context. \n");
         uninit(af);
         return AF_ERROR;
     }
@@ -402,7 +399,7 @@ const struct af_info af_info_lavrresample = {
         OPT_FLAG("linear", opts.linear, 0),
         OPT_DOUBLE("cutoff", opts.cutoff, M_OPT_RANGE, .min = 0, .max = 1),
         OPT_FLAG("detach", allow_detach, 0),
-        OPT_STRING("o", avopts, 0),
+        OPT_KEYVALUELIST("o", avopts, 0),
         {0}
     },
 };

audio/filter/af_volume.c

@@ -79,7 +79,7 @@ static int control(struct af_instance *af, int cmd, void *arg)
             if (!s->rgain_clip) // clipping prevention
                 s->rgain = MPMIN(s->rgain, 1.0 / peak);
         }
-        if (s->detach && fabs(s->level + s->rgain - 2.0) < 0.00001)
+        if (s->detach && fabs(s->level * s->rgain - 1.0) < 0.00001)
             return AF_DETACH;
         return af_test_output(af, in);
     }
@@ -97,7 +97,7 @@ static void filter_plane(struct af_instance *af, void *ptr, int num_samples)
 {
     struct priv *s = af->priv;
 
-    float level = s->level + s->rgain - 1.0;
+    float level = s->level * s->rgain;
 
     if (af_fmt_from_planar(af->data->format) == AF_FORMAT_S16) {
         int16_t *a = ptr;

audio/format.c

@@ -30,9 +30,6 @@
 
 int af_fmt2bps(int format)
 {
-    if (AF_FORMAT_IS_AC3(format)) return 2;
-    if (format == AF_FORMAT_UNKNOWN)
-        return 0;
     switch (format & AF_FORMAT_BITS_MASK) {
     case AF_FORMAT_8BIT:  return 1;
     case AF_FORMAT_16BIT: return 2;
@@ -62,7 +59,7 @@ static int bits_to_mask(int bits)
 
 int af_fmt_change_bits(int format, int bits)
 {
-    if (!af_fmt_is_valid(format) || (format & AF_FORMAT_SPECIAL_MASK))
+    if (!af_fmt_is_valid(format))
         return 0;
     int mask = bits_to_mask(bits);
     format = (format & ~AF_FORMAT_BITS_MASK) | mask;
@@ -109,35 +106,31 @@ bool af_fmt_is_planar(int format)
     return !!(format & AF_FORMAT_PLANAR);
 }
 
-#define FMT(string, id)                                 \
-    {string,            id},
-
-#define FMT_ENDIAN(string, id)                          \
-    {string,            id},                            \
-    {string "le",       MP_CONCAT(id, _LE)},            \
-    {string "be",       MP_CONCAT(id, _BE)},            \
-
 const struct af_fmt_entry af_fmtstr_table[] = {
-    FMT("mpeg2",                AF_FORMAT_MPEG2)
-    FMT_ENDIAN("ac3",           AF_FORMAT_AC3)
-    FMT_ENDIAN("iec61937",      AF_FORMAT_IEC61937)
+    {"u8",          AF_FORMAT_U8},
+    {"s8",          AF_FORMAT_S8},
+    {"u16",         AF_FORMAT_U16},
+    {"s16",         AF_FORMAT_S16},
+    {"u24",         AF_FORMAT_U24},
+    {"s24",         AF_FORMAT_S24},
+    {"u32",         AF_FORMAT_U32},
+    {"s32",         AF_FORMAT_S32},
+    {"float",       AF_FORMAT_FLOAT},
+    {"double",      AF_FORMAT_DOUBLE},
 
-    FMT("u8",                   AF_FORMAT_U8)
-    FMT("s8",                   AF_FORMAT_S8)
-    FMT_ENDIAN("u16",           AF_FORMAT_U16)
-    FMT_ENDIAN("s16",           AF_FORMAT_S16)
-    FMT_ENDIAN("u24",           AF_FORMAT_U24)
-    FMT_ENDIAN("s24",           AF_FORMAT_S24)
-    FMT_ENDIAN("u32",           AF_FORMAT_U32)
-    FMT_ENDIAN("s32",           AF_FORMAT_S32)
-    FMT_ENDIAN("float",         AF_FORMAT_FLOAT)
-    FMT_ENDIAN("double",        AF_FORMAT_DOUBLE)
+    {"u8p",         AF_FORMAT_U8P},
+    {"s16p",        AF_FORMAT_S16P},
+    {"s32p",        AF_FORMAT_S32P},
+    {"floatp",      AF_FORMAT_FLOATP},
+    {"doublep",     AF_FORMAT_DOUBLEP},
 
-    FMT("u8p",                  AF_FORMAT_U8P)
-    FMT("s16p",                 AF_FORMAT_S16P)
-    FMT("s32p",                 AF_FORMAT_S32P)
-    FMT("floatp",               AF_FORMAT_FLOATP)
-    FMT("doublep",              AF_FORMAT_DOUBLEP)
+    {"spdif-aac",   AF_FORMAT_S_AAC},
+    {"spdif-ac3",   AF_FORMAT_S_AC3},
+    {"spdif-dts",   AF_FORMAT_S_DTS},
+    {"spdif-dtshd", AF_FORMAT_S_DTSHD},
+    {"spdif-eac3",  AF_FORMAT_S_EAC3},
+    {"spdif-mp3",   AF_FORMAT_S_MP3},
+    {"spdif-truehd",AF_FORMAT_S_TRUEHD},
 
     {0}
 };
@@ -199,22 +192,17 @@ int af_format_conversion_score(int dst_format, int src_format)
         return INT_MIN;
     if (dst_format == src_format)
         return 1024;
-    // Just endian swapping (separate because it works for special formats)
-    if ((dst_format & ~AF_FORMAT_END_MASK) == (src_format & ~AF_FORMAT_END_MASK))
-        return 1024 - 2;
     // Can't be normally converted
     if (AF_FORMAT_IS_SPECIAL(dst_format) || AF_FORMAT_IS_SPECIAL(src_format))
         return INT_MIN;
     int score = 1024;
     if (FMT_DIFF(AF_FORMAT_INTERLEAVING_MASK, dst_format, src_format))
         score -= 1;     // has to (de-)planarize
-    if (FMT_DIFF(AF_FORMAT_END_MASK, dst_format, src_format))
-        score -= 2;     // has to swap endian
     if (FMT_DIFF(AF_FORMAT_SIGN_MASK, dst_format, src_format))
         score -= 4;     // has to swap sign
-    if (FMT_DIFF(AF_FORMAT_POINT_MASK, dst_format, src_format)) {
+    if (FMT_DIFF(AF_FORMAT_TYPE_MASK, dst_format, src_format)) {
         int dst_bits = dst_format & AF_FORMAT_BITS_MASK;
-        if ((dst_format & AF_FORMAT_POINT_MASK) == AF_FORMAT_F) {
+        if ((dst_format & AF_FORMAT_TYPE_MASK) == AF_FORMAT_F) {
             // For int->float, always prefer 32 bit float.
             score -= dst_bits == AF_FORMAT_32BIT ? 8 : 0;
         } else {
@@ -230,7 +218,7 @@ int af_format_conversion_score(int dst_format, int src_format)
         }
     }
     // Consider this the worst case.
-    if (FMT_DIFF(AF_FORMAT_POINT_MASK, dst_format, src_format))
+    if (FMT_DIFF(AF_FORMAT_TYPE_MASK, dst_format, src_format))
         score -= 2048;  // has to convert float<->int
     return score;
 }

audio/format.h

@@ -24,122 +24,78 @@
 #define MPLAYER_AF_FORMAT_H
 
 #include <stdbool.h>
-#include <sys/types.h>
 
-#include "bstr/bstr.h"
-
-#if BYTE_ORDER == BIG_ENDIAN
-#define AF_SELECT_LE_BE(LE, BE) BE
-#else
-#define AF_SELECT_LE_BE(LE, BE) LE
-#endif
-
-// Endianness
-#define AF_FORMAT_BE            (0<<0) // Big Endian
-#define AF_FORMAT_LE            (1<<0) // Little Endian
-#define AF_FORMAT_END_MASK      (1<<0)
-
-#define AF_FORMAT_NE AF_SELECT_LE_BE(AF_FORMAT_LE, AF_FORMAT_BE)
+#include "misc/bstr.h"
 
 // Signed/unsigned
-#define AF_FORMAT_SI            (0<<1) // Signed
-#define AF_FORMAT_US            (1<<1) // Unsigned
-#define AF_FORMAT_SIGN_MASK     (1<<1)
+#define AF_FORMAT_SI            (0<<0) // Signed
+#define AF_FORMAT_US            (1<<0) // Unsigned
+#define AF_FORMAT_SIGN_MASK     (1<<0)
 
 // Bits used
 // Some code assumes they're sorted by size.
-#define AF_FORMAT_8BIT          (0<<3)
-#define AF_FORMAT_16BIT         (1<<3)
-#define AF_FORMAT_24BIT         (2<<3)
-#define AF_FORMAT_32BIT         (3<<3)
-#define AF_FORMAT_64BIT         (4<<3)
-#define AF_FORMAT_BITS_MASK     (7<<3)
+#define AF_FORMAT_8BIT          (0<<1)
+#define AF_FORMAT_16BIT         (1<<1)
+#define AF_FORMAT_24BIT         (2<<1)
+#define AF_FORMAT_32BIT         (3<<1)
+#define AF_FORMAT_64BIT         (4<<1)
+#define AF_FORMAT_BITS_MASK     (7<<1)
 
-// Special flags refering to non pcm data (note: 1<<6, 2<<6, 5<<6 unused)
-#define AF_FORMAT_S_MPEG2       (3<<6) // MPEG(2) audio
-#define AF_FORMAT_S_AC3         (4<<6) // Dolby Digital AC3
-#define AF_FORMAT_S_IEC61937    (6<<6)
-#define AF_FORMAT_SPECIAL_MASK  (7<<6)
-
-// Fixed or floating point
-#define AF_FORMAT_I             (1<<9) // Int
-#define AF_FORMAT_F             (2<<9) // Foating point
-#define AF_FORMAT_POINT_MASK    (3<<9)
+// Fixed/floating point/special
+#define AF_FORMAT_I             (1<<4) // Int
+#define AF_FORMAT_F             (2<<4) // Foating point
+#define AF_FORMAT_S             (4<<4) // special (IEC61937)
+#define AF_FORMAT_TYPE_MASK     (7<<4)
 
 // Interleaving (planar formats have data for each channel in separate planes)
-#define AF_FORMAT_INTERLEAVED        (0<<11) // must be 0
-#define AF_FORMAT_PLANAR             (1<<11)
-#define AF_FORMAT_INTERLEAVING_MASK  (1<<11)
+#define AF_FORMAT_INTERLEAVED        (0<<7) // must be 0
+#define AF_FORMAT_PLANAR             (1<<7)
+#define AF_FORMAT_INTERLEAVING_MASK  (1<<7)
+
+#define AF_FORMAT_S_CODEC(n)    ((n)<<8)
+#define AF_FORMAT_S_CODEC_MASK  (15 <<8) // 16 codecs max.
 
 #define AF_FORMAT_MASK          ((1<<12)-1)
 
 #define AF_INTP (AF_FORMAT_I|AF_FORMAT_PLANAR)
 #define AF_FLTP (AF_FORMAT_F|AF_FORMAT_PLANAR)
+#define AF_FORMAT_S_(n) (AF_FORMAT_S_CODEC(n)|AF_FORMAT_S|AF_FORMAT_16BIT)
 
 // actual sample formats
 enum af_format {
     AF_FORMAT_UNKNOWN   = 0,
 
-    AF_FORMAT_U8        = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_8BIT|AF_FORMAT_NE),
-    AF_FORMAT_S8        = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_8BIT|AF_FORMAT_NE),
-    AF_FORMAT_U16_LE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_16BIT|AF_FORMAT_LE),
-    AF_FORMAT_U16_BE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_16BIT|AF_FORMAT_BE),
-    AF_FORMAT_S16_LE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_16BIT|AF_FORMAT_LE),
-    AF_FORMAT_S16_BE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_16BIT|AF_FORMAT_BE),
-    AF_FORMAT_U24_LE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_24BIT|AF_FORMAT_LE),
-    AF_FORMAT_U24_BE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_24BIT|AF_FORMAT_BE),
-    AF_FORMAT_S24_LE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_24BIT|AF_FORMAT_LE),
-    AF_FORMAT_S24_BE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_24BIT|AF_FORMAT_BE),
-    AF_FORMAT_U32_LE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_32BIT|AF_FORMAT_LE),
-    AF_FORMAT_U32_BE    = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_32BIT|AF_FORMAT_BE),
-    AF_FORMAT_S32_LE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_32BIT|AF_FORMAT_LE),
-    AF_FORMAT_S32_BE    = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_32BIT|AF_FORMAT_BE),
+    AF_FORMAT_U8        = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_8BIT),
+    AF_FORMAT_S8        = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_8BIT),
+    AF_FORMAT_U16       = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_16BIT),
+    AF_FORMAT_S16       = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_16BIT),
+    AF_FORMAT_U24       = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_24BIT),
+    AF_FORMAT_S24       = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_24BIT),
+    AF_FORMAT_U32       = (AF_FORMAT_I|AF_FORMAT_US|AF_FORMAT_32BIT),
+    AF_FORMAT_S32       = (AF_FORMAT_I|AF_FORMAT_SI|AF_FORMAT_32BIT),
 
-    AF_FORMAT_FLOAT_LE  = (AF_FORMAT_F|AF_FORMAT_32BIT|AF_FORMAT_LE),
-    AF_FORMAT_FLOAT_BE  = (AF_FORMAT_F|AF_FORMAT_32BIT|AF_FORMAT_BE),
-
-    AF_FORMAT_DOUBLE_LE = (AF_FORMAT_F|AF_FORMAT_64BIT|AF_FORMAT_LE),
-    AF_FORMAT_DOUBLE_BE = (AF_FORMAT_F|AF_FORMAT_64BIT|AF_FORMAT_BE),
-
-    AF_FORMAT_AC3_LE    = (AF_FORMAT_S_AC3|AF_FORMAT_16BIT|AF_FORMAT_LE),
-    AF_FORMAT_AC3_BE    = (AF_FORMAT_S_AC3|AF_FORMAT_16BIT|AF_FORMAT_BE),
-
-    AF_FORMAT_IEC61937_LE = (AF_FORMAT_S_IEC61937|AF_FORMAT_16BIT|AF_FORMAT_LE),
-    AF_FORMAT_IEC61937_BE = (AF_FORMAT_S_IEC61937|AF_FORMAT_16BIT|AF_FORMAT_BE),
-
-    AF_FORMAT_MPEG2     = (AF_FORMAT_S_MPEG2),
+    AF_FORMAT_FLOAT     = (AF_FORMAT_F|AF_FORMAT_32BIT),
+    AF_FORMAT_DOUBLE    = (AF_FORMAT_F|AF_FORMAT_64BIT),
 
     // Planar variants
-    AF_FORMAT_U8P       = (AF_INTP|AF_FORMAT_US|AF_FORMAT_8BIT|AF_FORMAT_NE),
-    AF_FORMAT_S16P      = (AF_INTP|AF_FORMAT_SI|AF_FORMAT_16BIT|AF_FORMAT_NE),
-    AF_FORMAT_S32P      = (AF_INTP|AF_FORMAT_SI|AF_FORMAT_32BIT|AF_FORMAT_NE),
-    AF_FORMAT_FLOATP    = (AF_FLTP|AF_FORMAT_32BIT|AF_FORMAT_NE),
-    AF_FORMAT_DOUBLEP   = (AF_FLTP|AF_FORMAT_64BIT|AF_FORMAT_NE),
+    AF_FORMAT_U8P       = (AF_INTP|AF_FORMAT_US|AF_FORMAT_8BIT),
+    AF_FORMAT_S16P      = (AF_INTP|AF_FORMAT_SI|AF_FORMAT_16BIT),
+    AF_FORMAT_S32P      = (AF_INTP|AF_FORMAT_SI|AF_FORMAT_32BIT),
+    AF_FORMAT_FLOATP    = (AF_FLTP|AF_FORMAT_32BIT),
+    AF_FORMAT_DOUBLEP   = (AF_FLTP|AF_FORMAT_64BIT),
 
-    // Native endian variants
-    AF_FORMAT_U16       = AF_SELECT_LE_BE(AF_FORMAT_U16_LE, AF_FORMAT_U16_BE),
-    AF_FORMAT_S16       = AF_SELECT_LE_BE(AF_FORMAT_S16_LE, AF_FORMAT_S16_BE),
-    AF_FORMAT_U24       = AF_SELECT_LE_BE(AF_FORMAT_U24_LE, AF_FORMAT_U24_BE),
-    AF_FORMAT_S24       = AF_SELECT_LE_BE(AF_FORMAT_S24_LE, AF_FORMAT_S24_BE),
-    AF_FORMAT_U32       = AF_SELECT_LE_BE(AF_FORMAT_U32_LE, AF_FORMAT_U32_BE),
-    AF_FORMAT_S32       = AF_SELECT_LE_BE(AF_FORMAT_S32_LE, AF_FORMAT_S32_BE),
-
-    AF_FORMAT_FLOAT     = AF_SELECT_LE_BE(AF_FORMAT_FLOAT_LE, AF_FORMAT_FLOAT_BE),
-    AF_FORMAT_DOUBLE    = AF_SELECT_LE_BE(AF_FORMAT_DOUBLE_LE, AF_FORMAT_DOUBLE_BE),
-
-    AF_FORMAT_AC3       = AF_SELECT_LE_BE(AF_FORMAT_AC3_LE, AF_FORMAT_AC3_BE),
-    AF_FORMAT_IEC61937  = AF_SELECT_LE_BE(AF_FORMAT_IEC61937_LE, AF_FORMAT_IEC61937_BE),
+    // All of these use IEC61937 framing, and otherwise pretend to be like PCM.
+    AF_FORMAT_S_AAC     = AF_FORMAT_S_(0),
+    AF_FORMAT_S_AC3     = AF_FORMAT_S_(1),
+    AF_FORMAT_S_DTS     = AF_FORMAT_S_(2),
+    AF_FORMAT_S_DTSHD   = AF_FORMAT_S_(3),
+    AF_FORMAT_S_EAC3    = AF_FORMAT_S_(4),
+    AF_FORMAT_S_MP3     = AF_FORMAT_S_(5),
+    AF_FORMAT_S_TRUEHD  = AF_FORMAT_S_(6),
 };
 
-#define AF_FORMAT_IS_AC3(fmt) \
-    (((fmt) & AF_FORMAT_SPECIAL_MASK) == AF_FORMAT_S_AC3)
-
-#define AF_FORMAT_IS_IEC61937(fmt) \
-    (AF_FORMAT_IS_AC3(fmt) ||      \
-     ((fmt) & AF_FORMAT_SPECIAL_MASK) == AF_FORMAT_S_IEC61937)
-
-#define AF_FORMAT_IS_SPECIAL(fmt) \
-    ((fmt & AF_FORMAT_SPECIAL_MASK) != 0)
+#define AF_FORMAT_IS_IEC61937(f) (((f) & AF_FORMAT_TYPE_MASK) == AF_FORMAT_S)
+#define AF_FORMAT_IS_SPECIAL(f) AF_FORMAT_IS_IEC61937(f)
 
 struct af_fmt_entry {
     const char *name;

audio/mixer.c

@@ -37,6 +37,7 @@ struct mixer {
     struct af_stream *af;
     // Static, dependent on ao/softvol settings
     bool softvol;                       // use AO (true) or af_volume (false)
+    bool ao_softvol;                    // AO has private or per-app volume
     bool emulate_mute;                  // if true, emulate mute with volume=0
     // Last known values (possibly out of sync with reality)
     float vol_l, vol_r;
@@ -245,11 +246,13 @@ char *mixer_get_volume_restore_data(struct mixer *mixer)
 
 static void probe_softvol(struct mixer *mixer)
 {
+    mixer->ao_softvol =
+        ao_control(mixer->ao, AOCONTROL_HAS_SOFT_VOLUME, 0) == 1 ||
+        ao_control(mixer->ao, AOCONTROL_HAS_PER_APP_VOLUME, 0) == 1;
+
     if (mixer->opts->softvol == SOFTVOL_AUTO) {
         // No system-wide volume => fine with AO volume control.
-        mixer->softvol =
-            ao_control(mixer->ao, AOCONTROL_HAS_TEMP_VOLUME, 0) != 1 ||
-            ao_control(mixer->ao, AOCONTROL_HAS_PER_APP_VOLUME, 0) != 1;
+        mixer->softvol = !mixer->ao_softvol;
     } else {
         mixer->softvol = mixer->opts->softvol == SOFTVOL_YES;
     }
@@ -282,8 +285,7 @@ static void restore_volume(struct mixer *mixer)
     const char *prev_driver = mixer->driver;
     mixer->driver = mixer->softvol ? "softvol" : ao_get_name(ao);
 
-    bool restore =
-        mixer->softvol || ao_control(ao, AOCONTROL_HAS_TEMP_VOLUME, 0) == 1;
+    bool restore = mixer->softvol || mixer->ao_softvol;
 
     // Restore old parameters if volume won't survive reinitialization.
     // But not if volume scale is possibly different.
@@ -361,7 +363,7 @@ void mixer_uninit_audio(struct mixer *mixer)
         return;
 
     checkvolume(mixer);
-    if (mixer->muted_by_us) {
+    if (mixer->muted_by_us && !mixer->softvol && !mixer->ao_softvol) {
         /* Current audio output API combines playing the remaining buffered
          * audio and uninitializing the AO into one operation, even though
          * ideally unmute would happen between those two steps. We can't do

audio/out/ao.c

@@ -31,13 +31,13 @@
 
 #include "options/options.h"
 #include "options/m_config.h"
-#include "osdep/timer.h"
 #include "common/msg.h"
 #include "common/common.h"
 #include "common/global.h"
 
 extern const struct ao_driver audio_out_oss;
 extern const struct ao_driver audio_out_coreaudio;
+extern const struct ao_driver audio_out_coreaudio_exclusive;
 extern const struct ao_driver audio_out_rsound;
 extern const struct ao_driver audio_out_sndio;
 extern const struct ao_driver audio_out_pulse;
@@ -60,9 +60,6 @@ static const struct ao_driver * const audio_out_drivers[] = {
 #if HAVE_PULSE
     &audio_out_pulse,
 #endif
-#if HAVE_SNDIO
-    &audio_out_sndio,
-#endif
 #if HAVE_ALSA
     &audio_out_alsa,
 #endif
@@ -87,6 +84,9 @@ static const struct ao_driver * const audio_out_drivers[] = {
 #endif
 #if HAVE_SDL1 || HAVE_SDL2
     &audio_out_sdl,
+#endif
+#if HAVE_SNDIO
+    &audio_out_sndio,
 #endif
     &audio_out_null,
     // should not be auto-selected:
@@ -94,6 +94,9 @@ static const struct ao_driver * const audio_out_drivers[] = {
 #if HAVE_ENCODING
     &audio_out_lavc,
 #endif
+#if HAVE_COREAUDIO
+    &audio_out_coreaudio_exclusive,
+#endif
 #if HAVE_RSOUND
     &audio_out_rsound,
 #endif
@@ -149,6 +152,7 @@ static struct ao *ao_create(bool probing, struct mpv_global *global,
         .channels = channels,
         .format = format,
         .log = mp_log_new(ao, log, name),
+        .def_buffer = global->opts->audio_buffer,
     };
     if (ao->driver->encode != !!ao->encode_lavc_ctx)
         goto error;
@@ -184,7 +188,7 @@ static struct ao *ao_create(bool probing, struct mpv_global *global,
         ao->device_buffer = ao->driver->get_space(ao);
         MP_VERBOSE(ao, "device buffer: %d samples.\n", ao->device_buffer);
     }
-    ao->buffer = MPMAX(ao->device_buffer, MIN_BUFFER * ao->samplerate);
+    ao->buffer = MPMAX(ao->device_buffer, ao->def_buffer * ao->samplerate);
     MP_VERBOSE(ao, "using soft-buffer of %d samples.\n", ao->buffer);
 
     if (ao->api->init(ao) < 0)
@@ -255,16 +259,7 @@ int ao_play(struct ao *ao, void **data, int samples, int flags)
 
 int ao_control(struct ao *ao, enum aocontrol cmd, void *arg)
 {
-    switch (cmd) {
-    case AOCONTROL_HAS_TEMP_VOLUME:
-        return !ao->no_persistent_volume;
-    case AOCONTROL_HAS_PER_APP_VOLUME:
-        return !!ao->per_application_mixer;
-    default:
-        if (ao->api->control)
-            return ao->api->control(ao, cmd, arg);
-    }
-    return CONTROL_UNKNOWN;
+    return ao->api->control ? ao->api->control(ao, cmd, arg) : CONTROL_UNKNOWN;
 }
 
 // Return size of the buffered data in seconds. Can include the device latency.
@@ -309,27 +304,16 @@ void ao_resume(struct ao *ao)
         ao->api->resume(ao);
 }
 
-// Be careful with locking
-void ao_wait_drain(struct ao *ao)
-{
-    // This is probably not entirely accurate, but good enough.
-    mp_sleep_us(ao_get_delay(ao) * 1000000);
-    ao_reset(ao);
-}
-
 // Block until the current audio buffer has played completely.
 void ao_drain(struct ao *ao)
 {
-    if (ao->api->drain) {
+    if (ao->api->drain)
         ao->api->drain(ao);
-    } else {
-        ao_wait_drain(ao);
-    }
 }
 
 bool ao_eof_reached(struct ao *ao)
 {
-    return ao_get_delay(ao) < AO_EOF_DELAY;
+    return ao->api->get_eof ? ao->api->get_eof(ao) : true;
 }
 
 bool ao_chmap_sel_adjust(struct ao *ao, const struct mp_chmap_sel *s,

audio/out/ao.h

@@ -21,7 +21,7 @@
 
 #include <stdbool.h>
 
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 #include "common/common.h"
 #include "audio/chmap.h"
 #include "audio/chmap_sel.h"
@@ -36,7 +36,9 @@ enum aocontrol {
     AOCONTROL_SET_MUTE,
     // Has char* as argument, which contains the desired stream title.
     AOCONTROL_UPDATE_STREAM_TITLE,
-    AOCONTROL_HAS_TEMP_VOLUME,
+    // the AO does the equivalent of af_volume (return CONTROL_TRUE if yes)
+    AOCONTROL_HAS_SOFT_VOLUME,
+    // like above, but volume persists (per app), mpv won't restore volume
     AOCONTROL_HAS_PER_APP_VOLUME,
 };
 

audio/out/ao_alsa.c

@@ -31,7 +31,6 @@
 #include <sys/time.h>
 #include <stdlib.h>
 #include <stdarg.h>
-#include <ctype.h>
 #include <math.h>
 #include <string.h>
 
@@ -39,6 +38,7 @@
 #include "options/options.h"
 #include "options/m_option.h"
 #include "common/msg.h"
+#include "osdep/endian.h"
 
 #define ALSA_PCM_NEW_HW_PARAMS_API
 #define ALSA_PCM_NEW_SW_PARAMS_API
@@ -64,6 +64,7 @@ struct priv {
     char *cfg_mixer_name;
     int cfg_mixer_index;
     int cfg_resample;
+    int cfg_ni;
 };
 
 #define BUFFER_TIME 250000  // 250ms
@@ -99,7 +100,7 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
         long get_vol, set_vol;
         float f_multi;
 
-        if (AF_FORMAT_IS_IEC61937(ao->format))
+        if (AF_FORMAT_IS_SPECIAL(ao->format))
             return CONTROL_FALSE;
 
         //allocate simple id
@@ -208,25 +209,15 @@ alsa_error:
 static const int mp_to_alsa_format[][2] = {
     {AF_FORMAT_S8,          SND_PCM_FORMAT_S8},
     {AF_FORMAT_U8,          SND_PCM_FORMAT_U8},
-    {AF_FORMAT_U16_LE,      SND_PCM_FORMAT_U16_LE},
-    {AF_FORMAT_U16_BE,      SND_PCM_FORMAT_U16_BE},
-    {AF_FORMAT_S16_LE,      SND_PCM_FORMAT_S16_LE},
-    {AF_FORMAT_S16_BE,      SND_PCM_FORMAT_S16_BE},
-    {AF_FORMAT_U32_LE,      SND_PCM_FORMAT_U32_LE},
-    {AF_FORMAT_U32_BE,      SND_PCM_FORMAT_U32_BE},
-    {AF_FORMAT_S32_LE,      SND_PCM_FORMAT_S32_LE},
-    {AF_FORMAT_S32_BE,      SND_PCM_FORMAT_S32_BE},
-    {AF_FORMAT_U24_LE,      SND_PCM_FORMAT_U24_3LE},
-    {AF_FORMAT_U24_BE,      SND_PCM_FORMAT_U24_3BE},
-    {AF_FORMAT_S24_LE,      SND_PCM_FORMAT_S24_3LE},
-    {AF_FORMAT_S24_BE,      SND_PCM_FORMAT_S24_3BE},
-    {AF_FORMAT_FLOAT_LE,    SND_PCM_FORMAT_FLOAT_LE},
-    {AF_FORMAT_FLOAT_BE,    SND_PCM_FORMAT_FLOAT_BE},
-    {AF_FORMAT_AC3_LE,      SND_PCM_FORMAT_S16_LE},
-    {AF_FORMAT_AC3_BE,      SND_PCM_FORMAT_S16_BE},
-    {AF_FORMAT_IEC61937_LE, SND_PCM_FORMAT_S16_LE},
-    {AF_FORMAT_IEC61937_BE, SND_PCM_FORMAT_S16_BE},
-    {AF_FORMAT_MPEG2,       SND_PCM_FORMAT_MPEG},
+    {AF_FORMAT_U16,         SND_PCM_FORMAT_U16},
+    {AF_FORMAT_S16,         SND_PCM_FORMAT_S16},
+    {AF_FORMAT_U32,         SND_PCM_FORMAT_U32},
+    {AF_FORMAT_S32,         SND_PCM_FORMAT_S32},
+    {AF_FORMAT_U24,
+            MP_SELECT_LE_BE(SND_PCM_FORMAT_U24_3LE, SND_PCM_FORMAT_U24_3BE)},
+    {AF_FORMAT_S24,
+            MP_SELECT_LE_BE(SND_PCM_FORMAT_S24_3LE, SND_PCM_FORMAT_S24_3BE)},
+    {AF_FORMAT_FLOAT,       SND_PCM_FORMAT_FLOAT},
     {AF_FORMAT_UNKNOWN,     SND_PCM_FORMAT_UNKNOWN},
 };
 
@@ -353,6 +344,9 @@ static int init(struct ao *ao)
 
     struct priv *p = ao->priv;
 
+    if (!p->cfg_ni)
+        ao->format = af_fmt_from_planar(ao->format);
+
     /* switch for spdif
      * sets opening sequence for SPDIF
      * sets also the playback and other switches 'on the fly'
@@ -406,7 +400,15 @@ static int init(struct ao *ao)
     err = snd_pcm_hw_params_any(p->alsa, alsa_hwparams);
     CHECK_ALSA_ERROR("Unable to get initial parameters");
 
-    p->alsa_fmt = find_alsa_format(ao->format);
+    if (AF_FORMAT_IS_IEC61937(ao->format)) {
+        if (ao->format == AF_FORMAT_S_MP3) {
+            p->alsa_fmt = SND_PCM_FORMAT_MPEG;
+        } else {
+            p->alsa_fmt = SND_PCM_FORMAT_S16;
+        }
+    } else {
+        p->alsa_fmt = find_alsa_format(ao->format);
+    }
     if (p->alsa_fmt == SND_PCM_FORMAT_UNKNOWN) {
         p->alsa_fmt = SND_PCM_FORMAT_S16;
         ao->format = AF_FORMAT_S16;
@@ -414,15 +416,12 @@ static int init(struct ao *ao)
 
     err = snd_pcm_hw_params_test_format(p->alsa, alsa_hwparams, p->alsa_fmt);
     if (err < 0) {
+        if (AF_FORMAT_IS_IEC61937(ao->format))
+            CHECK_ALSA_ERROR("Unable to set IEC61937 format");
         MP_INFO(ao, "Format %s is not supported by hardware, trying default.\n",
                 af_fmt_to_str(ao->format));
-        p->alsa_fmt = SND_PCM_FORMAT_S16_LE;
-        if (AF_FORMAT_IS_AC3(ao->format))
-            ao->format = AF_FORMAT_AC3_LE;
-        else if (AF_FORMAT_IS_IEC61937(ao->format))
-            ao->format = AF_FORMAT_IEC61937_LE;
-        else
-            ao->format = AF_FORMAT_S16_LE;
+        p->alsa_fmt = SND_PCM_FORMAT_S16;
+        ao->format = AF_FORMAT_S16;
     }
 
     err = snd_pcm_hw_params_set_format(p->alsa, alsa_hwparams, p->alsa_fmt);
@@ -672,7 +671,7 @@ static int get_space(struct ao *ao)
     unsigned space = snd_pcm_status_get_avail(status);
     if (space > p->buffersize) // Buffer underrun?
         space = p->buffersize;
-    return space;
+    return space / p->outburst * p->outburst;
 
 alsa_error:
     return 0;
@@ -755,6 +754,7 @@ const struct ao_driver audio_out_alsa = {
         .cfg_mixer_device = "default",
         .cfg_mixer_name = "Master",
         .cfg_mixer_index = 0,
+        .cfg_ni = 0,
     },
     .options = (const struct m_option[]) {
         OPT_STRING("device", cfg_device, 0),
@@ -763,6 +763,7 @@ const struct ao_driver audio_out_alsa = {
         OPT_STRING("mixer-device", cfg_mixer_device, 0),
         OPT_STRING("mixer-name", cfg_mixer_name, 0),
         OPT_INTRANGE("mixer-index", cfg_mixer_index, 0, 0, 99),
+        OPT_FLAG("non-interleaved", cfg_ni, 0),
         {0}
     },
 };

audio/out/ao_coreaudio.c

@@ -1,39 +1,21 @@
 /*
- * CoreAudio audio output driver for Mac OS X
+ * This file is part of mpv.
  *
- * original copyright (C) Timothy J. Wood - Aug 2000
- * ported to MPlayer libao2 by Dan Christiansen
- *
- * Chris Roccati
- * Stefano Pigozzi
- *
- * The S/PDIF part of the code is based on the auhal audio output
- * module from VideoLAN:
- * Copyright (c) 2006 Derk-Jan Hartman <hartman at videolan dot org>
- *
- * This file is part of MPlayer.
- *
- * MPlayer is free software; you can redistribute it and/or modify
+ * mpv is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
  * the Free Software Foundation; either version 2 of the License, or
  * (at your option) any later version.
  *
- * MPlayer is distributed in the hope that it will be useful,
+ * mpv is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * GNU General Public License for more details.
  *
  * You should have received a copy of the GNU General Public License along
- * along with MPlayer; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ * with mpv.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-/*
- * The MacOS X CoreAudio framework doesn't mesh as simply as some
- * simpler frameworks do.  This is due to the fact that CoreAudio pulls
- * audio samples rather than having them pushed at it (which is nice
- * when you are wanting to do good buffering of audio).
- */
+#include <CoreAudio/HostTime.h>
 
 #include "config.h"
 #include "ao.h"
@@ -46,61 +28,64 @@
 #include "audio/out/ao_coreaudio_properties.h"
 #include "audio/out/ao_coreaudio_utils.h"
 
-static void audio_pause(struct ao *ao);
-static void audio_resume(struct ao *ao);
-static void reset(struct ao *ao);
-
-static void print_buffer(struct ao *ao, struct mp_ring *buffer)
-{
-    void *tctx = talloc_new(NULL);
-    MP_VERBOSE(ao, "%s\n", mp_ring_repr(buffer, tctx));
-    talloc_free(tctx);
-}
-
-struct priv_d {
-    // digital render callback
-    AudioDeviceIOProcID render_cb;
-
-    // pid set for hog mode, (-1) means that hog mode on the device was
-    // released. hog mode is exclusive access to a device
-    pid_t hog_pid;
-
-    // stream selected for digital playback by the detection in init
-    AudioStreamID stream;
-
-    // stream index in an AudioBufferList
-    int stream_idx;
-
-    // format we changed the stream to: for the digital case each application
-    // sets the stream format for a device to what it needs
-    AudioStreamBasicDescription stream_asbd;
-    AudioStreamBasicDescription original_asbd;
-
-    bool changed_mixing;
-    int stream_asbd_changed;
-    bool muted;
-};
-
 struct priv {
-    AudioDeviceID device;   // selected device
-    bool is_digital;        // running in digital mode?
+    AudioDeviceID device;
+    AudioUnit audio_unit;
 
-    AudioUnit audio_unit;   // AudioUnit for lpcm output
-
-    bool paused;
-
-    struct mp_ring *buffer;
-    struct priv_d *digital;
+    uint64_t hw_latency_us;
 
     // options
     int opt_device_id;
     int opt_list;
 };
 
-static int get_ring_size(struct ao *ao)
+bool ca_layout_to_mp_chmap(struct ao *ao, AudioChannelLayout *layout,
+                           struct mp_chmap *chmap);
+
+static int64_t ca_frames_to_us(struct ao *ao, uint32_t frames)
 {
-    return af_fmt_seconds_to_bytes(
-            ao->format, 0.5, ao->channels.num, ao->samplerate);
+    return frames / (float) ao->samplerate * 1e6;
+}
+
+static int64_t ca_get_hardware_latency(struct ao *ao) {
+    struct priv *p = ao->priv;
+
+    double audiounit_latency_sec = 0.0;
+    uint32_t size = sizeof(audiounit_latency_sec);
+    OSStatus err = AudioUnitGetProperty(
+            p->audio_unit,
+            kAudioUnitProperty_Latency,
+            kAudioUnitScope_Global,
+            0,
+            &audiounit_latency_sec,
+            &size);
+    CHECK_CA_ERROR("cannot get audio unit latency");
+
+    uint32_t frames = 0;
+    err = CA_GET_O(p->device, kAudioDevicePropertyLatency, &frames);
+    CHECK_CA_ERROR("cannot get device latency");
+
+    uint64_t audiounit_latency_us = audiounit_latency_sec * 1e6;
+    uint64_t device_latency_us    = ca_frames_to_us(ao, frames);
+
+    MP_VERBOSE(ao, "audiounit latency [us]: %lld\n", audiounit_latency_us);
+    MP_VERBOSE(ao, "device latency [us]: %lld\n", device_latency_us);
+
+    return audiounit_latency_us + device_latency_us;
+
+coreaudio_error:
+    return 0;
+}
+
+static int64_t ca_get_latency(const AudioTimeStamp *ts)
+{
+    uint64_t out = AudioConvertHostTimeToNanos(ts->mHostTime);
+    uint64_t now = AudioConvertHostTimeToNanos(AudioGetCurrentHostTime());
+
+    if (now > out)
+        return 0;
+
+    return (out - now) * 1e-3;
 }
 
 static OSStatus render_cb_lpcm(void *ctx, AudioUnitRenderActionFlags *aflags,
@@ -109,259 +94,126 @@ static OSStatus render_cb_lpcm(void *ctx, AudioUnitRenderActionFlags *aflags,
 {
     struct ao *ao   = ctx;
     struct priv *p  = ao->priv;
-
     AudioBuffer buf = buffer_list->mBuffers[0];
-    int requested   = buf.mDataByteSize;
 
-    if (mp_ring_buffered(p->buffer) < requested) {
-        MP_VERBOSE(ao, "buffer underrun\n");
-        audio_pause(ao);
-        memset(buf.mData, 0, requested);
-    } else {
-        mp_ring_read(p->buffer, buf.mData, requested);
-    }
+    int64_t end = mp_time_us();
+    end += p->hw_latency_us + ca_get_latency(ts) + ca_frames_to_us(ao, frames);
 
+    ao_read_data(ao, &buf.mData, frames, end);
     return noErr;
 }
 
-static OSStatus render_cb_digital(
-        AudioDeviceID device, const AudioTimeStamp *ts,
-        const void *in_data, const AudioTimeStamp *in_ts,
-        AudioBufferList *out_data, const AudioTimeStamp *out_ts, void *ctx)
-{
-    struct ao *ao    = ctx;
-    struct priv *p   = ao->priv;
-    struct priv_d *d = p->digital;
-    AudioBuffer buf  = out_data->mBuffers[d->stream_idx];
-    int requested    = buf.mDataByteSize;
+static int get_volume(struct ao *ao, struct ao_control_vol *vol) {
+    struct priv *p = ao->priv;
+    float auvol;
+    OSStatus err =
+        AudioUnitGetParameter(p->audio_unit, kHALOutputParam_Volume,
+                              kAudioUnitScope_Global, 0, &auvol);
 
-    if (d->muted)
-        mp_ring_drain(p->buffer, requested);
-    else
-        mp_ring_read(p->buffer, buf.mData, requested);
+    CHECK_CA_ERROR("could not get HAL output volume");
+    vol->left = vol->right = auvol * 100.0;
+    return CONTROL_TRUE;
+coreaudio_error:
+    return CONTROL_ERROR;
+}
 
-    return noErr;
+static int set_volume(struct ao *ao, struct ao_control_vol *vol) {
+    struct priv *p = ao->priv;
+    float auvol = (vol->left + vol->right) / 200.0;
+    OSStatus err =
+        AudioUnitSetParameter(p->audio_unit, kHALOutputParam_Volume,
+                              kAudioUnitScope_Global, 0, auvol, 0);
+    CHECK_CA_ERROR("could not set HAL output volume");
+    return CONTROL_TRUE;
+coreaudio_error:
+    return CONTROL_ERROR;
 }
 
 static int control(struct ao *ao, enum aocontrol cmd, void *arg)
 {
-    struct priv *p = ao->priv;
-    ao_control_vol_t *control_vol;
-    OSStatus err;
-    Float32 vol;
     switch (cmd) {
     case AOCONTROL_GET_VOLUME:
-        control_vol = (ao_control_vol_t *)arg;
-        if (p->is_digital) {
-            struct priv_d *d = p->digital;
-            // Digital output has no volume adjust.
-            int digitalvol = d->muted ? 0 : 100;
-            *control_vol = (ao_control_vol_t) {
-                .left = digitalvol, .right = digitalvol,
-            };
-            return CONTROL_TRUE;
-        }
-
-        err = AudioUnitGetParameter(p->audio_unit, kHALOutputParam_Volume,
-                                    kAudioUnitScope_Global, 0, &vol);
-
-        CHECK_CA_ERROR("could not get HAL output volume");
-        control_vol->left = control_vol->right = vol * 100.0;
-        return CONTROL_TRUE;
-
+        return get_volume(ao, arg);
     case AOCONTROL_SET_VOLUME:
-        control_vol = (ao_control_vol_t *)arg;
-
-        if (p->is_digital) {
-            struct priv_d *d = p->digital;
-            // Digital output can not set volume. Here we have to return true
-            // to make mixer forget it. Else mixer will add a soft filter,
-            // that's not we expected and the filter not support ac3 stream
-            // will cause mplayer die.
-
-            // Although not support set volume, but at least we support mute.
-            // MPlayer set mute by set volume to zero, we handle it.
-            if (control_vol->left == 0 && control_vol->right == 0)
-                d->muted = true;
-            else
-                d->muted = false;
-            return CONTROL_TRUE;
-        }
-
-        vol = (control_vol->left + control_vol->right) / 200.0;
-        err = AudioUnitSetParameter(p->audio_unit, kHALOutputParam_Volume,
-                                    kAudioUnitScope_Global, 0, vol, 0);
-
-        CHECK_CA_ERROR("could not set HAL output volume");
+        return set_volume(ao, arg);
+    case AOCONTROL_HAS_SOFT_VOLUME:
+        return CONTROL_TRUE;
+    case AOCONTROL_HAS_PER_APP_VOLUME:
         return CONTROL_TRUE;
-
-    } // end switch
-    return CONTROL_UNKNOWN;
-
-coreaudio_error:
-    return CONTROL_ERROR;
-}
-
-static void print_list(struct ao *ao)
-{
-    char *help = talloc_strdup(NULL, "Available output devices:\n");
-
-    AudioDeviceID *devs;
-    size_t n_devs;
-
-    OSStatus err =
-        CA_GET_ARY(kAudioObjectSystemObject, kAudioHardwarePropertyDevices,
-                   &devs, &n_devs);
-
-    CHECK_CA_ERROR("Failed to get list of output devices.");
-
-    for (int i = 0; i < n_devs; i++) {
-        char *name;
-        err = CA_GET_STR(devs[i], kAudioObjectPropertyName, &name);
-
-        if (err == noErr)
-            talloc_steal(devs, name);
-        else
-            name = "Unknown";
-
-        help = talloc_asprintf_append(
-                help, "  * %s (id: %" PRIu32 ")\n", name, devs[i]);
     }
-
-    talloc_free(devs);
-
-coreaudio_error:
-    MP_INFO(ao, "%s", help);
-    talloc_free(help);
+    return CONTROL_UNKNOWN;
 }
 
-static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd);
-static int init_digital(struct ao *ao, AudioStreamBasicDescription asbd);
+static bool init_chmap(struct ao *ao);
+static bool init_audiounit(struct ao *ao, AudioStreamBasicDescription asbd);
 
 static int init(struct ao *ao)
 {
-    OSStatus err;
-    struct priv *p   = ao->priv;
+    struct priv *p = ao->priv;
 
-    if (p->opt_list) print_list(ao);
+    if (p->opt_list) ca_print_device_list(ao);
 
-    struct priv_d *d = talloc_zero(p, struct priv_d);
+    OSStatus err = ca_select_device(ao, p->opt_device_id, &p->device);
+    CHECK_CA_ERROR("failed to select device");
 
-    *d = (struct priv_d) {
-        .muted = false,
-        .stream_asbd_changed = 0,
-        .hog_pid = -1,
-        .stream = 0,
-        .stream_idx = -1,
-        .changed_mixing = false,
-    };
-
-    p->digital = d;
-
-    ao->per_application_mixer = true;
-    ao->no_persistent_volume  = true;
-
-    AudioDeviceID selected_device = 0;
-    if (p->opt_device_id < 0) {
-        // device not set by user, get the default one
-        err = CA_GET(kAudioObjectSystemObject,
-                     kAudioHardwarePropertyDefaultOutputDevice,
-                     &selected_device);
-        CHECK_CA_ERROR("could not get default audio device");
-    } else {
-        selected_device = p->opt_device_id;
-    }
-
-    if (mp_msg_test(ao->log, MSGL_V)) {
-        char *name;
-        err = CA_GET_STR(selected_device, kAudioObjectPropertyName, &name);
-        CHECK_CA_ERROR("could not get selected audio device name");
-
-        MP_VERBOSE(ao, "selected audio output device: %s (%" PRIu32 ")\n",
-                       name, selected_device);
-
-        talloc_free(name);
-    }
-
-    // Save selected device id
-    p->device = selected_device;
+    if (!init_chmap(ao))
+        goto coreaudio_error;
 
     ao->format = af_fmt_from_planar(ao->format);
 
-    bool supports_digital = false;
-    /* Probe whether device support S/PDIF stream output if input is AC3 stream. */
-    if (AF_FORMAT_IS_AC3(ao->format)) {
-        if (ca_device_supports_digital(ao, selected_device))
-            supports_digital = true;
-    }
-
-    if (!supports_digital) {
-        AudioChannelLayout *layouts;
-        size_t n_layouts;
-        err = CA_GET_ARY_O(selected_device,
-                           kAudioDevicePropertyPreferredChannelLayout,
-                           &layouts, &n_layouts);
-        CHECK_CA_ERROR("could not get audio device prefered layouts");
-
-        struct mp_chmap_sel chmap_sel = {0};
-        for (int i = 0; i < n_layouts; i++) {
-            struct mp_chmap chmap = {0};
-            if (ca_layout_to_mp_chmap(ao, &layouts[i], &chmap))
-                mp_chmap_sel_add_map(&chmap_sel, &chmap);
-        }
-
-        talloc_free(layouts);
-
-        if (ao->channels.num < 3) {
-            struct mp_chmap chmap;
-            mp_chmap_from_channels(&chmap, ao->channels.num);
-            mp_chmap_sel_add_map(&chmap_sel, &chmap);
-        }
-
-        if (!ao_chmap_sel_adjust(ao, &chmap_sel, &ao->channels)) {
-            MP_ERR(ao, "could not select a suitable channel map among the "
-                       "hardware supported ones. Make sure to configure your "
-                       "output device correctly in 'Audio MIDI Setup.app'\n");
-            goto coreaudio_error;
-        }
-
-    } // closes if (!supports_digital)
-
-    // Build ASBD for the input format
     AudioStreamBasicDescription asbd;
-    asbd.mSampleRate       = ao->samplerate;
-    asbd.mFormatID         = supports_digital ?
-                             kAudioFormat60958AC3 : kAudioFormatLinearPCM;
-    asbd.mChannelsPerFrame = ao->channels.num;
-    asbd.mBitsPerChannel   = af_fmt2bits(ao->format);
-    asbd.mFormatFlags      = kAudioFormatFlagIsPacked;
+    ca_fill_asbd(ao, &asbd);
 
-    if ((ao->format & AF_FORMAT_POINT_MASK) == AF_FORMAT_F)
-        asbd.mFormatFlags |= kAudioFormatFlagIsFloat;
+    if (!init_audiounit(ao, asbd))
+        goto coreaudio_error;
 
-    if ((ao->format & AF_FORMAT_SIGN_MASK) == AF_FORMAT_SI)
-        asbd.mFormatFlags |= kAudioFormatFlagIsSignedInteger;
-
-    if ((ao->format & AF_FORMAT_END_MASK) == AF_FORMAT_BE)
-        asbd.mFormatFlags |= kAudioFormatFlagIsBigEndian;
-
-    asbd.mFramesPerPacket = 1;
-    asbd.mBytesPerPacket = asbd.mBytesPerFrame =
-        asbd.mFramesPerPacket * asbd.mChannelsPerFrame *
-        (asbd.mBitsPerChannel / 8);
-
-    ca_print_asbd(ao, "source format:", &asbd);
-
-    if (supports_digital)
-        return init_digital(ao, asbd);
-    else
-        return init_lpcm(ao, asbd);
+    return CONTROL_OK;
 
 coreaudio_error:
     return CONTROL_ERROR;
 }
 
-static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd)
+static bool init_chmap(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    OSStatus err;
+    AudioChannelLayout *layouts;
+    size_t n_layouts;
+
+    err = CA_GET_ARY_O(p->device,
+                       kAudioDevicePropertyPreferredChannelLayout,
+                       &layouts, &n_layouts);
+    CHECK_CA_ERROR("could not get audio device prefered layouts");
+
+    struct mp_chmap_sel chmap_sel = {0};
+    for (int i = 0; i < n_layouts; i++) {
+        struct mp_chmap chmap = {0};
+        if (ca_layout_to_mp_chmap(ao, &layouts[i], &chmap))
+            mp_chmap_sel_add_map(&chmap_sel, &chmap);
+    }
+
+    talloc_free(layouts);
+
+    if (ao->channels.num < 3) {
+        struct mp_chmap chmap;
+        mp_chmap_from_channels(&chmap, ao->channels.num);
+        mp_chmap_sel_add_map(&chmap_sel, &chmap);
+    }
+
+    if (!ao_chmap_sel_adjust(ao, &chmap_sel, &ao->channels)) {
+        MP_ERR(ao, "could not select a suitable channel map among the "
+                   "hardware supported ones. Make sure to configure your "
+                   "output device correctly in 'Audio MIDI Setup.app'\n");
+        goto coreaudio_error;
+    }
+
+    return true;
+
+coreaudio_error:
+    return false;
+}
+
+static bool init_audiounit(struct ao *ao, AudioStreamBasicDescription asbd)
 {
     OSStatus err;
     uint32_t size;
@@ -386,7 +238,6 @@ static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd)
     err = AudioComponentInstanceNew(comp, &(p->audio_unit));
     CHECK_CA_ERROR("unable to open audio component");
 
-    // Initialize AudioUnit
     err = AudioUnitInitialize(p->audio_unit);
     CHECK_CA_ERROR_L(coreaudio_error_component,
                      "unable to initialize audio unit");
@@ -399,7 +250,6 @@ static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd)
     CHECK_CA_ERROR_L(coreaudio_error_audiounit,
                      "unable to set the input format on the audio unit");
 
-    //Set the Current Device to the Default Output Unit.
     err = AudioUnitSetProperty(p->audio_unit,
                                kAudioOutputUnitProperty_CurrentDevice,
                                kAudioUnitScope_Global, 0, &p->device,
@@ -407,8 +257,7 @@ static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd)
     CHECK_CA_ERROR_L(coreaudio_error_audiounit,
                      "can't link audio unit to selected device");
 
-    p->buffer = mp_ring_new(p, get_ring_size(ao));
-    print_buffer(ao, p->buffer);
+    p->hw_latency_us = ca_get_hardware_latency(ao);
 
     AURenderCallbackStruct render_cb = (AURenderCallbackStruct) {
         .inputProc       = render_cb_lpcm,
@@ -423,279 +272,174 @@ static int init_lpcm(struct ao *ao, AudioStreamBasicDescription asbd)
     CHECK_CA_ERROR_L(coreaudio_error_audiounit,
                      "unable to set render callback on audio unit");
 
-    reset(ao);
-    return CONTROL_OK;
+    return true;
 
 coreaudio_error_audiounit:
     AudioUnitUninitialize(p->audio_unit);
 coreaudio_error_component:
     AudioComponentInstanceDispose(p->audio_unit);
 coreaudio_error:
-    return CONTROL_ERROR;
+    return false;
 }
 
-static int init_digital(struct ao *ao, AudioStreamBasicDescription asbd)
+static void stop(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    struct priv_d *d = p->digital;
-    OSStatus err = noErr;
-
-    uint32_t is_alive = 1;
-    err = CA_GET(p->device, kAudioDevicePropertyDeviceIsAlive, &is_alive);
-    CHECK_CA_WARN("could not check whether device is alive");
-
-    if (!is_alive)
-        MP_WARN(ao , "device is not alive\n");
-
-    p->is_digital = 1;
-
-    err = ca_lock_device(p->device, &d->hog_pid);
-    CHECK_CA_WARN("failed to set hogmode");
-
-    err = ca_disable_mixing(ao, p->device, &d->changed_mixing);
-    CHECK_CA_WARN("failed to disable mixing");
-
-    AudioStreamID *streams;
-    size_t n_streams;
-
-    /* Get a list of all the streams on this device. */
-    err = CA_GET_ARY_O(p->device, kAudioDevicePropertyStreams,
-                       &streams, &n_streams);
-
-    CHECK_CA_ERROR("could not get number of streams");
-
-    for (int i = 0; i < n_streams && d->stream_idx < 0; i++) {
-        bool digital = ca_stream_supports_digital(ao, streams[i]);
-
-        if (digital) {
-            err = CA_GET(streams[i], kAudioStreamPropertyPhysicalFormat,
-                         &d->original_asbd);
-            if (!CHECK_CA_WARN("could not get stream's physical format to "
-                               "revert to, getting the next one"))
-                continue;
-
-            AudioStreamRangedDescription *formats;
-            size_t n_formats;
-
-            err = CA_GET_ARY(streams[i],
-                             kAudioStreamPropertyAvailablePhysicalFormats,
-                             &formats, &n_formats);
-
-            if (!CHECK_CA_WARN("could not get number of stream formats"))
-                continue; // try next one
-
-            int req_rate_format = -1;
-            int max_rate_format = -1;
-
-            d->stream = streams[i];
-            d->stream_idx = i;
-
-            for (int j = 0; j < n_formats; j++)
-                if (ca_format_is_digital(formats[j].mFormat)) {
-                    // select the digital format that has exactly the same
-                    // samplerate. If an exact match cannot be found, select
-                    // the format with highest samplerate as backup.
-                    if (formats[j].mFormat.mSampleRate == asbd.mSampleRate) {
-                        req_rate_format = j;
-                        break;
-                    } else if (max_rate_format < 0 ||
-                        formats[j].mFormat.mSampleRate >
-                        formats[max_rate_format].mFormat.mSampleRate)
-                        max_rate_format = j;
-                }
-
-            if (req_rate_format >= 0)
-                d->stream_asbd = formats[req_rate_format].mFormat;
-            else
-                d->stream_asbd = formats[max_rate_format].mFormat;
-
-            talloc_free(formats);
-        }
-    }
-
-    talloc_free(streams);
-
-    if (d->stream_idx < 0) {
-        MP_WARN(ao , "can't find any digital output stream format\n");
-        goto coreaudio_error;
-    }
-
-    if (!ca_change_format(ao, d->stream, d->stream_asbd))
-        goto coreaudio_error;
-
-    void *changed = (void *) &(d->stream_asbd_changed);
-    err = ca_enable_device_listener(p->device, changed);
-    CHECK_CA_ERROR("cannot install format change listener during init");
-
-#if BYTE_ORDER == BIG_ENDIAN
-    if (!(p->stream_asdb.mFormatFlags & kAudioFormatFlagIsBigEndian))
-#else
-    /* tell mplayer that we need a byteswap on AC3 streams, */
-    if (d->stream_asbd.mFormatID & kAudioFormat60958AC3)
-        ao->format = AF_FORMAT_AC3_LE;
-    else if (d->stream_asbd.mFormatFlags & kAudioFormatFlagIsBigEndian)
-#endif
-        MP_WARN(ao, "stream has non-native byte order, output may fail\n");
-
-    ao->samplerate = d->stream_asbd.mSampleRate;
-    ao->bps = ao->samplerate *
-                  (d->stream_asbd.mBytesPerPacket /
-                   d->stream_asbd.mFramesPerPacket);
-
-    p->buffer = mp_ring_new(p, get_ring_size(ao));
-    print_buffer(ao, p->buffer);
-
-    err = AudioDeviceCreateIOProcID(p->device,
-                                    (AudioDeviceIOProc)render_cb_digital,
-                                    (void *)ao,
-                                    &d->render_cb);
-
-    CHECK_CA_ERROR("failed to register digital render callback");
-
-    reset(ao);
-
-    return CONTROL_TRUE;
-
-coreaudio_error:
-    err = ca_unlock_device(p->device, &d->hog_pid);
-    CHECK_CA_WARN("can't release hog mode");
-    return CONTROL_ERROR;
+    OSStatus err = AudioOutputUnitStop(p->audio_unit);
+    CHECK_CA_WARN("can't stop audio unit");
 }
 
-static int play(struct ao *ao, void **data, int samples, int flags)
-{
-    struct priv *p   = ao->priv;
-    struct priv_d *d = p->digital;
-    void *output_samples = data[0];
-    int num_bytes = samples * ao->sstride;
-
-    // Check whether we need to reset the digital output stream.
-    if (p->is_digital && d->stream_asbd_changed) {
-        d->stream_asbd_changed = 0;
-        if (ca_stream_supports_digital(ao, d->stream)) {
-            if (!ca_change_format(ao, d->stream, d->stream_asbd)) {
-                MP_WARN(ao , "can't restore digital output\n");
-            } else {
-                MP_WARN(ao, "restoring digital output succeeded.\n");
-                reset(ao);
-            }
-        }
-    }
-
-    int wrote = mp_ring_write(p->buffer, output_samples, num_bytes);
-    audio_resume(ao);
-
-    return wrote / ao->sstride;
-}
-
-static void reset(struct ao *ao)
+static void start(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    audio_pause(ao);
-    mp_ring_reset(p->buffer);
+    OSStatus err = AudioOutputUnitStart(p->audio_unit);
+    CHECK_CA_WARN("can't start audio unit");
 }
 
-static int get_space(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    return mp_ring_available(p->buffer) / ao->sstride;
-}
-
-static float get_delay(struct ao *ao)
-{
-    // FIXME: should also report the delay of coreaudio itself (hardware +
-    // internal buffers)
-    struct priv *p = ao->priv;
-    return mp_ring_buffered(p->buffer) / (float)ao->bps;
-}
 
 static void uninit(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    OSStatus err = noErr;
+    AudioOutputUnitStop(p->audio_unit);
+    AudioUnitUninitialize(p->audio_unit);
+    AudioComponentInstanceDispose(p->audio_unit);
+}
 
-    if (!p->is_digital) {
-        AudioOutputUnitStop(p->audio_unit);
-        AudioUnitUninitialize(p->audio_unit);
-        AudioComponentInstanceDispose(p->audio_unit);
-    } else {
-        struct priv_d *d = p->digital;
+// Channel Mapping functions
+static const int speaker_map[][2] = {
+    { kAudioChannelLabel_Left,                 MP_SPEAKER_ID_FL   },
+    { kAudioChannelLabel_Right,                MP_SPEAKER_ID_FR   },
+    { kAudioChannelLabel_Center,               MP_SPEAKER_ID_FC   },
+    { kAudioChannelLabel_LFEScreen,            MP_SPEAKER_ID_LFE  },
+    { kAudioChannelLabel_LeftSurround,         MP_SPEAKER_ID_BL   },
+    { kAudioChannelLabel_RightSurround,        MP_SPEAKER_ID_BR   },
+    { kAudioChannelLabel_LeftCenter,           MP_SPEAKER_ID_FLC  },
+    { kAudioChannelLabel_RightCenter,          MP_SPEAKER_ID_FRC  },
+    { kAudioChannelLabel_CenterSurround,       MP_SPEAKER_ID_BC   },
+    { kAudioChannelLabel_LeftSurroundDirect,   MP_SPEAKER_ID_SL   },
+    { kAudioChannelLabel_RightSurroundDirect,  MP_SPEAKER_ID_SR   },
+    { kAudioChannelLabel_TopCenterSurround,    MP_SPEAKER_ID_TC   },
+    { kAudioChannelLabel_VerticalHeightLeft,   MP_SPEAKER_ID_TFL  },
+    { kAudioChannelLabel_VerticalHeightCenter, MP_SPEAKER_ID_TFC  },
+    { kAudioChannelLabel_VerticalHeightRight,  MP_SPEAKER_ID_TFR  },
+    { kAudioChannelLabel_TopBackLeft,          MP_SPEAKER_ID_TBL  },
+    { kAudioChannelLabel_TopBackCenter,        MP_SPEAKER_ID_TBC  },
+    { kAudioChannelLabel_TopBackRight,         MP_SPEAKER_ID_TBR  },
 
-        void *changed = (void *) &(d->stream_asbd_changed);
-        err = ca_disable_device_listener(p->device, changed);
-        CHECK_CA_WARN("can't remove device listener, this may cause a crash");
+    // unofficial extensions
+    { kAudioChannelLabel_RearSurroundLeft,     MP_SPEAKER_ID_SDL  },
+    { kAudioChannelLabel_RearSurroundRight,    MP_SPEAKER_ID_SDR  },
+    { kAudioChannelLabel_LeftWide,             MP_SPEAKER_ID_WL   },
+    { kAudioChannelLabel_RightWide,            MP_SPEAKER_ID_WR   },
+    { kAudioChannelLabel_LFE2,                 MP_SPEAKER_ID_LFE2 },
 
-        err = AudioDeviceStop(p->device, d->render_cb);
-        CHECK_CA_WARN("failed to stop audio device");
+    { kAudioChannelLabel_HeadphonesLeft,       MP_SPEAKER_ID_DL   },
+    { kAudioChannelLabel_HeadphonesRight,      MP_SPEAKER_ID_DR   },
 
-        err = AudioDeviceDestroyIOProcID(p->device, d->render_cb);
-        CHECK_CA_WARN("failed to remove device render callback");
+    { kAudioChannelLabel_Unknown,              -1 },
+};
 
-        if (!ca_change_format(ao, d->stream, d->original_asbd))
-            MP_WARN(ao, "can't revert to original device format");
+static int ca_label_to_mp_speaker_id(AudioChannelLabel label)
+{
+    for (int i = 0; speaker_map[i][1] >= 0; i++)
+        if (speaker_map[i][0] == label)
+            return speaker_map[i][1];
+    return -1;
+}
 
-        err = ca_enable_mixing(ao, p->device, d->changed_mixing);
-        CHECK_CA_WARN("can't re-enable mixing");
+static void ca_log_layout(struct ao *ao, AudioChannelLayout *layout)
+{
+    if (!mp_msg_test(ao->log, MSGL_V))
+        return;
 
-        err = ca_unlock_device(p->device, &d->hog_pid);
-        CHECK_CA_WARN("can't release hog mode");
+    AudioChannelDescription *descs = layout->mChannelDescriptions;
+
+    MP_VERBOSE(ao, "layout: tag: <%d>, bitmap: <%d>, "
+                   "descriptions <%d>\n",
+                   layout->mChannelLayoutTag,
+                   layout->mChannelBitmap,
+                   layout->mNumberChannelDescriptions);
+
+    for (int i = 0; i < layout->mNumberChannelDescriptions; i++) {
+        AudioChannelDescription d = descs[i];
+        MP_VERBOSE(ao, " - description %d: label <%d, %d>, flags: <%u>, "
+                       "coords: <%f, %f, %f>\n", i,
+                       d.mChannelLabel,
+                       ca_label_to_mp_speaker_id(d.mChannelLabel),
+                       d.mChannelFlags,
+                       d.mCoordinates[0],
+                       d.mCoordinates[1],
+                       d.mCoordinates[2]);
     }
 }
 
-static void audio_pause(struct ao *ao)
+bool ca_layout_to_mp_chmap(struct ao *ao, AudioChannelLayout *layout,
+                           struct mp_chmap *chmap)
 {
-    struct priv *p = ao->priv;
-    OSErr err = noErr;
+    AudioChannelLayoutTag tag  = layout->mChannelLayoutTag;
+    uint32_t layout_size       = sizeof(layout);
+    OSStatus err;
 
-    if (p->paused)
-        return;
-
-    if (!p->is_digital) {
-        err = AudioOutputUnitStop(p->audio_unit);
-        CHECK_CA_WARN("can't stop audio unit");
-    } else {
-        struct priv_d *d = p->digital;
-        err = AudioDeviceStop(p->device, d->render_cb);
-        CHECK_CA_WARN("can't stop digital device");
+    if (tag == kAudioChannelLayoutTag_UseChannelBitmap) {
+        err = AudioFormatGetProperty(kAudioFormatProperty_ChannelLayoutForBitmap,
+                                     sizeof(uint32_t),
+                                     &layout->mChannelBitmap,
+                                     &layout_size,
+                                     layout);
+        CHECK_CA_ERROR("failed to convert channel bitmap to descriptions");
+    } else if (tag != kAudioChannelLayoutTag_UseChannelDescriptions) {
+        err = AudioFormatGetProperty(kAudioFormatProperty_ChannelLayoutForTag,
+                                     sizeof(AudioChannelLayoutTag),
+                                     &layout->mChannelLayoutTag,
+                                     &layout_size,
+                                     layout);
+        CHECK_CA_ERROR("failed to convert channel tag to descriptions");
     }
 
-    p->paused = true;
-}
+    ca_log_layout(ao, layout);
 
-static void audio_resume(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    OSErr err = noErr;
+    // If the channel layout uses channel descriptions, from my
+    // experiments there are there three possibile cases:
+    // * The description has a label kAudioChannelLabel_Unknown:
+    //   Can't do anything about this (looks like non surround
+    //   layouts are like this).
+    // * The description uses positional information: this in
+    //   theory could be used but one would have to map spatial
+    //   positions to labels which is not really feasible.
+    // * The description has a well known label which can be mapped
+    //   to the waveextensible definition: this is the kind of
+    //   descriptions we process here.
 
-    if (!p->paused)
-        return;
-
-    if (!p->is_digital) {
-        err = AudioOutputUnitStart(p->audio_unit);
-        CHECK_CA_WARN("can't start audio unit");
-    } else {
-        struct priv_d *d = p->digital;
-        err = AudioDeviceStart(p->device, d->render_cb);
-        CHECK_CA_WARN("can't start digital device");
+    for (int n = 0; n < layout->mNumberChannelDescriptions; n++) {
+        AudioChannelLabel label = layout->mChannelDescriptions[n].mChannelLabel;
+        uint8_t speaker = ca_label_to_mp_speaker_id(label);
+        if (label == kAudioChannelLabel_Unknown)
+            continue;
+        if (speaker < 0) {
+            MP_VERBOSE(ao, "channel label=%d unusable to build channel "
+                           "bitmap, skipping layout\n", label);
+        } else {
+            chmap->speaker[n] = speaker;
+            chmap->num = n + 1;
+        }
     }
 
-    p->paused = false;
+    return chmap->num > 0;
+coreaudio_error:
+    ca_log_layout(ao, layout);
+    return false;
 }
 
 #define OPT_BASE_STRUCT struct priv
 
 const struct ao_driver audio_out_coreaudio = {
-    .description = "CoreAudio (OS X Audio Output)",
+    .description = "CoreAudio AudioUnit",
     .name      = "coreaudio",
     .uninit    = uninit,
     .init      = init,
-    .play      = play,
     .control   = control,
-    .get_space = get_space,
-    .get_delay = get_delay,
-    .reset     = reset,
-    .pause     = audio_pause,
-    .resume    = audio_resume,
+    .pause     = stop,
+    .resume    = start,
     .priv_size = sizeof(struct priv),
     .options = (const struct m_option[]) {
         OPT_INT("device_id", opt_device_id, 0, OPTDEF_INT(-1)),

audio/out/ao_coreaudio_device.c

@@ -0,0 +1,677 @@
+/*
+ * CoreAudio audio output driver for Mac OS X
+ *
+ * original copyright (C) Timothy J. Wood - Aug 2000
+ * ported to MPlayer libao2 by Dan Christiansen
+ *
+ * Chris Roccati
+ * Stefano Pigozzi
+ *
+ * The S/PDIF part of the code is based on the auhal audio output
+ * module from VideoLAN:
+ * Copyright (c) 2006 Derk-Jan Hartman <hartman at videolan dot org>
+ *
+ * This file is part of MPlayer.
+ *
+ * MPlayer is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * MPlayer is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * along with MPlayer; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * The MacOS X CoreAudio framework doesn't mesh as simply as some
+ * simpler frameworks do.  This is due to the fact that CoreAudio pulls
+ * audio samples rather than having them pushed at it (which is nice
+ * when you are wanting to do good buffering of audio).
+ */
+
+#include "config.h"
+#include "ao.h"
+#include "internal.h"
+#include "audio/format.h"
+#include "osdep/timer.h"
+#include "options/m_option.h"
+#include "misc/ring.h"
+#include "common/msg.h"
+#include "audio/out/ao_coreaudio_properties.h"
+#include "audio/out/ao_coreaudio_utils.h"
+
+static void audio_pause(struct ao *ao);
+static void audio_resume(struct ao *ao);
+static void reset(struct ao *ao);
+
+static bool ca_format_is_digital(AudioStreamBasicDescription asbd)
+{
+    switch (asbd.mFormatID)
+    case 'IAC3':
+    case 'iac3':
+    case  kAudioFormat60958AC3:
+    case  kAudioFormatAC3:
+        return true;
+    return false;
+}
+
+static bool ca_stream_supports_digital(struct ao *ao, AudioStreamID stream)
+{
+    AudioStreamRangedDescription *formats = NULL;
+    size_t n_formats;
+
+    OSStatus err =
+        CA_GET_ARY(stream, kAudioStreamPropertyAvailablePhysicalFormats,
+                   &formats, &n_formats);
+
+    CHECK_CA_ERROR("Could not get number of stream formats.");
+
+    for (int i = 0; i < n_formats; i++) {
+        AudioStreamBasicDescription asbd = formats[i].mFormat;
+        ca_print_asbd(ao, "supported format:", &(asbd));
+        if (ca_format_is_digital(asbd)) {
+            talloc_free(formats);
+            return true;
+        }
+    }
+
+    talloc_free(formats);
+coreaudio_error:
+    return false;
+}
+
+static bool ca_device_supports_digital(struct ao *ao, AudioDeviceID device)
+{
+    AudioStreamID *streams = NULL;
+    size_t n_streams;
+
+    /* Retrieve all the output streams. */
+    OSStatus err =
+        CA_GET_ARY_O(device, kAudioDevicePropertyStreams, &streams, &n_streams);
+
+    CHECK_CA_ERROR("could not get number of streams.");
+
+    for (int i = 0; i < n_streams; i++) {
+        if (ca_stream_supports_digital(ao, streams[i])) {
+            talloc_free(streams);
+            return true;
+        }
+    }
+
+    talloc_free(streams);
+
+coreaudio_error:
+    return false;
+}
+
+static OSStatus ca_property_listener(
+    AudioObjectPropertySelector selector,
+    AudioObjectID object, uint32_t n_addresses,
+    const AudioObjectPropertyAddress addresses[],
+    void *data)
+{
+    void *talloc_ctx = talloc_new(NULL);
+
+    for (int i = 0; i < n_addresses; i++) {
+        if (addresses[i].mSelector == selector) {
+            if (data) *(volatile int *)data = 1;
+            break;
+        }
+    }
+    talloc_free(talloc_ctx);
+    return noErr;
+}
+
+static OSStatus ca_stream_listener(
+    AudioObjectID object, uint32_t n_addresses,
+    const AudioObjectPropertyAddress addresses[],
+    void *data)
+{
+    return ca_property_listener(kAudioStreamPropertyPhysicalFormat,
+                                object, n_addresses, addresses, data);
+}
+
+static OSStatus ca_device_listener(
+    AudioObjectID object, uint32_t n_addresses,
+    const AudioObjectPropertyAddress addresses[],
+    void *data)
+{
+    return ca_property_listener(kAudioDevicePropertyDeviceHasChanged,
+                                object, n_addresses, addresses, data);
+}
+
+static OSStatus ca_lock_device(AudioDeviceID device, pid_t *pid) {
+    *pid = getpid();
+    OSStatus err = CA_SET(device, kAudioDevicePropertyHogMode, pid);
+    if (err != noErr)
+        *pid = -1;
+
+    return err;
+}
+
+static OSStatus ca_unlock_device(AudioDeviceID device, pid_t *pid) {
+    if (*pid == getpid()) {
+        *pid = -1;
+        return CA_SET(device, kAudioDevicePropertyHogMode, &pid);
+    }
+    return noErr;
+}
+
+static OSStatus ca_change_mixing(struct ao *ao, AudioDeviceID device,
+                                 uint32_t val, bool *changed) {
+    *changed = false;
+
+    AudioObjectPropertyAddress p_addr = (AudioObjectPropertyAddress) {
+        .mSelector = kAudioDevicePropertySupportsMixing,
+        .mScope    = kAudioObjectPropertyScopeGlobal,
+        .mElement  = kAudioObjectPropertyElementMaster,
+    };
+
+    if (AudioObjectHasProperty(device, &p_addr)) {
+        OSStatus err;
+        Boolean writeable = 0;
+        err = CA_SETTABLE(device, kAudioDevicePropertySupportsMixing,
+                          &writeable);
+
+        if (!CHECK_CA_WARN("can't tell if mixing property is settable")) {
+            return err;
+        }
+
+        if (!writeable)
+            return noErr;
+
+        err = CA_SET(device, kAudioDevicePropertySupportsMixing, &val);
+        if (err != noErr)
+            return err;
+
+        if (!CHECK_CA_WARN("can't set mix mode")) {
+            return err;
+        }
+
+        *changed = true;
+    }
+
+    return noErr;
+}
+
+static OSStatus ca_disable_mixing(struct ao *ao,
+                                  AudioDeviceID device, bool *changed) {
+    return ca_change_mixing(ao, device, 0, changed);
+}
+
+static OSStatus ca_enable_mixing(struct ao *ao,
+                                 AudioDeviceID device, bool changed) {
+    if (changed) {
+        bool dont_care = false;
+        return ca_change_mixing(ao, device, 1, &dont_care);
+    }
+
+    return noErr;
+}
+
+static OSStatus ca_change_device_listening(AudioDeviceID device,
+                                           void *flag, bool enabled)
+{
+    AudioObjectPropertyAddress p_addr = (AudioObjectPropertyAddress) {
+        .mSelector = kAudioDevicePropertyDeviceHasChanged,
+        .mScope    = kAudioObjectPropertyScopeGlobal,
+        .mElement  = kAudioObjectPropertyElementMaster,
+    };
+
+    if (enabled) {
+        return AudioObjectAddPropertyListener(
+            device, &p_addr, ca_device_listener, flag);
+    } else {
+        return AudioObjectRemovePropertyListener(
+            device, &p_addr, ca_device_listener, flag);
+    }
+}
+
+static OSStatus ca_enable_device_listener(AudioDeviceID device, void *flag) {
+    return ca_change_device_listening(device, flag, true);
+}
+
+static OSStatus ca_disable_device_listener(AudioDeviceID device, void *flag) {
+    return ca_change_device_listening(device, flag, false);
+}
+
+static bool ca_change_format(struct ao *ao, AudioStreamID stream,
+                             AudioStreamBasicDescription change_format)
+{
+    OSStatus err = noErr;
+    AudioObjectPropertyAddress p_addr;
+    volatile int stream_format_changed = 0;
+
+    ca_print_asbd(ao, "setting stream format:", &change_format);
+
+    /* Install the callback. */
+    p_addr = (AudioObjectPropertyAddress) {
+        .mSelector = kAudioStreamPropertyPhysicalFormat,
+        .mScope    = kAudioObjectPropertyScopeGlobal,
+        .mElement  = kAudioObjectPropertyElementMaster,
+    };
+
+    err = AudioObjectAddPropertyListener(stream, &p_addr, ca_stream_listener,
+                                         (void *)&stream_format_changed);
+    if (!CHECK_CA_WARN("can't add property listener during format change")) {
+        return false;
+    }
+
+    /* Change the format. */
+    err = CA_SET(stream, kAudioStreamPropertyPhysicalFormat, &change_format);
+    if (!CHECK_CA_WARN("error changing physical format")) {
+        return false;
+    }
+
+    /* The AudioStreamSetProperty is not only asynchronious,
+     * it is also not Atomic, in its behaviour.
+     * Therefore we check 5 times before we really give up. */
+    bool format_set = false;
+    for (int i = 0; !format_set && i < 5; i++) {
+        for (int j = 0; !stream_format_changed && j < 50; j++)
+            mp_sleep_us(10000);
+
+        if (stream_format_changed) {
+            stream_format_changed = 0;
+        } else {
+            MP_VERBOSE(ao, "reached timeout\n");
+        }
+
+        AudioStreamBasicDescription actual_format;
+        err = CA_GET(stream, kAudioStreamPropertyPhysicalFormat, &actual_format);
+
+        ca_print_asbd(ao, "actual format in use:", &actual_format);
+        if (actual_format.mSampleRate == change_format.mSampleRate &&
+            actual_format.mFormatID == change_format.mFormatID &&
+            actual_format.mFramesPerPacket == change_format.mFramesPerPacket) {
+            format_set = true;
+        }
+    }
+
+    err = AudioObjectRemovePropertyListener(stream, &p_addr, ca_stream_listener,
+                                            (void *)&stream_format_changed);
+
+    if (!CHECK_CA_WARN("can't remove property listener")) {
+        return false;
+    }
+
+    return format_set;
+}
+
+
+struct priv {
+    AudioDeviceID device;   // selected device
+
+    bool paused;
+
+    struct mp_ring *buffer;
+
+    // digital render callback
+    AudioDeviceIOProcID render_cb;
+
+    // pid set for hog mode, (-1) means that hog mode on the device was
+    // released. hog mode is exclusive access to a device
+    pid_t hog_pid;
+
+    // stream selected for digital playback by the detection in init
+    AudioStreamID stream;
+
+    // stream index in an AudioBufferList
+    int stream_idx;
+
+    // format we changed the stream to: for the digital case each application
+    // sets the stream format for a device to what it needs
+    AudioStreamBasicDescription stream_asbd;
+    AudioStreamBasicDescription original_asbd;
+
+    bool changed_mixing;
+    int stream_asbd_changed;
+    bool muted;
+
+    // options
+    int opt_device_id;
+    int opt_list;
+};
+
+static int get_ring_size(struct ao *ao)
+{
+    return af_fmt_seconds_to_bytes(
+            ao->format, 0.5, ao->channels.num, ao->samplerate);
+}
+
+static OSStatus render_cb_digital(
+        AudioDeviceID device, const AudioTimeStamp *ts,
+        const void *in_data, const AudioTimeStamp *in_ts,
+        AudioBufferList *out_data, const AudioTimeStamp *out_ts, void *ctx)
+{
+    struct ao *ao    = ctx;
+    struct priv *p   = ao->priv;
+    AudioBuffer buf  = out_data->mBuffers[p->stream_idx];
+    int requested    = buf.mDataByteSize;
+
+    if (p->muted)
+        mp_ring_drain(p->buffer, requested);
+    else
+        mp_ring_read(p->buffer, buf.mData, requested);
+
+    return noErr;
+}
+
+static int control(struct ao *ao, enum aocontrol cmd, void *arg)
+{
+    struct priv *p = ao->priv;
+    ao_control_vol_t *control_vol;
+    switch (cmd) {
+    case AOCONTROL_GET_VOLUME:
+        control_vol = (ao_control_vol_t *)arg;
+        // Digital output has no volume adjust.
+        int digitalvol = p->muted ? 0 : 100;
+        *control_vol = (ao_control_vol_t) {
+            .left = digitalvol, .right = digitalvol,
+        };
+        return CONTROL_TRUE;
+
+    case AOCONTROL_SET_VOLUME:
+        control_vol = (ao_control_vol_t *)arg;
+        // Digital output can not set volume. Here we have to return true
+        // to make mixer forget it. Else mixer will add a soft filter,
+        // that's not we expected and the filter not support ac3 stream
+        // will cause mplayer die.
+
+        // Although not support set volume, but at least we support mute.
+        // MPlayer set mute by set volume to zero, we handle it.
+        if (control_vol->left == 0 && control_vol->right == 0)
+            p->muted = true;
+        else
+            p->muted = false;
+        return CONTROL_TRUE;
+
+    } // end switch
+    return CONTROL_UNKNOWN;
+}
+
+static int init_digital(struct ao *ao, AudioStreamBasicDescription asbd);
+
+static int init(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+
+    if (p->opt_list) ca_print_device_list(ao);
+
+    OSStatus err = ca_select_device(ao, p->opt_device_id, &p->device);
+    CHECK_CA_ERROR("failed to select device");
+
+    ao->format = af_fmt_from_planar(ao->format);
+
+    bool supports_digital = false;
+    /* Probe whether device support S/PDIF stream output if input is AC3 stream,
+     * or anything else IEC61937-framed. */
+    if (AF_FORMAT_IS_IEC61937(ao->format)) {
+        if (ca_device_supports_digital(ao, p->device))
+            supports_digital = true;
+    }
+
+    if (!supports_digital) {
+        MP_ERR(ao, "selected device doesn't support digital formats\n");
+        goto coreaudio_error;
+    } // closes if (!supports_digital)
+
+    // Build ASBD for the input format
+    AudioStreamBasicDescription asbd;
+    ca_fill_asbd(ao, &asbd);
+
+    return init_digital(ao, asbd);
+
+coreaudio_error:
+    return CONTROL_ERROR;
+}
+
+static int init_digital(struct ao *ao, AudioStreamBasicDescription asbd)
+{
+    struct priv *p = ao->priv;
+    OSStatus err = noErr;
+
+    uint32_t is_alive = 1;
+    err = CA_GET(p->device, kAudioDevicePropertyDeviceIsAlive, &is_alive);
+    CHECK_CA_WARN("could not check whether device is alive");
+
+    if (!is_alive)
+        MP_WARN(ao , "device is not alive\n");
+
+    err = ca_lock_device(p->device, &p->hog_pid);
+    CHECK_CA_WARN("failed to set hogmode");
+
+    err = ca_disable_mixing(ao, p->device, &p->changed_mixing);
+    CHECK_CA_WARN("failed to disable mixing");
+
+    AudioStreamID *streams;
+    size_t n_streams;
+
+    /* Get a list of all the streams on this device. */
+    err = CA_GET_ARY_O(p->device, kAudioDevicePropertyStreams,
+                       &streams, &n_streams);
+
+    CHECK_CA_ERROR("could not get number of streams");
+
+    for (int i = 0; i < n_streams && p->stream_idx < 0; i++) {
+        bool digital = ca_stream_supports_digital(ao, streams[i]);
+
+        if (digital) {
+            err = CA_GET(streams[i], kAudioStreamPropertyPhysicalFormat,
+                         &p->original_asbd);
+            if (!CHECK_CA_WARN("could not get stream's physical format to "
+                               "revert to, getting the next one"))
+                continue;
+
+            AudioStreamRangedDescription *formats;
+            size_t n_formats;
+
+            err = CA_GET_ARY(streams[i],
+                             kAudioStreamPropertyAvailablePhysicalFormats,
+                             &formats, &n_formats);
+
+            if (!CHECK_CA_WARN("could not get number of stream formats"))
+                continue; // try next one
+
+            int req_rate_format = -1;
+            int max_rate_format = -1;
+
+            p->stream = streams[i];
+            p->stream_idx = i;
+
+            for (int j = 0; j < n_formats; j++)
+                if (ca_format_is_digital(formats[j].mFormat)) {
+                    // select the digital format that has exactly the same
+                    // samplerate. If an exact match cannot be found, select
+                    // the format with highest samplerate as backup.
+                    if (formats[j].mFormat.mSampleRate == asbd.mSampleRate) {
+                        req_rate_format = j;
+                        break;
+                    } else if (max_rate_format < 0 ||
+                        formats[j].mFormat.mSampleRate >
+                        formats[max_rate_format].mFormat.mSampleRate)
+                        max_rate_format = j;
+                }
+
+            if (req_rate_format >= 0)
+                p->stream_asbd = formats[req_rate_format].mFormat;
+            else
+                p->stream_asbd = formats[max_rate_format].mFormat;
+
+            talloc_free(formats);
+        }
+    }
+
+    talloc_free(streams);
+
+    if (p->stream_idx < 0) {
+        MP_WARN(ao , "can't find any digital output stream format\n");
+        goto coreaudio_error;
+    }
+
+    if (!ca_change_format(ao, p->stream, p->stream_asbd))
+        goto coreaudio_error;
+
+    void *changed = (void *) &(p->stream_asbd_changed);
+    err = ca_enable_device_listener(p->device, changed);
+    CHECK_CA_ERROR("cannot install format change listener during init");
+
+    if (p->stream_asbd.mFormatFlags & kAudioFormatFlagIsBigEndian)
+        MP_WARN(ao, "stream has non-native byte order, output may fail\n");
+
+    ao->samplerate = p->stream_asbd.mSampleRate;
+    ao->bps = ao->samplerate *
+                  (p->stream_asbd.mBytesPerPacket /
+                   p->stream_asbd.mFramesPerPacket);
+
+    p->buffer = mp_ring_new(p, get_ring_size(ao));
+
+    err = AudioDeviceCreateIOProcID(p->device,
+                                    (AudioDeviceIOProc)render_cb_digital,
+                                    (void *)ao,
+                                    &p->render_cb);
+
+    CHECK_CA_ERROR("failed to register digital render callback");
+
+    reset(ao);
+
+    return CONTROL_TRUE;
+
+coreaudio_error:
+    err = ca_unlock_device(p->device, &p->hog_pid);
+    CHECK_CA_WARN("can't release hog mode");
+    return CONTROL_ERROR;
+}
+
+static int play(struct ao *ao, void **data, int samples, int flags)
+{
+    struct priv *p   = ao->priv;
+    void *output_samples = data[0];
+    int num_bytes = samples * ao->sstride;
+
+    // Check whether we need to reset the digital output stream.
+    if (p->stream_asbd_changed) {
+        p->stream_asbd_changed = 0;
+        if (ca_stream_supports_digital(ao, p->stream)) {
+            if (!ca_change_format(ao, p->stream, p->stream_asbd)) {
+                MP_WARN(ao , "can't restore digital output\n");
+            } else {
+                MP_WARN(ao, "restoring digital output succeeded.\n");
+                reset(ao);
+            }
+        }
+    }
+
+    int wrote = mp_ring_write(p->buffer, output_samples, num_bytes);
+    audio_resume(ao);
+
+    return wrote / ao->sstride;
+}
+
+static void reset(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    audio_pause(ao);
+    mp_ring_reset(p->buffer);
+}
+
+static int get_space(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    return mp_ring_available(p->buffer) / ao->sstride;
+}
+
+static float get_delay(struct ao *ao)
+{
+    // FIXME: should also report the delay of coreaudio itself (hardware +
+    // internal buffers)
+    struct priv *p = ao->priv;
+    return mp_ring_buffered(p->buffer) / (float)ao->bps;
+}
+
+static void uninit(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    OSStatus err = noErr;
+
+    void *changed = (void *) &(p->stream_asbd_changed);
+    err = ca_disable_device_listener(p->device, changed);
+    CHECK_CA_WARN("can't remove device listener, this may cause a crash");
+
+    err = AudioDeviceStop(p->device, p->render_cb);
+    CHECK_CA_WARN("failed to stop audio device");
+
+    err = AudioDeviceDestroyIOProcID(p->device, p->render_cb);
+    CHECK_CA_WARN("failed to remove device render callback");
+
+    if (!ca_change_format(ao, p->stream, p->original_asbd))
+        MP_WARN(ao, "can't revert to original device format");
+
+    err = ca_enable_mixing(ao, p->device, p->changed_mixing);
+    CHECK_CA_WARN("can't re-enable mixing");
+
+    err = ca_unlock_device(p->device, &p->hog_pid);
+    CHECK_CA_WARN("can't release hog mode");
+}
+
+static void audio_pause(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+
+    if (p->paused)
+        return;
+
+    OSStatus err = AudioDeviceStop(p->device, p->render_cb);
+    CHECK_CA_WARN("can't stop digital device");
+
+    p->paused = true;
+}
+
+static void audio_resume(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+
+    if (!p->paused)
+        return;
+
+    OSStatus err = AudioDeviceStart(p->device, p->render_cb);
+    CHECK_CA_WARN("can't start digital device");
+
+    p->paused = false;
+}
+
+#define OPT_BASE_STRUCT struct priv
+
+const struct ao_driver audio_out_coreaudio_exclusive = {
+    .description = "CoreAudio Exclusive Mode",
+    .name      = "coreaudio_exclusive",
+    .uninit    = uninit,
+    .init      = init,
+    .play      = play,
+    .control   = control,
+    .get_space = get_space,
+    .get_delay = get_delay,
+    .reset     = reset,
+    .pause     = audio_pause,
+    .resume    = audio_resume,
+    .priv_size = sizeof(struct priv),
+    .priv_defaults = &(const struct priv){
+        .muted = false,
+        .stream_asbd_changed = 0,
+        .hog_pid = -1,
+        .stream = 0,
+        .stream_idx = -1,
+        .changed_mixing = false,
+    },
+    .options = (const struct m_option[]) {
+        OPT_INT("device_id", opt_device_id, 0, OPTDEF_INT(-1)),
+        OPT_FLAG("list", opt_list, 0),
+        {0}
+    },
+};

audio/out/ao_coreaudio_properties.h

@@ -37,6 +37,7 @@ OSStatus ca_set(AudioObjectID id, ca_scope scope, ca_sel selector,
 
 #define CA_GET(id, sel, data) ca_get(id, CA_GLOBAL, sel, sizeof(*(data)), data)
 #define CA_SET(id, sel, data) ca_set(id, CA_GLOBAL, sel, sizeof(*(data)), data)
+#define CA_GET_O(id, sel, data) ca_get(id, CA_OUTPUT, sel, sizeof(*(data)), data)
 
 OSStatus ca_get_ary(AudioObjectID id, ca_scope scope, ca_sel selector,
                     uint32_t element_size, void **data, size_t *elements);

audio/out/ao_coreaudio_utils.c

@@ -25,12 +25,76 @@
 #include "audio/out/ao_coreaudio_utils.h"
 #include "audio/out/ao_coreaudio_properties.h"
 #include "osdep/timer.h"
+#include "osdep/endian.h"
+#include "audio/format.h"
+
+void ca_print_device_list(struct ao *ao)
+{
+    char *help = talloc_strdup(NULL, "Available output devices:\n");
+
+    AudioDeviceID *devs;
+    size_t n_devs;
+
+    OSStatus err =
+        CA_GET_ARY(kAudioObjectSystemObject, kAudioHardwarePropertyDevices,
+                   &devs, &n_devs);
+
+    CHECK_CA_ERROR("Failed to get list of output devices.");
+
+    for (int i = 0; i < n_devs; i++) {
+        char *name;
+        err = CA_GET_STR(devs[i], kAudioObjectPropertyName, &name);
+
+        if (err == noErr)
+            talloc_steal(devs, name);
+        else
+            name = "Unknown";
+
+        help = talloc_asprintf_append(
+                help, "  * %s (id: %" PRIu32 ")\n", name, devs[i]);
+    }
+
+    talloc_free(devs);
+
+coreaudio_error:
+    MP_INFO(ao, "%s", help);
+    talloc_free(help);
+}
+
+OSStatus ca_select_device(struct ao *ao, int selection, AudioDeviceID *device)
+{
+    OSStatus err = noErr;
+    *device = 0;
+    if (selection < 0) {
+        // device not set by user, get the default one
+        err = CA_GET(kAudioObjectSystemObject,
+                     kAudioHardwarePropertyDefaultOutputDevice,
+                     device);
+        CHECK_CA_ERROR("could not get default audio device");
+    } else {
+        *device = selection;
+    }
+
+    if (mp_msg_test(ao->log, MSGL_V)) {
+        char *name;
+        err = CA_GET_STR(*device, kAudioObjectPropertyName, &name);
+        CHECK_CA_ERROR("could not get selected audio device name");
+
+        MP_VERBOSE(ao, "selected audio output device: %s (%" PRIu32 ")\n",
+                       name, *device);
+
+        talloc_free(name);
+    }
+
+coreaudio_error:
+    return err;
+}
 
 char *fourcc_repr(void *talloc_ctx, uint32_t code)
 {
     // Extract FourCC letters from the uint32_t and finde out if it's a valid
     // code that is made of letters.
-    char fcc[4] = {
+    unsigned char fcc[4] = {
         (code >> 24) & 0xFF,
         (code >> 16) & 0xFF,
         (code >> 8)  & 0xFF,
@@ -39,7 +103,7 @@ char *fourcc_repr(void *talloc_ctx, uint32_t code)
 
     bool valid_fourcc = true;
     for (int i = 0; i < 4; i++)
-        if (!isprint(fcc[i]))
+        if (fcc[i] >= 32 && fcc[i] < 128)
             valid_fourcc = false;
 
     char *repr;
@@ -63,6 +127,32 @@ bool check_ca_st(struct ao *ao, int level, OSStatus code, const char *message)
     return false;
 }
 
+void ca_fill_asbd(struct ao *ao, AudioStreamBasicDescription *asbd)
+{
+    asbd->mSampleRate       = ao->samplerate;
+    // Set "AC3" for other spdif formats too - unknown if that works.
+    asbd->mFormatID         = AF_FORMAT_IS_IEC61937(ao->format) ?
+                              kAudioFormat60958AC3 :
+                              kAudioFormatLinearPCM;
+    asbd->mChannelsPerFrame = ao->channels.num;
+    asbd->mBitsPerChannel   = af_fmt2bits(ao->format);
+    asbd->mFormatFlags      = kAudioFormatFlagIsPacked;
+
+    if ((ao->format & AF_FORMAT_TYPE_MASK) == AF_FORMAT_F)
+        asbd->mFormatFlags |= kAudioFormatFlagIsFloat;
+
+    if ((ao->format & AF_FORMAT_SIGN_MASK) == AF_FORMAT_SI)
+        asbd->mFormatFlags |= kAudioFormatFlagIsSignedInteger;
+
+    if (BYTE_ORDER == BIG_ENDIAN)
+        asbd->mFormatFlags |= kAudioFormatFlagIsBigEndian;
+
+    asbd->mFramesPerPacket = 1;
+    asbd->mBytesPerPacket = asbd->mBytesPerFrame =
+        asbd->mFramesPerPacket * asbd->mChannelsPerFrame *
+        (asbd->mBitsPerChannel / 8);
+}
+
 void ca_print_asbd(struct ao *ao, const char *description,
                    const AudioStreamBasicDescription *asbd)
 {
@@ -87,375 +177,3 @@ void ca_print_asbd(struct ao *ao, const char *description,
     talloc_free(format);
 }
 
-bool ca_format_is_digital(AudioStreamBasicDescription asbd)
-{
-    switch (asbd.mFormatID)
-    case 'IAC3':
-    case 'iac3':
-    case  kAudioFormat60958AC3:
-    case  kAudioFormatAC3:
-        return true;
-    return false;
-}
-
-bool ca_stream_supports_digital(struct ao *ao, AudioStreamID stream)
-{
-    AudioStreamRangedDescription *formats = NULL;
-    size_t n_formats;
-
-    OSStatus err =
-        CA_GET_ARY(stream, kAudioStreamPropertyAvailablePhysicalFormats,
-                   &formats, &n_formats);
-
-    CHECK_CA_ERROR("Could not get number of stream formats.");
-
-    for (int i = 0; i < n_formats; i++) {
-        AudioStreamBasicDescription asbd = formats[i].mFormat;
-        ca_print_asbd(ao, "supported format:", &(asbd));
-        if (ca_format_is_digital(asbd)) {
-            talloc_free(formats);
-            return true;
-        }
-    }
-
-    talloc_free(formats);
-coreaudio_error:
-    return false;
-}
-
-bool ca_device_supports_digital(struct ao *ao, AudioDeviceID device)
-{
-    AudioStreamID *streams = NULL;
-    size_t n_streams;
-
-    /* Retrieve all the output streams. */
-    OSStatus err =
-        CA_GET_ARY_O(device, kAudioDevicePropertyStreams, &streams, &n_streams);
-
-    CHECK_CA_ERROR("could not get number of streams.");
-
-    for (int i = 0; i < n_streams; i++) {
-        if (ca_stream_supports_digital(ao, streams[i])) {
-            talloc_free(streams);
-            return true;
-        }
-    }
-
-    talloc_free(streams);
-
-coreaudio_error:
-    return false;
-}
-
-OSStatus ca_property_listener(AudioObjectPropertySelector selector,
-                              AudioObjectID object, uint32_t n_addresses,
-                              const AudioObjectPropertyAddress addresses[],
-                              void *data)
-{
-    void *talloc_ctx = talloc_new(NULL);
-
-    for (int i = 0; i < n_addresses; i++) {
-        if (addresses[i].mSelector == selector) {
-            if (data) *(volatile int *)data = 1;
-            break;
-        }
-    }
-    talloc_free(talloc_ctx);
-    return noErr;
-}
-
-OSStatus ca_stream_listener(AudioObjectID object, uint32_t n_addresses,
-                            const AudioObjectPropertyAddress addresses[],
-                            void *data)
-{
-    return ca_property_listener(kAudioStreamPropertyPhysicalFormat,
-                                object, n_addresses, addresses, data);
-}
-
-OSStatus ca_device_listener(AudioObjectID object, uint32_t n_addresses,
-                            const AudioObjectPropertyAddress addresses[],
-                            void *data)
-{
-    return ca_property_listener(kAudioDevicePropertyDeviceHasChanged,
-                                object, n_addresses, addresses, data);
-}
-
-OSStatus ca_lock_device(AudioDeviceID device, pid_t *pid) {
-    *pid = getpid();
-    OSStatus err = CA_SET(device, kAudioDevicePropertyHogMode, pid);
-    if (err != noErr)
-        *pid = -1;
-
-    return err;
-}
-
-OSStatus ca_unlock_device(AudioDeviceID device, pid_t *pid) {
-    if (*pid == getpid()) {
-        *pid = -1;
-        return CA_SET(device, kAudioDevicePropertyHogMode, &pid);
-    }
-    return noErr;
-}
-
-static OSStatus ca_change_mixing(struct ao *ao, AudioDeviceID device,
-                                 uint32_t val, bool *changed) {
-    *changed = false;
-
-    AudioObjectPropertyAddress p_addr = (AudioObjectPropertyAddress) {
-        .mSelector = kAudioDevicePropertySupportsMixing,
-        .mScope    = kAudioObjectPropertyScopeGlobal,
-        .mElement  = kAudioObjectPropertyElementMaster,
-    };
-
-    if (AudioObjectHasProperty(device, &p_addr)) {
-        OSStatus err;
-        Boolean writeable = 0;
-        err = CA_SETTABLE(device, kAudioDevicePropertySupportsMixing,
-                          &writeable);
-
-        if (!CHECK_CA_WARN("can't tell if mixing property is settable")) {
-            return err;
-        }
-
-        if (!writeable)
-            return noErr;
-
-        err = CA_SET(device, kAudioDevicePropertySupportsMixing, &val);
-        if (err != noErr)
-            return err;
-
-        if (!CHECK_CA_WARN("can't set mix mode")) {
-            return err;
-        }
-
-        *changed = true;
-    }
-
-    return noErr;
-}
-
-OSStatus ca_disable_mixing(struct ao *ao, AudioDeviceID device, bool *changed) {
-    return ca_change_mixing(ao, device, 0, changed);
-}
-
-OSStatus ca_enable_mixing(struct ao *ao, AudioDeviceID device, bool changed) {
-    if (changed) {
-        bool dont_care = false;
-        return ca_change_mixing(ao, device, 1, &dont_care);
-    }
-
-    return noErr;
-}
-
-static OSStatus ca_change_device_listening(AudioDeviceID device,
-                                           void *flag, bool enabled)
-{
-    AudioObjectPropertyAddress p_addr = (AudioObjectPropertyAddress) {
-        .mSelector = kAudioDevicePropertyDeviceHasChanged,
-        .mScope    = kAudioObjectPropertyScopeGlobal,
-        .mElement  = kAudioObjectPropertyElementMaster,
-    };
-
-    if (enabled) {
-        return AudioObjectAddPropertyListener(
-            device, &p_addr, ca_device_listener, flag);
-    } else {
-        return AudioObjectRemovePropertyListener(
-            device, &p_addr, ca_device_listener, flag);
-    }
-}
-
-OSStatus ca_enable_device_listener(AudioDeviceID device, void *flag) {
-    return ca_change_device_listening(device, flag, true);
-}
-
-OSStatus ca_disable_device_listener(AudioDeviceID device, void *flag) {
-    return ca_change_device_listening(device, flag, false);
-}
-
-bool ca_change_format(struct ao *ao, AudioStreamID stream,
-                      AudioStreamBasicDescription change_format)
-{
-    OSStatus err = noErr;
-    AudioObjectPropertyAddress p_addr;
-    volatile int stream_format_changed = 0;
-
-    ca_print_asbd(ao, "setting stream format:", &change_format);
-
-    /* Install the callback. */
-    p_addr = (AudioObjectPropertyAddress) {
-        .mSelector = kAudioStreamPropertyPhysicalFormat,
-        .mScope    = kAudioObjectPropertyScopeGlobal,
-        .mElement  = kAudioObjectPropertyElementMaster,
-    };
-
-    err = AudioObjectAddPropertyListener(stream, &p_addr, ca_stream_listener,
-                                         (void *)&stream_format_changed);
-    if (!CHECK_CA_WARN("can't add property listener during format change")) {
-        return false;
-    }
-
-    /* Change the format. */
-    err = CA_SET(stream, kAudioStreamPropertyPhysicalFormat, &change_format);
-    if (!CHECK_CA_WARN("error changing physical format")) {
-        return false;
-    }
-
-    /* The AudioStreamSetProperty is not only asynchronious,
-     * it is also not Atomic, in its behaviour.
-     * Therefore we check 5 times before we really give up. */
-    bool format_set = false;
-    for (int i = 0; !format_set && i < 5; i++) {
-        for (int j = 0; !stream_format_changed && j < 50; j++)
-            mp_sleep_us(10000);
-
-        if (stream_format_changed) {
-            stream_format_changed = 0;
-        } else {
-            MP_VERBOSE(ao, "reached timeout\n");
-        }
-
-        AudioStreamBasicDescription actual_format;
-        err = CA_GET(stream, kAudioStreamPropertyPhysicalFormat, &actual_format);
-
-        ca_print_asbd(ao, "actual format in use:", &actual_format);
-        if (actual_format.mSampleRate == change_format.mSampleRate &&
-            actual_format.mFormatID == change_format.mFormatID &&
-            actual_format.mFramesPerPacket == change_format.mFramesPerPacket) {
-            format_set = true;
-        }
-    }
-
-    err = AudioObjectRemovePropertyListener(stream, &p_addr, ca_stream_listener,
-                                            (void *)&stream_format_changed);
-
-    if (!CHECK_CA_WARN("can't remove property listener")) {
-        return false;
-    }
-
-    return format_set;
-}
-
-static const int speaker_map[][2] = {
-    { kAudioChannelLabel_Left,                 MP_SPEAKER_ID_FL   },
-    { kAudioChannelLabel_Right,                MP_SPEAKER_ID_FR   },
-    { kAudioChannelLabel_Center,               MP_SPEAKER_ID_FC   },
-    { kAudioChannelLabel_LFEScreen,            MP_SPEAKER_ID_LFE  },
-    { kAudioChannelLabel_LeftSurround,         MP_SPEAKER_ID_BL   },
-    { kAudioChannelLabel_RightSurround,        MP_SPEAKER_ID_BR   },
-    { kAudioChannelLabel_LeftCenter,           MP_SPEAKER_ID_FLC  },
-    { kAudioChannelLabel_RightCenter,          MP_SPEAKER_ID_FRC  },
-    { kAudioChannelLabel_CenterSurround,       MP_SPEAKER_ID_BC   },
-    { kAudioChannelLabel_LeftSurroundDirect,   MP_SPEAKER_ID_SL   },
-    { kAudioChannelLabel_RightSurroundDirect,  MP_SPEAKER_ID_SR   },
-    { kAudioChannelLabel_TopCenterSurround,    MP_SPEAKER_ID_TC   },
-    { kAudioChannelLabel_VerticalHeightLeft,   MP_SPEAKER_ID_TFL  },
-    { kAudioChannelLabel_VerticalHeightCenter, MP_SPEAKER_ID_TFC  },
-    { kAudioChannelLabel_VerticalHeightRight,  MP_SPEAKER_ID_TFR  },
-    { kAudioChannelLabel_TopBackLeft,          MP_SPEAKER_ID_TBL  },
-    { kAudioChannelLabel_TopBackCenter,        MP_SPEAKER_ID_TBC  },
-    { kAudioChannelLabel_TopBackRight,         MP_SPEAKER_ID_TBR  },
-
-    // unofficial extensions
-    { kAudioChannelLabel_RearSurroundLeft,     MP_SPEAKER_ID_SDL  },
-    { kAudioChannelLabel_RearSurroundRight,    MP_SPEAKER_ID_SDR  },
-    { kAudioChannelLabel_LeftWide,             MP_SPEAKER_ID_WL   },
-    { kAudioChannelLabel_RightWide,            MP_SPEAKER_ID_WR   },
-    { kAudioChannelLabel_LFE2,                 MP_SPEAKER_ID_LFE2 },
-
-    { kAudioChannelLabel_HeadphonesLeft,       MP_SPEAKER_ID_DL   },
-    { kAudioChannelLabel_HeadphonesRight,      MP_SPEAKER_ID_DR   },
-
-    { kAudioChannelLabel_Unknown,              -1 },
-};
-
-static int ca_label_to_mp_speaker_id(AudioChannelLabel label)
-{
-    for (int i = 0; speaker_map[i][1] >= 0; i++)
-        if (speaker_map[i][0] == label)
-            return speaker_map[i][1];
-    return -1;
-}
-
-static void ca_log_layout(struct ao *ao, AudioChannelLayout *layout)
-{
-    if (!mp_msg_test(ao->log, MSGL_V))
-        return;
-
-    AudioChannelDescription *descs = layout->mChannelDescriptions;
-
-    MP_VERBOSE(ao, "layout: tag: <%d>, bitmap: <%d>, "
-                   "descriptions <%d>\n",
-                   layout->mChannelLayoutTag,
-                   layout->mChannelBitmap,
-                   layout->mNumberChannelDescriptions);
-
-    for (int i = 0; i < layout->mNumberChannelDescriptions; i++) {
-        AudioChannelDescription d = descs[i];
-        MP_VERBOSE(ao, " - description %d: label <%d, %d>, flags: <%u>, "
-                       "coords: <%f, %f, %f>\n", i,
-                       d.mChannelLabel,
-                       ca_label_to_mp_speaker_id(d.mChannelLabel),
-                       d.mChannelFlags,
-                       d.mCoordinates[0],
-                       d.mCoordinates[1],
-                       d.mCoordinates[2]);
-    }
-}
-
-bool ca_layout_to_mp_chmap(struct ao *ao, AudioChannelLayout *layout,
-                           struct mp_chmap *chmap)
-{
-    AudioChannelLayoutTag tag  = layout->mChannelLayoutTag;
-    uint32_t layout_size       = sizeof(layout);
-    OSStatus err;
-
-    if (tag == kAudioChannelLayoutTag_UseChannelBitmap) {
-        err = AudioFormatGetProperty(kAudioFormatProperty_ChannelLayoutForBitmap,
-                                     sizeof(uint32_t),
-                                     &layout->mChannelBitmap,
-                                     &layout_size,
-                                     layout);
-        CHECK_CA_ERROR("failed to convert channel bitmap to descriptions");
-    } else if (tag != kAudioChannelLayoutTag_UseChannelDescriptions) {
-        err = AudioFormatGetProperty(kAudioFormatProperty_ChannelLayoutForTag,
-                                     sizeof(AudioChannelLayoutTag),
-                                     &layout->mChannelLayoutTag,
-                                     &layout_size,
-                                     layout);
-        CHECK_CA_ERROR("failed to convert channel tag to descriptions");
-    }
-
-    ca_log_layout(ao, layout);
-
-    // If the channel layout uses channel descriptions, from my
-    // experiments there are there three possibile cases:
-    // * The description has a label kAudioChannelLabel_Unknown:
-    //   Can't do anything about this (looks like non surround
-    //   layouts are like this).
-    // * The description uses positional information: this in
-    //   theory could be used but one would have to map spatial
-    //   positions to labels which is not really feasible.
-    // * The description has a well known label which can be mapped
-    //   to the waveextensible definition: this is the kind of
-    //   descriptions we process here.
-
-    for (int n = 0; n < layout->mNumberChannelDescriptions; n++) {
-        AudioChannelLabel label = layout->mChannelDescriptions[n].mChannelLabel;
-        uint8_t speaker = ca_label_to_mp_speaker_id(label);
-        if (label == kAudioChannelLabel_Unknown)
-            continue;
-        if (speaker < 0) {
-            MP_VERBOSE(ao, "channel label=%d unusable to build channel "
-                           "bitmap, skipping layout\n", label);
-        } else {
-            chmap->speaker[n] = speaker;
-            chmap->num = n + 1;
-        }
-    }
-
-    return chmap->num > 0;
-coreaudio_error:
-    ca_log_layout(ao, layout);
-    return false;
-}

audio/out/ao_coreaudio_utils.h

@@ -46,38 +46,11 @@ bool check_ca_st(struct ao *ao, int level, OSStatus code, const char *message);
         if (err != noErr) goto label; \
     } while (0)
 
+void ca_print_device_list(struct ao *ao);
+OSStatus ca_select_device(struct ao *ao, int selection, AudioDeviceID *device);
+
+void ca_fill_asbd(struct ao *ao, AudioStreamBasicDescription *asbd);
 void ca_print_asbd(struct ao *ao, const char *description,
                    const AudioStreamBasicDescription *asbd);
 
-bool ca_format_is_digital(AudioStreamBasicDescription asbd);
-bool ca_stream_supports_digital(struct ao *ao, AudioStreamID stream);
-bool ca_device_supports_digital(struct ao *ao, AudioDeviceID device);
-
-OSStatus ca_property_listener(AudioObjectPropertySelector selector,
-                              AudioObjectID object, uint32_t n_addresses,
-                              const AudioObjectPropertyAddress addresses[],
-                              void *data);
-
-OSStatus ca_stream_listener(AudioObjectID object, uint32_t n_addresses,
-                            const AudioObjectPropertyAddress addresses[],
-                            void *data);
-
-OSStatus ca_device_listener(AudioObjectID object, uint32_t n_addresses,
-                            const AudioObjectPropertyAddress addresses[],
-                            void *data);
-
-OSStatus ca_lock_device(AudioDeviceID device, pid_t *pid);
-OSStatus ca_unlock_device(AudioDeviceID device, pid_t *pid);
-OSStatus ca_disable_mixing(struct ao *ao, AudioDeviceID device, bool *changed);
-OSStatus ca_enable_mixing(struct ao *ao, AudioDeviceID device, bool changed);
-
-OSStatus ca_enable_device_listener(AudioDeviceID device, void *flag);
-OSStatus ca_disable_device_listener(AudioDeviceID device, void *flag);
-
-bool ca_change_format(struct ao *ao, AudioStreamID stream,
-                      AudioStreamBasicDescription change_format);
-
-bool ca_layout_to_mp_chmap(struct ao *ao, AudioChannelLayout *layout,
-                           struct mp_chmap *chmap);
-
 #endif /* MPV_COREAUDIO_UTILS_H */

audio/out/ao_dsound.c

@@ -98,6 +98,7 @@ struct priv {
     int outburst;                   ///play in multiple of chunks of this size
 
     int cfg_device;
+    int cfg_buffersize;
 };
 
 static float get_delay(struct ao *ao);
@@ -350,6 +351,8 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
         IDirectSoundBuffer_SetVolume(p->hdsbuf, volume);
         return CONTROL_OK;
     }
+    case AOCONTROL_HAS_SOFT_VOLUME:
+        return CONTROL_TRUE;
     }
     return CONTROL_UNKNOWN;
 }
@@ -370,7 +373,6 @@ static int init(struct ao *ao)
     if (!InitDirectSound(ao))
         return -1;
 
-    ao->no_persistent_volume = true;
     p->audio_volume = 100;
 
     // ok, now create the buffers
@@ -380,34 +382,33 @@ static int init(struct ao *ao)
     int format = af_fmt_from_planar(ao->format);
     int rate = ao->samplerate;
 
-    if (AF_FORMAT_IS_AC3(format))
-        format = AF_FORMAT_AC3;
-    else {
+    if (!AF_FORMAT_IS_IEC61937(format)) {
         struct mp_chmap_sel sel = {0};
         mp_chmap_sel_add_waveext(&sel);
         if (!ao_chmap_sel_adjust(ao, &sel, &ao->channels))
             return -1;
     }
     switch (format) {
-    case AF_FORMAT_AC3:
-    case AF_FORMAT_S24_LE:
-    case AF_FORMAT_S16_LE:
+    case AF_FORMAT_S24:
+    case AF_FORMAT_S16:
     case AF_FORMAT_U8:
         break;
     default:
+        if (AF_FORMAT_IS_IEC61937(format))
+            break;
         MP_VERBOSE(ao, "format %s not supported defaulting to Signed 16-bit Little-Endian\n",
                    af_fmt_to_str(format));
-        format = AF_FORMAT_S16_LE;
+        format = AF_FORMAT_S16;
     }
     //set our audio parameters
     ao->samplerate = rate;
     ao->format = format;
     ao->bps = ao->channels.num * rate * af_fmt2bps(format);
-    int buffersize = ao->bps; // space for 1 sec
+    int buffersize = ao->bps * p->cfg_buffersize / 1000;
     MP_VERBOSE(ao, "Samplerate:%iHz Channels:%i Format:%s\n", rate,
                ao->channels.num, af_fmt_to_str(format));
-    MP_VERBOSE(ao, "Buffersize:%d bytes (%d msec)\n",
-               buffersize, buffersize / ao->bps * 1000);
+    MP_VERBOSE(ao, "Buffersize:%d bytes (%f msec)\n",
+               buffersize, buffersize * 1000.0 / ao->bps);
 
     //fill waveformatex
     ZeroMemory(&wformat, sizeof(WAVEFORMATEXTENSIBLE));
@@ -415,7 +416,8 @@ static int init(struct ao *ao)
                     ? sizeof(WAVEFORMATEXTENSIBLE) - sizeof(WAVEFORMATEX) : 0;
     wformat.Format.nChannels = ao->channels.num;
     wformat.Format.nSamplesPerSec = rate;
-    if (AF_FORMAT_IS_AC3(format)) {
+    if (AF_FORMAT_IS_IEC61937(format)) {
+        // Whether it also works with e.g. DTS is unknown, but probably does.
         wformat.Format.wFormatTag = WAVE_FORMAT_DOLBY_AC3_SPDIF;
         wformat.Format.wBitsPerSample = 16;
         wformat.Format.nBlockAlign = 4;
@@ -583,7 +585,7 @@ static int get_space(struct ao *ao)
     int space = check_free_buffer_size(ao);
     if (space < p->min_free_space)
         return 0;
-    return (space - p->min_free_space) / ao->sstride;
+    return (space - p->min_free_space) / p->outburst * p->outburst / ao->sstride;
 }
 
 /**
@@ -636,6 +638,7 @@ const struct ao_driver audio_out_dsound = {
     .priv_size = sizeof(struct priv),
     .options = (const struct m_option[]) {
         OPT_INT("device", cfg_device, 0),
+        OPT_INTRANGE("buffersize", cfg_buffersize, 0, 1, 10000, OPTDEF_INT(200)),
         {0}
     },
 };

audio/out/ao_lavc.c

@@ -28,7 +28,6 @@
 #include <libavutil/common.h>
 #include <libavutil/audioconvert.h>
 
-#include "compat/libav.h"
 #include "config.h"
 #include "options/options.h"
 #include "common/common.h"
@@ -71,7 +70,7 @@ static void select_format(struct ao *ao, AVCodec *codec)
     {
         int fmt = af_from_avformat(*sampleformat);
         if (!fmt) {
-            MP_WARN(ao, "unsupported lavc format %s",
+            MP_WARN(ao, "unsupported lavc format %s\n",
                     av_get_sample_fmt_name(*sampleformat));
             continue;
         }
@@ -114,12 +113,11 @@ static int init(struct ao *ao)
 
     codec = encode_lavc_get_codec(ao->encode_lavc_ctx, ac->stream);
 
-    // ac->stream->time_base.num = 1;
-    // ac->stream->time_base.den = ao->samplerate;
-    // doing this breaks mpeg2ts in ffmpeg
-    // which doesn't properly force the time base to be 90000
-    // furthermore, ffmpeg.c doesn't do this either and works
-
+    // TODO: Remove this redundancy with encode_lavc_alloc_stream also
+    // setting the time base.
+    // Using codec->time_bvase is deprecated, but needed for older lavf.
+    ac->stream->time_base.num = 1;
+    ac->stream->time_base.den = ao->samplerate;
     ac->stream->codec->time_base.num = 1;
     ac->stream->codec->time_base.den = ao->samplerate;
 
@@ -342,8 +340,7 @@ static int play(struct ao *ao, void **data, int samples, int flags)
     }
 
     double pts = ectx->last_audio_in_pts;
-    pts += ectx->samples_since_last_pts / ao->samplerate;
-    ectx->samples_since_last_pts += samples;
+    pts += ectx->samples_since_last_pts / (double)ao->samplerate;
 
     size_t num_planes = af_fmt_is_planar(ao->format) ? ao->channels.num : 1;
 
@@ -451,6 +448,10 @@ static int play(struct ao *ao, void **data, int samples, int flags)
     }
 
     talloc_free(tempdata);
+
+    int taken = FFMIN(bufpos, orig_samples);
+    ectx->samples_since_last_pts += taken;
+
     pthread_mutex_unlock(&ectx->lock);
 
     if (flags & AOPLAY_FINAL_CHUNK) {
@@ -463,7 +464,7 @@ static int play(struct ao *ao, void **data, int samples, int flags)
         }
     }
 
-    return FFMIN(bufpos, orig_samples);
+    return taken;
 }
 
 static void drain(struct ao *ao)

audio/out/ao_null.c

@@ -90,7 +90,7 @@ static int init(struct ao *ao)
     struct mp_chmap_sel sel = {0};
     mp_chmap_sel_add_any(&sel);
     if (!ao_chmap_sel_adjust(ao, &sel, &ao->channels))
-        return -1;
+        mp_chmap_from_channels(&ao->channels, 2);
 
     priv->latency = priv->latency_sec * ao->samplerate;
 
@@ -148,7 +148,8 @@ static int get_space(struct ao *ao)
     struct priv *priv = ao->priv;
 
     drain(ao);
-    return priv->buffersize - priv->latency - priv->buffered;
+    int samples = priv->buffersize - priv->latency - priv->buffered;
+    return samples / priv->outburst * priv->outburst;
 }
 
 static int play(struct ao *ao, void **data, int samples, int flags)
@@ -210,7 +211,6 @@ const struct ao_driver audio_out_null = {
     .priv_size = sizeof(struct priv),
     .priv_defaults = &(const struct priv) {
         .bufferlen = 0.2,
-        .latency_sec = 0.5,
         .outburst = 256,
         .speed = 1,
     },

audio/out/ao_openal.c

@@ -76,6 +76,8 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
         vol->left = vol->right = volume * 100;
         return CONTROL_TRUE;
     }
+    case AOCONTROL_HAS_SOFT_VOLUME:
+        return CONTROL_TRUE;
     }
     return CONTROL_UNKNOWN;
 }
@@ -132,7 +134,6 @@ static int init(struct ao *ao)
         return -1;
     }
     ao_data = ao;
-    ao->no_persistent_volume = true;
     struct mp_chmap_sel sel = {0};
     for (i = 0; speaker_pos[i].id != -1; i++)
         mp_chmap_sel_add_speaker(&sel, speaker_pos[i].id);

audio/out/ao_oss.c

@@ -31,12 +31,16 @@
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <fcntl.h>
+#include <poll.h>
 #include <errno.h>
 #include <string.h>
+#include <strings.h>
 
 #include "config.h"
 #include "options/options.h"
 #include "common/msg.h"
+#include "osdep/timer.h"
+#include "osdep/endian.h"
 
 #if HAVE_SYS_SOUNDCARD_H
 #include <sys/soundcard.h>
@@ -51,14 +55,20 @@
 #include "ao.h"
 #include "internal.h"
 
+// Define to 0 if the device must be reopened to reset it (stop all playback,
+// clear the buffer), and the device should be closed when unused.
+// Define to 1 if SNDCTL_DSP_RESET should be used to reset without close.
+#define KEEP_DEVICE (defined(SNDCTL_DSP_RESET) && !defined(__NetBSD__))
+
 struct priv {
     int audio_fd;
-    int prepause_space;
+    int prepause_samples;
     int oss_mixer_channel;
-    audio_buf_info zz;
     int audio_delay_method;
     int buffersize;
     int outburst;
+    bool device_failed;
+    double audio_end;
 
     char *dsp;
     char *oss_mixer_device;
@@ -80,53 +90,60 @@ static const struct mp_chmap oss_layouts[MP_NUM_CHANNELS + 1] = {
     MP_CHMAP8(FL, FR, BL, BR, FC, LFE, SL, SR), // 7.1
 };
 
+#if !defined(AFMT_S16_NE) && defined(AFMT_S16_LE) && defined(AFMT_S16_BE)
+#define AFMT_S16_NE MP_SELECT_LE_BE(AFMT_S16_LE, AFMT_S16_BE)
+#endif
+
+#if !defined(AFMT_U16_NE) && defined(AFMT_U16_LE) && defined(AFMT_U16_BE)
+#define AFMT_U16_NE MP_SELECT_LE_BE(AFMT_U16_LE, AFMT_U16_BE)
+#endif
+
+#if !defined(AFMT_U24_NE) && defined(AFMT_U24_LE) && defined(AFMT_U24_BE)
+#define AFMT_U24_NE MP_SELECT_LE_BE(AFMT_U24_LE, AFMT_U24_BE)
+#endif
+
+#if !defined(AFMT_S24_NE) && defined(AFMT_S24_LE) && defined(AFMT_S24_BE)
+#define AFMT_S24_NE MP_SELECT_LE_BE(AFMT_S24_LE, AFMT_S24_BE)
+#endif
+
+#if !defined(AFMT_U32_NE) && defined(AFMT_U32_LE) && defined(AFMT_U32_BE)
+#define AFMT_U32MP_SELECT_LE_BE(AFMT_U32_LE, AFMT_U32_BE)
+#endif
+
+#if !defined(AFMT_S32_NE) && defined(AFMT_S32_LE) && defined(AFMT_S32_BE)
+#define AFMT_S32MP_SELECT_LE_BE(AFMT_S32_LE, AFMT_S32_BE)
+#endif
+
 static const int format_table[][2] = {
     {AFMT_U8,           AF_FORMAT_U8},
     {AFMT_S8,           AF_FORMAT_S8},
-    {AFMT_U16_LE,       AF_FORMAT_U16_LE},
-    {AFMT_U16_BE,       AF_FORMAT_U16_BE},
-    {AFMT_S16_LE,       AF_FORMAT_S16_LE},
-    {AFMT_S16_BE,       AF_FORMAT_S16_BE},
-#ifdef AFMT_S24_PACKED
-    {AFMT_S24_PACKED,   AF_FORMAT_S24_LE},
+    {AFMT_U16_NE,       AF_FORMAT_U16},
+    {AFMT_S16_NE,       AF_FORMAT_S16},
+#ifdef AFMT_U24_NE
+    {AFMT_U24_NE,       AF_FORMAT_U24},
 #endif
-#ifdef AFMT_U24_LE
-    {AFMT_U24_LE,       AF_FORMAT_U24_LE},
+#ifdef AFMT_S24_NE
+    {AFMT_S24_NE,       AF_FORMAT_S24},
 #endif
-#ifdef AFMT_U24_BE
-    {AFMT_U24_BE,       AF_FORMAT_U24_BE},
+#ifdef AFMT_U32_NE
+    {AFMT_U32_NE,       AF_FORMAT_U32},
 #endif
-#ifdef AFMT_S24_LE
-    {AFMT_S24_LE,       AF_FORMAT_S24_LE},
-#endif
-#ifdef AFMT_S24_BE
-    {AFMT_S24_BE,       AF_FORMAT_S24_BE},
-#endif
-#ifdef AFMT_U32_LE
-    {AFMT_U32_LE,       AF_FORMAT_U32_LE},
-#endif
-#ifdef AFMT_U32_BE
-    {AFMT_U32_BE,       AF_FORMAT_U32_BE},
-#endif
-#ifdef AFMT_S32_LE
-    {AFMT_S32_LE,       AF_FORMAT_S32_LE},
-#endif
-#ifdef AFMT_S32_BE
-    {AFMT_S32_BE,       AF_FORMAT_S32_BE},
+#ifdef AFMT_S32_NE
+    {AFMT_S32_NE,       AF_FORMAT_S32},
 #endif
 #ifdef AFMT_FLOAT
     {AFMT_FLOAT,        AF_FORMAT_FLOAT},
 #endif
-    // SPECIALS
 #ifdef AFMT_MPEG
-    {AFMT_MPEG,         AF_FORMAT_MPEG2},
-#endif
-#ifdef AFMT_AC3
-    {AFMT_AC3,          AF_FORMAT_AC3},
+    {AFMT_MPEG,         AF_FORMAT_S_MP3},
 #endif
     {-1, -1}
 };
 
+#ifndef AFMT_AC3
+#define AFMT_AC3 -1
+#endif
+
 static int format2oss(int format)
 {
     for (int n = 0; format_table[n][0] != -1; n++) {
@@ -142,7 +159,7 @@ static int oss2format(int format)
         if (format_table[n][0] == format)
             return format_table[n][1];
     }
-    return -1;
+    return 0;
 }
 
 
@@ -177,8 +194,7 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
     struct priv *p = ao->priv;
     switch (cmd) {
     case AOCONTROL_GET_VOLUME:
-    case AOCONTROL_SET_VOLUME:
-    {
+    case AOCONTROL_SET_VOLUME: {
         ao_control_vol_t *vol = (ao_control_vol_t *)arg;
         int fd, v, devs;
 
@@ -188,7 +204,7 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
             return CONTROL_OK;
 #endif
 
-        if (AF_FORMAT_IS_AC3(ao->format))
+        if (AF_FORMAT_IS_SPECIAL(ao->format))
             return CONTROL_TRUE;
 
         if ((fd = open(p->oss_mixer_device, O_RDONLY)) != -1) {
@@ -209,22 +225,203 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
             close(fd);
             return CONTROL_OK;
         }
-    }
         return CONTROL_ERROR;
     }
+#ifdef SNDCTL_DSP_GETPLAYVOL
+    case AOCONTROL_HAS_SOFT_VOLUME:
+        return CONTROL_TRUE;
+#endif
+    }
     return CONTROL_UNKNOWN;
 }
 
+// 1: ok, 0: not writable, -1: error
+static int device_writable(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    struct pollfd fd = {.fd = p->audio_fd, .events = POLLOUT};
+    return poll(&fd, 1, 0);
+}
+
+static void close_device(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    p->device_failed = false;
+    if (p->audio_fd == -1)
+        return;
+#if defined(SNDCTL_DSP_RESET)
+    ioctl(p->audio_fd, SNDCTL_DSP_RESET, NULL);
+#endif
+    close(p->audio_fd);
+    p->audio_fd = -1;
+}
+
+// close audio device
+static void uninit(struct ao *ao)
+{
+    close_device(ao);
+}
+
+static bool try_format(struct ao *ao, int *format)
+{
+    struct priv *p = ao->priv;
+
+    int oss_format = format2oss(*format);
+    if (oss_format == -1 && AF_FORMAT_IS_IEC61937(*format))
+        oss_format = AFMT_AC3;
+
+    if (oss_format == -1) {
+        MP_VERBOSE(ao, "Unknown/not supported internal format: %s\n",
+                   af_fmt_to_str(*format));
+        *format = 0;
+        return false;
+    }
+
+    int actual_format = oss_format;
+    if (ioctl(p->audio_fd, SNDCTL_DSP_SETFMT, &actual_format) < 0)
+        actual_format = -1;
+
+    if (actual_format == oss_format)
+        return true;
+
+    MP_WARN(ao, "Can't set audio device to %s output.\n", af_fmt_to_str(*format));
+    *format = oss2format(actual_format);
+    if (actual_format != -1 && !*format)
+        MP_ERR(ao, "Unknown/Unsupported OSS format: 0x%x.\n", actual_format);
+    return false;
+}
+
+static int reopen_device(struct ao *ao, bool allow_format_changes)
+{
+    struct priv *p = ao->priv;
+
+    int samplerate = ao->samplerate;
+    int format = ao->format;
+    struct mp_chmap channels = ao->channels;
+
+#ifdef __linux__
+    p->audio_fd = open(p->dsp, O_WRONLY | O_NONBLOCK);
+#else
+    p->audio_fd = open(p->dsp, O_WRONLY);
+#endif
+    if (p->audio_fd < 0) {
+        MP_ERR(ao, "Can't open audio device %s: %s\n", p->dsp, strerror(errno));
+        goto fail;
+    }
+
+#ifdef __linux__
+    /* Remove the non-blocking flag */
+    if (fcntl(p->audio_fd, F_SETFL, 0) < 0) {
+        MP_ERR(ao, "Can't make file descriptor blocking: %s\n",  strerror(errno));
+        goto fail;
+    }
+#endif
+
+#if defined(FD_CLOEXEC) && defined(F_SETFD)
+    fcntl(p->audio_fd, F_SETFD, FD_CLOEXEC);
+#endif
+
+    if (AF_FORMAT_IS_IEC61937(format)) {
+        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &samplerate);
+        // Probably could be fixed by setting number of channels; needs testing.
+        if (channels.num != 2) {
+            MP_ERR(ao, "Format %s not implemented.\n", af_fmt_to_str(format));
+            goto fail;
+        }
+    }
+
+    int try_formats[] = {format, AF_FORMAT_S32, AF_FORMAT_S24, AF_FORMAT_S16, 0};
+    for (int n = 0; try_formats[n]; n++) {
+        format = try_formats[n];
+        if (try_format(ao, &format))
+            break;
+    }
+
+    if (!format) {
+        MP_ERR(ao, "Can't set sample format.\n");
+        goto fail;
+    }
+
+    MP_VERBOSE(ao, "sample format: %s\n", af_fmt_to_str(format));
+
+    if (!AF_FORMAT_IS_IEC61937(format)) {
+        struct mp_chmap_sel sel = {0};
+        for (int n = 0; n < MP_NUM_CHANNELS + 1; n++)
+            mp_chmap_sel_add_map(&sel, &oss_layouts[n]);
+        if (!ao_chmap_sel_adjust(ao, &sel, &channels))
+            goto fail;
+        int reqchannels = channels.num;
+        // We only use SNDCTL_DSP_CHANNELS for >2 channels, in case some drivers don't have it
+        if (reqchannels > 2) {
+            int nchannels = reqchannels;
+            if (ioctl(p->audio_fd, SNDCTL_DSP_CHANNELS, &nchannels) == -1 ||
+                nchannels != reqchannels)
+            {
+                MP_ERR(ao, "Failed to set audio device to %d channels.\n",
+                       reqchannels);
+                goto fail;
+            }
+        } else {
+            int c = reqchannels - 1;
+            if (ioctl(p->audio_fd, SNDCTL_DSP_STEREO, &c) == -1) {
+                MP_ERR(ao, "Failed to set audio device to %d channels.\n",
+                       reqchannels);
+                goto fail;
+            }
+            if (!ao_chmap_sel_get_def(ao, &sel, &channels, c + 1))
+                goto fail;
+        }
+        MP_VERBOSE(ao, "using %d channels (requested: %d)\n",
+                   channels.num, reqchannels);
+        // set rate
+        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &samplerate);
+        MP_VERBOSE(ao, "using %d Hz samplerate\n", samplerate);
+    }
+
+    audio_buf_info zz = {0};
+    if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &zz) == -1) {
+        int r = 0;
+        MP_WARN(ao, "driver doesn't support SNDCTL_DSP_GETOSPACE\n");
+        if (ioctl(p->audio_fd, SNDCTL_DSP_GETBLKSIZE, &r) == -1)
+            MP_VERBOSE(ao, "%d bytes/frag (config.h)\n", p->outburst);
+        else {
+            p->outburst = r;
+            MP_VERBOSE(ao, "%d bytes/frag (GETBLKSIZE)\n", p->outburst);
+        }
+    } else {
+        MP_VERBOSE(ao, "frags: %3d/%d  (%d bytes/frag)  free: %6d\n",
+                   zz.fragments, zz.fragstotal, zz.fragsize, zz.bytes);
+        p->buffersize = zz.bytes;
+        p->outburst = zz.fragsize;
+    }
+
+    if (allow_format_changes) {
+        ao->format = format;
+        ao->samplerate = samplerate;
+        ao->channels = channels;
+    } else {
+        if (format != ao->format || samplerate != ao->samplerate ||
+            !mp_chmap_equals(&channels, &ao->channels))
+        {
+            MP_ERR(ao, "Could not reselect previous audio format.\n");
+            goto fail;
+        }
+    }
+
+    p->outburst -= p->outburst % (channels.num * af_fmt2bps(format)); // round down
+
+    return 0;
+
+fail:
+    close_device(ao);
+    return -1;
+}
+
 // open & setup audio device
 // return: 0=success -1=fail
 static int init(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    int oss_format;
-
-#ifdef SNDCTL_DSP_GETPLAYVOL
-    ao->no_persistent_volume = true;
-#endif
 
     const char *mchan = NULL;
     if (p->cfg_oss_mixer_channel && p->cfg_oss_mixer_channel[0])
@@ -265,178 +462,36 @@ static int init(struct ao *ao)
     MP_VERBOSE(ao, "using '%s' mixer device\n", p->oss_mixer_device);
     MP_VERBOSE(ao, "using '%s' mixer device\n", mixer_channels[p->oss_mixer_channel]);
 
-#ifdef __linux__
-    p->audio_fd = open(p->dsp, O_WRONLY | O_NONBLOCK);
-#else
-    p->audio_fd = open(p->dsp, O_WRONLY);
-#endif
-    if (p->audio_fd < 0) {
-        MP_ERR(ao, "Can't open audio device %s: %s\n", p->dsp, strerror(errno));
-        return -1;
-    }
-
-#ifdef __linux__
-    /* Remove the non-blocking flag */
-    if (fcntl(p->audio_fd, F_SETFL, 0) < 0) {
-        MP_ERR(ao, "Can't make file descriptor blocking: %s\n",  strerror(errno));
-        return -1;
-    }
-#endif
-
-#if defined(FD_CLOEXEC) && defined(F_SETFD)
-    fcntl(p->audio_fd, F_SETFD, FD_CLOEXEC);
-#endif
-
     ao->format = af_fmt_from_planar(ao->format);
 
-    if (AF_FORMAT_IS_AC3(ao->format)) {
-        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &ao->samplerate);
-    }
-
-ac3_retry:
-    if (AF_FORMAT_IS_AC3(ao->format))
-        ao->format = AF_FORMAT_AC3;
-    oss_format = format2oss(ao->format);
-    if (oss_format == -1) {
-        MP_VERBOSE(ao, "Unknown/not supported internal format: %s\n",
-                   af_fmt_to_str(ao->format));
-#if defined(AFMT_S32_LE) && defined(AFMT_S32_BE)
-#if BYTE_ORDER == BIG_ENDIAN
-        oss_format = AFMT_S32_BE;
-#else
-        oss_format = AFMT_S32_LE;
-#endif
-        ao->format = AF_FORMAT_S32;
-#elif defined(AFMT_S24_LE) && defined(AFMT_S24_BE)
-#if BYTE_ORDER == BIG_ENDIAN
-        oss_format = AFMT_S24_BE;
-#else
-        oss_format = AFMT_S24_LE;
-#endif
-        ao->format = AF_FORMAT_S24;
-#else
-#if BYTE_ORDER == BIG_ENDIAN
-        oss_format = AFMT_S16_BE;
-#else
-        oss_format = AFMT_S16_LE;
-#endif
-        ao->format = AF_FORMAT_S16;
-#endif
-    }
-    if (ioctl(p->audio_fd, SNDCTL_DSP_SETFMT, &oss_format) < 0 ||
-        oss_format != format2oss(ao->format))
-    {
-        MP_WARN(ao, "Can't set audio device %s to %s output, trying %s...\n",
-                p->dsp, af_fmt_to_str(ao->format),
-                af_fmt_to_str(AF_FORMAT_S16));
-        ao->format = AF_FORMAT_S16;
-        goto ac3_retry;
-    }
-
-    ao->format = oss2format(oss_format);
-    if (ao->format == -1) {
-        MP_ERR(ao, "Unknown/Unsupported OSS format: %x.\n", oss_format);
-        return -1;
-    }
-
-    MP_VERBOSE(ao, "sample format: %s\n", af_fmt_to_str(ao->format));
-
-    if (!AF_FORMAT_IS_AC3(ao->format)) {
-        struct mp_chmap_sel sel = {0};
-        for (int n = 0; n < MP_NUM_CHANNELS + 1; n++)
-            mp_chmap_sel_add_map(&sel, &oss_layouts[n]);
-        if (!ao_chmap_sel_adjust(ao, &sel, &ao->channels))
-            return -1;
-        int reqchannels = ao->channels.num;
-        // We only use SNDCTL_DSP_CHANNELS for >2 channels, in case some drivers don't have it
-        if (reqchannels > 2) {
-            int nchannels = reqchannels;
-            if (ioctl(p->audio_fd, SNDCTL_DSP_CHANNELS, &nchannels) == -1 ||
-                nchannels != reqchannels)
-            {
-                MP_ERR(ao, "Failed to set audio device to %d channels.\n",
-                       reqchannels);
-                return -1;
-            }
-        } else {
-            int c = reqchannels - 1;
-            if (ioctl(p->audio_fd, SNDCTL_DSP_STEREO, &c) == -1) {
-                MP_ERR(ao, "Failed to set audio device to %d channels.\n",
-                       reqchannels);
-                return -1;
-            }
-            if (!ao_chmap_sel_get_def(ao, &sel, &ao->channels, c + 1))
-                return -1;
-        }
-        MP_VERBOSE(ao, "using %d channels (requested: %d)\n",
-                   ao->channels.num, reqchannels);
-        // set rate
-        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &ao->samplerate);
-        MP_VERBOSE(ao, "using %d Hz samplerate\n", ao->samplerate);
-    }
-
-    if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &p->zz) == -1) {
-        int r = 0;
-        MP_WARN(ao, "driver doesn't support SNDCTL_DSP_GETOSPACE\n");
-        if (ioctl(p->audio_fd, SNDCTL_DSP_GETBLKSIZE, &r) == -1)
-            MP_VERBOSE(ao, "%d bytes/frag (config.h)\n", p->outburst);
-        else {
-            p->outburst = r;
-            MP_VERBOSE(ao, "%d bytes/frag (GETBLKSIZE)\n", p->outburst);
-        }
-    } else {
-        MP_VERBOSE(ao, "frags: %3d/%d  (%d bytes/frag)  free: %6d\n",
-                   p->zz.fragments, p->zz.fragstotal, p->zz.fragsize, p->zz.bytes);
-        p->buffersize = p->zz.bytes;
-        p->outburst = p->zz.fragsize;
-    }
+    if (reopen_device(ao, true) < 0)
+        goto fail;
 
     if (p->buffersize == -1) {
         // Measuring buffer size:
-        void *data;
+        void *data = malloc(p->outburst);
+        if (!data) {
+            MP_ERR(ao, "Out of memory, or broken outburst size.\n");
+            goto fail;
+        }
         p->buffersize = 0;
-#if HAVE_AUDIO_SELECT
-        data = malloc(p->outburst);
         memset(data, 0, p->outburst);
-        while (p->buffersize < 0x40000) {
-            fd_set rfds;
-            struct timeval tv;
-            FD_ZERO(&rfds);
-            FD_SET(p->audio_fd, &rfds);
-            tv.tv_sec = 0;
-            tv.tv_usec = 0;
-            if (!select(p->audio_fd + 1, NULL, &rfds, NULL, &tv))
-                break;
+        while (p->buffersize < 0x40000 && device_writable(ao) > 0) {
             write(p->audio_fd, data, p->outburst);
             p->buffersize += p->outburst;
         }
         free(data);
         if (p->buffersize == 0) {
-            MP_ERR(ao, "***  Your audio driver DOES NOT support select()  ***\n");
-            MP_ERR(ao, "Recompile mpv with #define HAVE_AUDIO_SELECT 0 in config.h!\n");
-            return -1;
+            MP_ERR(ao, "Your OSS audio driver DOES NOT support poll().\n");
+            goto fail;
         }
-#endif
     }
 
-    ao->bps = ao->channels.num * af_fmt2bps(ao->format);
-    p->outburst -= p->outburst % ao->bps; // round down
-    ao->bps *= ao->samplerate;
-
     return 0;
-}
 
-// close audio device
-static void uninit(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    if (p->audio_fd == -1)
-        return;
-#ifdef SNDCTL_DSP_RESET
-    ioctl(p->audio_fd, SNDCTL_DSP_RESET, NULL);
-#endif
-    close(p->audio_fd);
-    p->audio_fd = -1;
+fail:
+    uninit(ao);
+    return -1;
 }
 
 static void drain(struct ao *ao)
@@ -449,104 +504,39 @@ static void drain(struct ao *ao)
 #endif
 }
 
-#ifndef SNDCTL_DSP_RESET
-static void close_device(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    close(p->audio_fd);
-    p->audio_fd = -1;
-}
-#endif
-
 // stop playing and empty buffers (for seeking/pause)
 static void reset(struct ao *ao)
 {
+#if KEEP_DEVICE
     struct priv *p = ao->priv;
-    int oss_format;
-#ifdef SNDCTL_DSP_RESET
-    ioctl(p->audio_fd, SNDCTL_DSP_RESET, NULL);
-#else
-    close_device(ao);
-    p->audio_fd = open(p->dsp, O_WRONLY);
-    if (p->audio_fd < 0) {
-        MP_ERR(ao, "Fatal error: *** CANNOT "
-               "RE-OPEN / RESET AUDIO DEVICE *** %s\n", strerror(errno));
-        return;
-    }
-
-#if defined(FD_CLOEXEC) && defined(F_SETFD)
-    fcntl(p->audio_fd, F_SETFD, FD_CLOEXEC);
-#endif
-#endif
-
-    oss_format = format2oss(ao->format);
-    if (AF_FORMAT_IS_AC3(ao->format))
-        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &ao->samplerate);
-    ioctl(p->audio_fd, SNDCTL_DSP_SETFMT, &oss_format);
-    if (!AF_FORMAT_IS_AC3(ao->format)) {
-        int c = ao->channels.num;
-        if (ao->channels.num > 2)
-            ioctl(p->audio_fd, SNDCTL_DSP_CHANNELS, &c);
-        else {
-            c--;
-            ioctl(p->audio_fd, SNDCTL_DSP_STEREO, &c);
-        }
-        ioctl(p->audio_fd, SNDCTL_DSP_SPEED, &ao->samplerate);
-    }
-}
-
-// return: how many bytes can be played without blocking
-static int get_space(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    int playsize = p->outburst;
-
-#ifdef SNDCTL_DSP_GETOSPACE
-    if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &p->zz) != -1) {
-        // calculate exact buffer space:
-        playsize = p->zz.fragments * p->zz.fragsize;
-        return playsize / ao->sstride;
-    }
-#endif
-
-    // check buffer
-#if HAVE_AUDIO_SELECT
-    {
-        fd_set rfds;
-        struct timeval tv;
-        FD_ZERO(&rfds);
-        FD_SET(p->audio_fd, &rfds);
-        tv.tv_sec = 0;
-        tv.tv_usec = 0;
-        if (!select(p->audio_fd + 1, NULL, &rfds, NULL, &tv))
-            return 0;                                            // not block!
-    }
-#endif
-
-    return p->outburst / ao->sstride;
-}
-
-// stop playing, keep buffers (for pause)
-static void audio_pause(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-    p->prepause_space = get_space(ao) * ao->sstride;
-#ifdef SNDCTL_DSP_RESET
     ioctl(p->audio_fd, SNDCTL_DSP_RESET, NULL);
 #else
     close_device(ao);
 #endif
 }
 
-// plays 'len' bytes of 'data'
+// plays 'len' samples of 'data'
 // it should round it down to outburst*n
-// return: number of bytes played
+// return: number of samples played
 static int play(struct ao *ao, void **data, int samples, int flags)
 {
     struct priv *p = ao->priv;
     int len = samples * ao->sstride;
     if (len == 0)
         return len;
+
+    if (p->audio_fd < 0 && !p->device_failed && reopen_device(ao, false) < 0)
+        MP_ERR(ao, "Fatal error: *** CANNOT RE-OPEN / RESET AUDIO DEVICE ***\n");
+    if (p->audio_fd < 0) {
+        // Let playback continue normally, even with a closed device.
+        p->device_failed = true;
+        double now = mp_time_sec();
+        if (p->audio_end < now)
+            p->audio_end = now;
+        p->audio_end += samples / (double)ao->samplerate;
+        return samples;
+    }
+
     if (len > p->outburst || !(flags & AOPLAY_FINAL_CHUNK)) {
         len /= p->outburst;
         len *= p->outburst;
@@ -555,22 +545,16 @@ static int play(struct ao *ao, void **data, int samples, int flags)
     return len / ao->sstride;
 }
 
-// resume playing, after audio_pause()
-static void audio_resume(struct ao *ao)
-{
-    struct priv *p = ao->priv;
-#ifndef SNDCTL_DSP_RESET
-    reset(ao);
-#endif
-    int fillframes = get_space(ao) - p->prepause_space / ao->sstride;
-    if (fillframes > 0)
-        ao_play_silence(ao, fillframes);
-}
-
 // return: delay in seconds between first and last sample in buffer
 static float get_delay(struct ao *ao)
 {
     struct priv *p = ao->priv;
+    if (p->audio_fd < 0) {
+        double rest = p->audio_end - mp_time_sec();
+        if (rest > 0)
+            return rest;
+        return 0;
+    }
     /* Calculate how many bytes/second is sent out */
     if (p->audio_delay_method == 2) {
 #ifdef SNDCTL_DSP_GETODELAY
@@ -581,16 +565,57 @@ static float get_delay(struct ao *ao)
         p->audio_delay_method = 1; // fallback if not supported
     }
     if (p->audio_delay_method == 1) {
-        // SNDCTL_DSP_GETOSPACE
-        if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &p->zz) != -1) {
-            return ((float)(p->buffersize -
-                            p->zz.bytes)) / (float)ao->bps;
+        audio_buf_info zz = {0};
+        if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &zz) != -1) {
+            return ((float)(p->buffersize - zz.bytes)) / (float)ao->bps;
         }
         p->audio_delay_method = 0; // fallback if not supported
     }
     return ((float)p->buffersize) / (float)ao->bps;
 }
 
+
+// return: how many samples can be played without blocking
+static int get_space(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+
+    audio_buf_info zz = {0};
+    if (ioctl(p->audio_fd, SNDCTL_DSP_GETOSPACE, &zz) != -1) {
+        // calculate exact buffer space:
+        return zz.fragments * zz.fragsize / ao->sstride;
+    }
+
+    if (p->audio_fd < 0 && p->device_failed && get_delay(ao) > 0.2)
+        return 0;
+
+    if (p->audio_fd < 0 || device_writable(ao) > 0)
+        return p->outburst / ao->sstride;
+
+    return 0;
+}
+
+// stop playing, keep buffers (for pause)
+static void audio_pause(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    p->prepause_samples = get_delay(ao) * ao->samplerate;
+#if KEEP_DEVICE
+    ioctl(p->audio_fd, SNDCTL_DSP_RESET, NULL);
+#else
+    close_device(ao);
+#endif
+}
+
+// resume playing, after audio_pause()
+static void audio_resume(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    p->audio_end = 0;
+    if (p->prepause_samples > 0)
+        ao_play_silence(ao, p->prepause_samples);
+}
+
 #define OPT_BASE_STRUCT struct priv
 
 const struct ao_driver audio_out_oss = {

audio/out/ao_pcm.c

@@ -35,6 +35,7 @@
 #include "ao.h"
 #include "internal.h"
 #include "common/msg.h"
+#include "osdep/endian.h"
 
 #ifdef __MINGW32__
 // for GetFileType to detect pipes
@@ -72,8 +73,7 @@ static void fput32le(uint32_t val, FILE *fp)
 static void write_wave_header(struct ao *ao, FILE *fp, uint64_t data_length)
 {
     bool use_waveex = true;
-    uint16_t fmt = ao->format == AF_FORMAT_FLOAT_LE ?
-        WAV_ID_FLOAT_PCM : WAV_ID_PCM;
+    uint16_t fmt = ao->format == AF_FORMAT_FLOAT ? WAV_ID_FLOAT_PCM : WAV_ID_PCM;
     uint32_t fmt_chunk_size = use_waveex ? 40 : 16;
     int bits = af_fmt2bits(ao->format);
 
@@ -124,16 +124,23 @@ static int init(struct ao *ao)
     if (priv->waveheader) {
         // WAV files must have one of the following formats
 
+        // And they don't work in big endian; fixing it would be simple, but
+        // nobody cares.
+        if (BYTE_ORDER == BIG_ENDIAN) {
+            MP_FATAL(ao, "Not supported on big endian.\n");
+            return -1;
+        }
+
         switch (ao->format) {
         case AF_FORMAT_U8:
-        case AF_FORMAT_S16_LE:
-        case AF_FORMAT_S24_LE:
-        case AF_FORMAT_S32_LE:
-        case AF_FORMAT_FLOAT_LE:
-        case AF_FORMAT_AC3_LE:
+        case AF_FORMAT_S16:
+        case AF_FORMAT_S24:
+        case AF_FORMAT_S32:
+        case AF_FORMAT_FLOAT:
              break;
         default:
-            ao->format = AF_FORMAT_S16_LE;
+            if (!AF_FORMAT_IS_IEC61937(ao->format))
+                ao->format = AF_FORMAT_S16;
             break;
         }
     }
@@ -203,7 +210,7 @@ static int play(struct ao *ao, void **data, int samples, int flags)
 
     fwrite(data[0], len, 1, priv->fp);
     priv->data_length += len;
-    return len / ao->sstride;
+    return samples;
 }
 
 #define OPT_BASE_STRUCT struct priv

audio/out/ao_pulse.c

@@ -53,7 +53,6 @@ struct priv {
     // temporary during control()
     struct pa_sink_input_info pi;
 
-    bool broken_pause;
     int retval;
 
     // for wakeup handling
@@ -64,6 +63,7 @@ struct priv {
     char *cfg_host;
     char *cfg_sink;
     int cfg_buffer;
+    int cfg_latency_hacks;
 };
 
 #define GENERIC_ERR_MSG(str) \
@@ -113,7 +113,7 @@ static void stream_request_cb(pa_stream *s, size_t length, void *userdata)
     pa_threaded_mainloop_signal(priv->mainloop, 0);
 }
 
-static int wait(struct ao *ao, pthread_mutex_t *lock)
+static int wait_audio(struct ao *ao, pthread_mutex_t *lock)
 {
     struct priv *priv = ao->priv;
     // We don't use this mutex, because pulse like to call stream_request_cb
@@ -169,16 +169,31 @@ static const struct format_map {
     int mp_format;
     pa_sample_format_t pa_format;
 } format_maps[] = {
-    {AF_FORMAT_S16_LE, PA_SAMPLE_S16LE},
-    {AF_FORMAT_S16_BE, PA_SAMPLE_S16BE},
-    {AF_FORMAT_S32_LE, PA_SAMPLE_S32LE},
-    {AF_FORMAT_S32_BE, PA_SAMPLE_S32BE},
-    {AF_FORMAT_FLOAT_LE, PA_SAMPLE_FLOAT32LE},
-    {AF_FORMAT_FLOAT_BE, PA_SAMPLE_FLOAT32BE},
+    {AF_FORMAT_S16, PA_SAMPLE_S16NE},
+    {AF_FORMAT_S32, PA_SAMPLE_S32NE},
+    {AF_FORMAT_FLOAT, PA_SAMPLE_FLOAT32NE},
     {AF_FORMAT_U8, PA_SAMPLE_U8},
     {AF_FORMAT_UNKNOWN, 0}
 };
 
+static pa_encoding_t map_digital_format(int format)
+{
+    switch (format) {
+    case AF_FORMAT_S_AC3:   return PA_ENCODING_AC3_IEC61937;
+    case AF_FORMAT_S_EAC3:  return PA_ENCODING_EAC3_IEC61937;
+    case AF_FORMAT_S_MP3:   return PA_ENCODING_MPEG_IEC61937;
+    case AF_FORMAT_S_DTS:
+    case AF_FORMAT_S_DTSHD: return PA_ENCODING_DTS_IEC61937;
+#ifdef PA_ENCODING_MPEG2_AAC_IEC61937
+    case AF_FORMAT_S_AAC:   return PA_ENCODING_MPEG2_AAC_IEC61937;
+#endif
+    default:
+        if (AF_FORMAT_IS_IEC61937(format))
+            return PA_ENCODING_ANY;
+        return PA_ENCODING_PCM;
+    }
+}
+
 static const int speaker_map[][2] = {
   {PA_CHANNEL_POSITION_FRONT_LEFT,              MP_SPEAKER_ID_FL},
   {PA_CHANNEL_POSITION_FRONT_RIGHT,             MP_SPEAKER_ID_FR},
@@ -274,32 +289,16 @@ static void uninit(struct ao *ao)
 
 static int init(struct ao *ao)
 {
-    struct pa_sample_spec ss;
     struct pa_channel_map map;
     pa_proplist *proplist = NULL;
+    pa_format_info *format = NULL;
     struct priv *priv = ao->priv;
     char *host = priv->cfg_host && priv->cfg_host[0] ? priv->cfg_host : NULL;
     char *sink = priv->cfg_sink && priv->cfg_sink[0] ? priv->cfg_sink : NULL;
-    const char *version = pa_get_library_version();
 
     pthread_mutex_init(&priv->wakeup_lock, NULL);
     pthread_cond_init(&priv->wakeup, NULL);
 
-    ao->per_application_mixer = true;
-
-    priv->broken_pause = false;
-    /* not sure which versions are affected, assume 0.9.11* to 0.9.14*
-     * known bad: 0.9.14, 0.9.13
-     * known good: 0.9.9, 0.9.10, 0.9.15
-     * To test: pause, wait ca. 5 seconds, framestep and see if MPlayer
-     * hangs somewhen. */
-    if (strncmp(version, "0.9.1", 5) == 0 && version[5] >= '1'
-        && version[5] <= '4') {
-        MP_WARN(ao, "working around probably broken pause functionality,\n"
-                    "        see http://www.pulseaudio.org/ticket/440\n");
-        priv->broken_pause = true;
-    }
-
     if (!(priv->mainloop = pa_threaded_mainloop_new())) {
         MP_ERR(ao, "Failed to allocate main loop\n");
         goto fail;
@@ -311,6 +310,12 @@ static int init(struct ao *ao)
         goto fail;
     }
 
+    MP_VERBOSE(ao, "Library version: %s\n", pa_get_library_version());
+    MP_VERBOSE(ao, "Proto: %lu\n",
+        (long)pa_context_get_protocol_version(priv->context));
+    MP_VERBOSE(ao, "Server proto: %lu\n",
+        (long)pa_context_get_server_protocol_version(priv->context));
+
     pa_context_set_state_callback(priv->context, context_state_cb, ao);
 
     if (pa_context_connect(priv->context, host, 0, NULL) < 0)
@@ -327,26 +332,26 @@ static int init(struct ao *ao)
     if (pa_context_get_state(priv->context) != PA_CONTEXT_READY)
         goto unlock_and_fail;
 
-    ss.channels = ao->channels.num;
-    ss.rate = ao->samplerate;
+    if (!(format = pa_format_info_new()))
+        goto unlock_and_fail;
 
     ao->format = af_fmt_from_planar(ao->format);
 
-    const struct format_map *fmt_map = format_maps;
-    while (fmt_map->mp_format != ao->format) {
-        if (fmt_map->mp_format == AF_FORMAT_UNKNOWN) {
-            MP_VERBOSE(ao, "Unsupported format, using default\n");
-            fmt_map = format_maps;
-            break;
-        }
-        fmt_map++;
-    }
-    ao->format = fmt_map->mp_format;
-    ss.format = fmt_map->pa_format;
+    format->encoding = map_digital_format(ao->format);
+    if (format->encoding == PA_ENCODING_PCM) {
+        const struct format_map *fmt_map = format_maps;
 
-    if (!pa_sample_spec_valid(&ss)) {
-        MP_ERR(ao, "Invalid sample spec\n");
-        goto unlock_and_fail;
+        while (fmt_map->mp_format != ao->format) {
+            if (fmt_map->mp_format == AF_FORMAT_UNKNOWN) {
+                MP_VERBOSE(ao, "Unsupported format, using default\n");
+                fmt_map = format_maps;
+                break;
+            }
+            fmt_map++;
+        }
+        ao->format = fmt_map->mp_format;
+
+        pa_format_info_set_sample_format(format, fmt_map->pa_format);
     }
 
     if (!select_chmap(ao, &map))
@@ -357,10 +362,20 @@ static int init(struct ao *ao)
         goto unlock_and_fail;
     }
     (void)pa_proplist_sets(proplist, PA_PROP_MEDIA_ROLE, "video");
+    (void)pa_proplist_sets(proplist, PA_PROP_MEDIA_ICON_NAME,
+                           PULSE_CLIENT_NAME);
 
-    if (!(priv->stream = pa_stream_new_with_proplist(priv->context,
-                                                     "audio stream", &ss,
-                                                     &map, proplist)))
+    pa_format_info_set_rate(format, ao->samplerate);
+    pa_format_info_set_channels(format, ao->channels.num);
+    pa_format_info_set_channel_map(format, &map);
+
+    if (!pa_format_info_valid(format)) {
+        MP_ERR(ao, "Invalid audio format\n");
+        goto unlock_and_fail;
+    }
+
+    if (!(priv->stream = pa_stream_new_extended(priv->context, "audio stream",
+                                                &format, 1, proplist)))
         goto unlock_and_fail;
 
     pa_proplist_free(proplist);
@@ -370,23 +385,38 @@ static int init(struct ao *ao)
     pa_stream_set_write_callback(priv->stream, stream_request_cb, ao);
     pa_stream_set_latency_update_callback(priv->stream,
                                           stream_latency_update_cb, ao);
+    int buf_size = af_fmt_seconds_to_bytes(ao->format, priv->cfg_buffer / 1000.0,
+                                           ao->channels.num, ao->samplerate);
     pa_buffer_attr bufattr = {
         .maxlength = -1,
-        .tlength = priv->cfg_buffer > 0 ?
-            pa_usec_to_bytes(priv->cfg_buffer * 1000, &ss) : (uint32_t)-1,
+        .tlength = buf_size > 0 ? buf_size : (uint32_t)-1,
         .prebuf = -1,
         .minreq = -1,
         .fragsize = -1,
     };
+
+    int flags = PA_STREAM_NOT_MONOTONIC;
+    if (!priv->cfg_latency_hacks)
+        flags |= PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_AUTO_TIMING_UPDATE;
+
     if (pa_stream_connect_playback(priv->stream, sink, &bufattr,
-                                   PA_STREAM_NOT_MONOTONIC, NULL, NULL) < 0)
+                                   flags, NULL, NULL) < 0)
         goto unlock_and_fail;
 
     /* Wait until the stream is ready */
-    pa_threaded_mainloop_wait(priv->mainloop);
+    while (1) {
+        int state = pa_stream_get_state(priv->stream);
+        if (state == PA_STREAM_READY)
+            break;
+        if (!PA_STREAM_IS_GOOD(state))
+            goto unlock_and_fail;
+        pa_threaded_mainloop_wait(priv->mainloop);
+    }
 
-    if (pa_stream_get_state(priv->stream) != PA_STREAM_READY)
+    if (pa_stream_is_suspended(priv->stream)) {
+        MP_ERR(ao, "The stream is suspended. Bailing out.\n");
         goto unlock_and_fail;
+    }
 
     pa_threaded_mainloop_unlock(priv->mainloop);
 
@@ -398,6 +428,8 @@ unlock_and_fail:
         pa_threaded_mainloop_unlock(priv->mainloop);
 
 fail:
+    if (format)
+        pa_format_info_free(format);
     if (priv->context) {
         if (!(pa_context_errno(priv->context) == PA_ERR_CONNECTIONREFUSED
               && ao->probing))
@@ -463,12 +495,6 @@ static void pause(struct ao *ao)
 // Resume the audio stream by uncorking it on the server
 static void resume(struct ao *ao)
 {
-    struct priv *priv = ao->priv;
-    /* Without this, certain versions will cause an infinite hang because
-     * pa_stream_writable_size returns 0 always.
-     * Note that this workaround causes A-V desync after pause. */
-    if (priv->broken_pause)
-        reset(ao);
     cork(ao, false);
 }
 
@@ -482,8 +508,7 @@ static int get_space(struct ao *ao)
     return space / ao->sstride;
 }
 
-// Return the current latency in seconds
-static float get_delay(struct ao *ao)
+static float get_delay_hackfixed(struct ao *ao)
 {
     /* This code basically does what pa_stream_get_latency() _should_
      * do, but doesn't due to multiple known bugs in PulseAudio (at
@@ -538,6 +563,34 @@ static float get_delay(struct ao *ao)
     return latency / 1e6;
 }
 
+static float get_delay_pulse(struct ao *ao)
+{
+    struct priv *priv = ao->priv;
+    pa_usec_t latency = (pa_usec_t) -1;
+    pa_threaded_mainloop_lock(priv->mainloop);
+    while (pa_stream_get_latency(priv->stream, &latency, NULL) < 0) {
+        if (pa_context_errno(priv->context) != PA_ERR_NODATA) {
+            GENERIC_ERR_MSG("pa_stream_get_latency() failed");
+            break;
+        }
+        /* Wait until latency data is available again */
+        pa_threaded_mainloop_wait(priv->mainloop);
+    }
+    pa_threaded_mainloop_unlock(priv->mainloop);
+    return latency == (pa_usec_t) -1 ? 0 : latency / 1000000.0;
+}
+
+// Return the current latency in seconds
+static float get_delay(struct ao *ao)
+{
+    struct priv *priv = ao->priv;
+    if (priv->cfg_latency_hacks) {
+        return get_delay_hackfixed(ao);
+    } else {
+        return get_delay_pulse(ao);
+    }
+}
+
 /* A callback function that is called when the
  * pa_context_get_sink_input_info() operation completes. Saves the
  * volume field of the specified structure to the global variable volume.
@@ -567,7 +620,7 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
         pa_threaded_mainloop_lock(priv->mainloop);
         if (!waitop(priv, pa_context_get_sink_input_info(priv->context, devidx,
                                                          info_func, ao))) {
-            GENERIC_ERR_MSG("pa_stream_get_sink_input_info() failed");
+            GENERIC_ERR_MSG("pa_context_get_sink_input_info() failed");
             return CONTROL_ERROR;
         }
         // Warning: some information in pi might be unaccessible, because
@@ -630,6 +683,9 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
         return CONTROL_OK;
     }
 
+    case AOCONTROL_HAS_PER_APP_VOLUME:
+        return CONTROL_TRUE;
+
     case AOCONTROL_UPDATE_STREAM_TITLE: {
         char *title = (char *)arg;
         pa_threaded_mainloop_lock(priv->mainloop);
@@ -662,16 +718,18 @@ const struct ao_driver audio_out_pulse = {
     .pause     = pause,
     .resume    = resume,
     .drain     = drain,
-    .wait      = wait,
+    .wait      = wait_audio,
     .wakeup    = wakeup,
     .priv_size = sizeof(struct priv),
     .priv_defaults = &(const struct priv) {
         .cfg_buffer = 250,
+        .cfg_latency_hacks = 1,
     },
     .options = (const struct m_option[]) {
         OPT_STRING("host", cfg_host, 0),
         OPT_STRING("sink", cfg_sink, 0),
-        OPT_CHOICE_OR_INT("buffer", cfg_buffer, 0, 1, 2000, ({"native", -1})),
+        OPT_CHOICE_OR_INT("buffer", cfg_buffer, 0, 1, 2000, ({"native", 0})),
+        OPT_FLAG("latency-hacks", cfg_latency_hacks, 0),
         {0}
     },
 };

audio/out/ao_rsound.c

@@ -52,40 +52,26 @@ static int set_format(struct ao *ao)
     case AF_FORMAT_S8:
         rsd_format = RSD_S8;
         break;
-    case AF_FORMAT_S16_LE:
-        rsd_format = RSD_S16_LE;
+    case AF_FORMAT_S16:
+        rsd_format = RSD_S16_NE;
         break;
-    case AF_FORMAT_S16_BE:
-        rsd_format = RSD_S16_BE;
+    case AF_FORMAT_U16:
+        rsd_format = RSD_U16_NE;
         break;
-    case AF_FORMAT_U16_LE:
-        rsd_format = RSD_U16_LE;
+    case AF_FORMAT_S24:
+    case AF_FORMAT_U24:
+        rsd_format = RSD_S32_NE;
+        ao->format = AF_FORMAT_S32;
         break;
-    case AF_FORMAT_U16_BE:
-        rsd_format = RSD_U16_BE;
+    case AF_FORMAT_S32:
+        rsd_format = RSD_S32_NE;
         break;
-    case AF_FORMAT_S24_LE:
-    case AF_FORMAT_S24_BE:
-    case AF_FORMAT_U24_LE:
-    case AF_FORMAT_U24_BE:
-        rsd_format = RSD_S32_LE;
-        ao->format = AF_FORMAT_S32_LE;
-        break;
-    case AF_FORMAT_S32_LE:
-        rsd_format = RSD_S32_LE;
-        break;
-    case AF_FORMAT_S32_BE:
-        rsd_format = RSD_S32_BE;
-        break;
-    case AF_FORMAT_U32_LE:
-        rsd_format = RSD_U32_LE;
-        break;
-    case AF_FORMAT_U32_BE:
-        rsd_format = RSD_U32_BE;
+    case AF_FORMAT_U32:
+        rsd_format = RSD_U32_NE;
         break;
     default:
-        rsd_format = RSD_S16_LE;
-        ao->format = AF_FORMAT_S16_LE;
+        rsd_format = RSD_S16_NE;
+        ao->format = AF_FORMAT_S16;
     }
 
     return rsd_format;

audio/out/ao_sdl.c

@@ -38,6 +38,20 @@ struct priv
     float buflen;
 };
 
+static int fmtmap[][2] = {
+    {AF_FORMAT_U8,      AUDIO_U8},
+    {AF_FORMAT_S8,      AUDIO_S8},
+    {AF_FORMAT_U16,     AUDIO_U16SYS},
+    {AF_FORMAT_S16,     AUDIO_S16SYS},
+#ifdef AUDIO_S32SYS
+    {AF_FORMAT_S32,     AUDIO_S32SYS},
+#endif
+#ifdef AUDIO_F32SYS
+    {AF_FORMAT_FLOAT,   AUDIO_F32SYS},
+#endif
+    {0}
+};
+
 static void audio_callback(void *userdata, Uint8 *stream, int len)
 {
     struct ao *ao = userdata;
@@ -104,26 +118,12 @@ static int init(struct ao *ao)
 
     SDL_AudioSpec desired, obtained;
 
-    switch (ao->format) {
-        case AF_FORMAT_U8: desired.format = AUDIO_U8; break;
-        case AF_FORMAT_S8: desired.format = AUDIO_S8; break;
-        case AF_FORMAT_U16_LE: desired.format = AUDIO_U16LSB; break;
-        case AF_FORMAT_U16_BE: desired.format = AUDIO_U16MSB; break;
-        default:
-        case AF_FORMAT_S16_LE: desired.format = AUDIO_S16LSB; break;
-        case AF_FORMAT_S16_BE: desired.format = AUDIO_S16MSB; break;
-#ifdef AUDIO_S32LSB
-        case AF_FORMAT_S32_LE: desired.format = AUDIO_S32LSB; break;
-#endif
-#ifdef AUDIO_S32MSB
-        case AF_FORMAT_S32_BE: desired.format = AUDIO_S32MSB; break;
-#endif
-#ifdef AUDIO_F32LSB
-        case AF_FORMAT_FLOAT_LE: desired.format = AUDIO_F32LSB; break;
-#endif
-#ifdef AUDIO_F32MSB
-        case AF_FORMAT_FLOAT_BE: desired.format = AUDIO_F32MSB; break;
-#endif
+    desired.format = AUDIO_S16SYS;
+    for (int n = 0; fmtmap[n][0]; n++) {
+        if (ao->format == fmtmap[n][0]) {
+            desired.format = fmtmap[n][1];
+            break;
+        }
     }
     desired.freq = ao->samplerate;
     desired.channels = ao->channels.num;
@@ -156,30 +156,18 @@ static int init(struct ao *ao)
     // large, this will help.
     ao->device_buffer = 3 * obtained.samples;
 
-    switch (obtained.format) {
-        case AUDIO_U8: ao->format = AF_FORMAT_U8; break;
-        case AUDIO_S8: ao->format = AF_FORMAT_S8; break;
-        case AUDIO_S16LSB: ao->format = AF_FORMAT_S16_LE; break;
-        case AUDIO_S16MSB: ao->format = AF_FORMAT_S16_BE; break;
-        case AUDIO_U16LSB: ao->format = AF_FORMAT_U16_LE; break;
-        case AUDIO_U16MSB: ao->format = AF_FORMAT_U16_BE; break;
-#ifdef AUDIO_S32LSB
-        case AUDIO_S32LSB: ao->format = AF_FORMAT_S32_LE; break;
-#endif
-#ifdef AUDIO_S32MSB
-        case AUDIO_S32MSB: ao->format = AF_FORMAT_S32_BE; break;
-#endif
-#ifdef AUDIO_F32LSB
-        case AUDIO_F32LSB: ao->format = AF_FORMAT_FLOAT_LE; break;
-#endif
-#ifdef AUDIO_F32MSB
-        case AUDIO_F32MSB: ao->format = AF_FORMAT_FLOAT_BE; break;
-#endif
-        default:
-            if (!ao->probing)
-                MP_ERR(ao, "could not find matching format\n");
-            uninit(ao);
-            return -1;
+    ao->format = 0;
+    for (int n = 0; fmtmap[n][0]; n++) {
+        if (obtained.format == fmtmap[n][1]) {
+            ao->format = fmtmap[n][0];
+            break;
+        }
+    }
+    if (!ao->format) {
+        if (!ao->probing)
+            MP_ERR(ao, "could not find matching format\n");
+        uninit(ao);
+        return -1;
     }
 
     if (!ao_chmap_sel_get_def(ao, &sel, &ao->channels, obtained.channels)) {

audio/out/ao_sndio.c

@@ -33,6 +33,7 @@ struct priv {
     struct sio_hdl *hdl;
     struct sio_par par;
     int delay;
+    bool playing;
     int vol;
     int havevol;
 #define SILENCE_NMAX 0x1000
@@ -60,6 +61,8 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
             return CONTROL_FALSE;
         sio_setvol(p->hdl, vol->left * SIO_MAXVOL / 100);
         break;
+    case AOCONTROL_HAS_SOFT_VOLUME:
+        return CONTROL_TRUE;
     default:
         return CONTROL_UNKNOWN;
     }
@@ -72,7 +75,7 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
 static void movecb(void *addr, int delta)
 {
     struct priv *p = addr;
-    p->delay -= delta * (int)(p->par.bps * p->par.pchan);
+    p->delay -= delta;
 }
 
 /*
@@ -107,26 +110,20 @@ static int init(struct ao *ao)
     struct priv *p = ao->priv;
 
     struct af_to_par {
-        int format, bits, sig, le;
+        int format, bits, sig;
     } static const af_to_par[] = {
-        {AF_FORMAT_U8,      8, 0, 0},
-        {AF_FORMAT_S8,      8, 1, 0},
-        {AF_FORMAT_U16_LE, 16, 0, 1},
-        {AF_FORMAT_U16_BE, 16, 0, 0},
-        {AF_FORMAT_S16_LE, 16, 1, 1},
-        {AF_FORMAT_S16_BE, 16, 1, 0},
-        {AF_FORMAT_U24_LE, 16, 0, 1},
-        {AF_FORMAT_U24_BE, 24, 0, 0},
-        {AF_FORMAT_S24_LE, 24, 1, 1},
-        {AF_FORMAT_S24_BE, 24, 1, 0},
-        {AF_FORMAT_U32_LE, 32, 0, 1},
-        {AF_FORMAT_U32_BE, 32, 0, 0},
-        {AF_FORMAT_S32_LE, 32, 1, 1},
-        {AF_FORMAT_S32_BE, 32, 1, 0}
+        {AF_FORMAT_U8,   8, 0},
+        {AF_FORMAT_S8,   8, 1},
+        {AF_FORMAT_U16, 16, 0},
+        {AF_FORMAT_S16, 16, 1},
+        {AF_FORMAT_U24, 24, 0},
+        {AF_FORMAT_S24, 24, 1},
+        {AF_FORMAT_U32, 32, 0},
+        {AF_FORMAT_S32, 32, 1},
     }, *ap;
     int i;
 
-    p->hdl = sio_open(p->dev, SIO_PLAY, 0);
+    p->hdl = sio_open(p->dev, SIO_PLAY, 1);
     if (p->hdl == NULL) {
         MP_ERR(ao, "can't open sndio %s\n", p->dev);
         goto error;
@@ -147,7 +144,7 @@ static int init(struct ao *ao)
             p->par.bits = ap->bits;
             p->par.sig = ap->sig;
             if (ap->bits > 8)
-                p->par.le = ap->le;
+                p->par.le = SIO_LE_NATIVE;
             if (ap->bits != SIO_BPS(ap->bits))
                 p->par.bps = ap->bits / 8;
             break;
@@ -173,30 +170,25 @@ static int init(struct ao *ao)
         MP_ERR(ao, "couldn't get params\n");
         goto error;
     }
+    if (p->par.bps > 1 && p->par.le != SIO_LE_NATIVE) {
+        MP_ERR(ao, "swapped endian output not supported\n");
+        goto error;
+    }
     if (p->par.bits == 8 && p->par.bps == 1) {
         ao->format = p->par.sig ? AF_FORMAT_S8 : AF_FORMAT_U8;
     } else if (p->par.bits == 16 && p->par.bps == 2) {
-        ao->format = p->par.sig ?
-            (p->par.le ? AF_FORMAT_S16_LE : AF_FORMAT_S16_BE) :
-            (p->par.le ? AF_FORMAT_U16_LE : AF_FORMAT_U16_BE);
+        ao->format = p->par.sig ? AF_FORMAT_S16 : AF_FORMAT_U16;
     } else if ((p->par.bits == 24 || p->par.msb) && p->par.bps == 3) {
-        ao->format = p->par.sig ?
-            (p->par.le ? AF_FORMAT_S24_LE : AF_FORMAT_S24_BE) :
-            (p->par.le ? AF_FORMAT_U24_LE : AF_FORMAT_U24_BE);
+        ao->format = p->par.sig ? AF_FORMAT_S24 : AF_FORMAT_U24;
     } else if ((p->par.bits == 32 || p->par.msb) && p->par.bps == 4) {
-        ao->format = p->par.sig ?
-            (p->par.le ? AF_FORMAT_S32_LE : AF_FORMAT_S32_BE) :
-            (p->par.le ? AF_FORMAT_U32_LE : AF_FORMAT_U32_BE);
+        ao->format = p->par.sig ? AF_FORMAT_S32 : AF_FORMAT_U32;
     } else {
         MP_ERR(ao, "couldn't set format\n");
         goto error;
     }
 
-    ao->bps = p->par.bps * p->par.pchan * p->par.rate;
-    ao->no_persistent_volume = true;
     p->havevol = sio_onvol(p->hdl, volcb, p);
     sio_onmove(p->hdl, movecb, p);
-    p->delay = 0;
     if (!sio_start(p->hdl))
         MP_ERR(ao, "init: couldn't start\n");
 
@@ -233,46 +225,56 @@ static void reset(struct ao *ao)
 {
     struct priv *p = ao->priv;
 
-    if (!sio_stop(p->hdl))
-        MP_ERR(ao, "reset: couldn't stop\n");
-    p->delay = 0;
-    if (!sio_start(p->hdl))
-        MP_ERR(ao, "reset: couldn't start\n");
+    if (p->playing) {
+        MP_WARN(ao, "Blocking until remaining audio is played... (sndio design bug).\n");
+
+        p->playing = false;
+
+        if (!sio_stop(p->hdl))
+            MP_ERR(ao, "reset: couldn't stop\n");
+        p->delay = 0;
+        if (!sio_start(p->hdl))
+            MP_ERR(ao, "reset: couldn't start\n");
+    }
 }
 
 /*
- * play given number of bytes until sio_write() blocks
+ * play given number of samples until sio_write() blocks
  */
 static int play(struct ao *ao, void **data, int samples, int flags)
 {
     struct priv *p = ao->priv;
     int n;
 
-    n = sio_write(p->hdl, data[0], samples * ao->sstride);
+    n = sio_write(p->hdl, data[0], samples * ao->sstride) / ao->sstride;
     p->delay += n;
-    if (flags & AOPLAY_FINAL_CHUNK)
-        reset(ao);
-    return n / ao->sstride;
+    p->playing = true;
+    /* on AOPLAY_FINAL_CHUNK, just let it underrun */
+    return n;
 }
 
 /*
- * how many bytes can be played without blocking
+ * make libsndio call movecb()
+ */
+static void update(struct ao *ao)
+{
+    struct priv *p = ao->priv;
+    int n = sio_pollfd(p->hdl, p->pfd, POLLOUT);
+    while (poll(p->pfd, n, 0) < 0 && errno == EINTR) {}
+    sio_revents(p->hdl, p->pfd);
+}
+
+/*
+ * how many samples can be played without blocking
  */
 static int get_space(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    int n;
 
-    /*
-     * call poll() and sio_revents(), so the
-     * delay counter is updated
-     */
-    n = sio_pollfd(p->hdl, p->pfd, POLLOUT);
-    while (poll(p->pfd, n, 0) < 0 && errno == EINTR)
-        ; /* nothing */
-    sio_revents(p->hdl, p->pfd);
+    update(ao);
 
-    return (p->par.bufsz * p->par.pchan * p->par.bps - p->delay) / ao->sstride;
+    int samples = p->par.bufsz - p->delay;
+    return samples / p->par.round * p->par.round;
 }
 
 /*
@@ -281,7 +283,10 @@ static int get_space(struct ao *ao)
 static float get_delay(struct ao *ao)
 {
     struct priv *p = ao->priv;
-    return (float)p->delay / (p->par.bps * p->par.pchan * p->par.rate);
+
+    update(ao);
+
+    return p->delay / (double)p->par.rate;
 }
 
 /*
@@ -301,8 +306,8 @@ static void audio_resume(struct ao *ao)
     int n, count, todo;
 
     /*
-     * we want to start with buffers full, because mplayer uses
-     * get_space() pointer as clock, which would cause video to
+     * we want to start with buffers full, because mpv uses
+     * get_delay() as clock, which would cause video to
      * accelerate while buffers are filled.
      */
     todo = p->par.bufsz * p->par.pchan * p->par.bps;

audio/out/ao_wasapi.c

@@ -194,7 +194,6 @@ static int init(struct ao *ao)
         wasapi_enumerate_devices(state->log);
     }
 
-    ao->per_application_mixer = true;
     if (state->opt_exclusive) {
         state->share_mode = AUDCLNT_SHAREMODE_EXCLUSIVE;
     } else {
@@ -284,6 +283,8 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
             ISimpleAudioVolume_SetMute(state->pAudioVolumeProxy, mute, NULL);
 
         return CONTROL_OK;
+    case AOCONTROL_HAS_PER_APP_VOLUME:
+        return CONTROL_TRUE;
     case AOCONTROL_UPDATE_STREAM_TITLE: {
         MP_VERBOSE(state, "Updating stream title to \"%s\"\n", (char*)arg);
         wchar_t *title = mp_from_utf8(NULL, (char*)arg);

audio/out/ao_wasapi.h

@@ -28,7 +28,7 @@
 #include <mmdeviceapi.h>
 #include <avrt.h>
 
-#include "compat/atomics.h"
+#include "osdep/atomics.h"
 
 typedef struct wasapi_state {
     struct mp_log *log;

audio/out/ao_wasapi_utils.c

@@ -142,47 +142,36 @@ static void set_format(WAVEFORMATEXTENSIBLE *wformat, WORD bytepersample,
     wformat->dwChannelMask = chanmask;
 }
 
-static int format_set_bits(int old_format, int bits, int fp) {
-    int format = old_format;
-    format &= (~AF_FORMAT_BITS_MASK) & (~AF_FORMAT_POINT_MASK) & (~AF_FORMAT_SIGN_MASK);
-    format |= AF_FORMAT_SI;
-
-    switch (bits) {
-    case 32:
-        format |= AF_FORMAT_32BIT;
-        break;
-    case 24:
-        format |= AF_FORMAT_24BIT;
-        break;
-    case 16:
-        format |= AF_FORMAT_16BIT;
-        break;
-    default:
-        abort(); // (should be) unreachable
-    }
-
+static int format_set_bits(int old_format, int bits, int fp)
+{
     if (fp) {
-        format |= AF_FORMAT_F;
-    } else {
-        format |= AF_FORMAT_I;
+        switch (bits) {
+        case 64: return AF_FORMAT_DOUBLE;
+        case 32: return AF_FORMAT_FLOAT;
+        default: return 0;
+        }
     }
 
-    return format;
+    return af_fmt_change_bits(old_format, bits);
 }
 
 static int set_ao_format(struct wasapi_state *state,
                          struct ao *const ao,
-                         WAVEFORMATEXTENSIBLE wformat) {
-    // .Data1 == 1 is PCM, .Data1 == 3 is IEEE_FLOAT
-    int format = format_set_bits(ao->format,
-        wformat.Format.wBitsPerSample, wformat.SubFormat.Data1 == 3);
-
+                         WAVEFORMATEXTENSIBLE wformat)
+{
     if (wformat.SubFormat.Data1 != 1 && wformat.SubFormat.Data1 != 3) {
         MP_ERR(ao, "unknown SubFormat %"PRIu32"\n",
                (uint32_t)wformat.SubFormat.Data1);
         return 0;
     }
 
+    // .Data1 == 1 is PCM, .Data1 == 3 is IEEE_FLOAT
+    int format = format_set_bits(ao->format,
+        wformat.Format.wBitsPerSample, wformat.SubFormat.Data1 == 3);
+
+    if (!format)
+        return 0;
+
     ao->samplerate = wformat.Format.nSamplesPerSec;
     ao->bps = wformat.Format.nAvgBytesPerSec;
     ao->format = format;
@@ -204,6 +193,8 @@ static int try_format(struct wasapi_state *state,
     set_format(&wformat, bits / 8, samplerate, channels.num, mp_chmap_to_waveext(&channels));
 
     int af_format = format_set_bits(ao->format, bits, bits == 32);
+    if (!af_format)
+        return 0;
 
     MP_VERBOSE(ao, "trying %dch %s @ %dhz\n",
                channels.num, af_fmt_to_str(af_format), samplerate);
@@ -292,14 +283,13 @@ static int try_passthrough(struct wasapi_state *state,
     union WAVEFMT u;
     u.extensible = &wformat;
 
-    MP_VERBOSE(ao, "trying passthrough for %s...\n",
-               af_fmt_to_str((ao->format&~AF_FORMAT_END_MASK) | AF_FORMAT_LE));
+    MP_VERBOSE(ao, "trying passthrough for %s...\n", af_fmt_to_str(ao->format));
 
     HRESULT hr = IAudioClient_IsFormatSupported(state->pAudioClient,
                                                 state->share_mode,
                                                 u.ex, NULL);
     if (!FAILED(hr)) {
-        ao->format = (ao->format&~AF_FORMAT_END_MASK) | AF_FORMAT_LE;
+        ao->format = ao->format;
         state->format = wformat;
         return 1;
     }
@@ -310,7 +300,7 @@ static int find_formats(struct ao *const ao)
 {
     struct wasapi_state *state = (struct wasapi_state *)ao->priv;
 
-    if (AF_FORMAT_IS_IEC61937(ao->format) || ao->format == AF_FORMAT_MPEG2) {
+    if (AF_FORMAT_IS_IEC61937(ao->format)) {
         if (try_passthrough(state, ao))
             return 0;
 

audio/out/internal.h

@@ -25,14 +25,6 @@
 #include "audio/chmap.h"
 #include "audio/chmap_sel.h"
 
-// Minimum buffer size in seconds.
-#define MIN_BUFFER 0.2
-
-// If ao_get_delay() reaches this value after ao_play() was called with the
-// AOPLAY_FINAL_CHUNK flag set, the playback core expects that the audio has
-// all been played.
-#define AO_EOF_DELAY 0.05
-
 /* global data used by ao.c and ao drivers */
 struct ao {
     int samplerate;
@@ -44,8 +36,6 @@ struct ao {
     int num_planes;
     bool probing;               // if true, don't fail loudly on init
     bool untimed;               // don't assume realtime playback
-    bool no_persistent_volume;  // the AO does the equivalent of af_volume
-    bool per_application_mixer; // like above, but volume persists (per app)
     int device_buffer;          // device buffer in samples (guessed by
                                 // common init code if not set by driver)
     const struct ao_driver *api; // entrypoints to the wrapper (push.c/pull.c)
@@ -56,6 +46,7 @@ struct ao {
     struct mp_log *log; // Using e.g. "[ao/coreaudio]" as prefix
 
     int buffer;
+    double def_buffer;
     void *api_priv;
 };
 
@@ -134,6 +125,8 @@ struct ao_driver {
     float (*get_delay)(struct ao *ao);
     // push based: block until all queued audio is played (optional)
     void (*drain)(struct ao *ao);
+    // Optional. Return true if audio has stopped in any way.
+    bool (*get_eof)(struct ao *ao);
     // Wait until the audio buffer needs to be refilled. The lock is the
     // internal mutex usually protecting the internal AO state (and used to
     // protect driver calls), and must be temporarily unlocked while waiting.
@@ -158,7 +151,6 @@ struct ao_driver {
 // These functions can be called by AOs.
 
 int ao_play_silence(struct ao *ao, int samples);
-void ao_wait_drain(struct ao *ao);
 int ao_read_data(struct ao *ao, void **data, int samples, int64_t out_time_us);
 struct pollfd;
 int ao_wait_poll(struct ao *ao, struct pollfd *fds, int num_fds,

audio/out/pull.c

@@ -30,7 +30,7 @@
 
 #include "osdep/timer.h"
 #include "osdep/threads.h"
-#include "compat/atomics.h"
+#include "osdep/atomics.h"
 #include "misc/ring.h"
 
 /*
@@ -193,6 +193,22 @@ static void resume(struct ao *ao)
     ao->driver->resume(ao);
 }
 
+static void drain(struct ao *ao)
+{
+    struct ao_pull_state *p = ao->api_priv;
+    if (atomic_load(&p->state) == AO_STATE_PLAY)
+        mp_sleep_us(get_delay(ao) * 1000000);
+    reset(ao);
+}
+
+static bool get_eof(struct ao *ao)
+{
+    struct ao_pull_state *p = ao->api_priv;
+    // For simplicity, ignore the latency. Otherwise, we would have to run an
+    // extra thread to time it.
+    return mp_ring_buffered(p->buffers[0]) == 0;
+}
+
 static void uninit(struct ao *ao)
 {
     ao->driver->uninit(ao);
@@ -212,10 +228,12 @@ const struct ao_driver ao_api_pull = {
     .init = init,
     .control = control,
     .uninit = uninit,
+    .drain = drain,
     .reset = reset,
     .get_space = get_space,
     .play = play,
     .get_delay = get_delay,
+    .get_eof = get_eof,
     .pause = pause,
     .resume = resume,
     .priv_size = sizeof(struct ao_pull_state),

audio/out/push.c

@@ -35,7 +35,7 @@
 
 #include "osdep/threads.h"
 #include "osdep/timer.h"
-#include "compat/atomics.h"
+#include "osdep/atomics.h"
 
 #include "audio/audio.h"
 #include "audio/audio_buffer.h"
@@ -44,18 +44,15 @@ struct ao_push_state {
     pthread_t thread;
     pthread_mutex_t lock;
     pthread_cond_t wakeup;
-    pthread_cond_t wakeup_drain;
 
     // --- protected by lock
 
     struct mp_audio_buffer *buffer;
 
     bool terminate;
-    bool drain;
-    bool buffers_full;
-    bool avoid_ao_wait;
+    bool wait_on_ao;
+    bool still_playing;
     bool need_wakeup;
-    bool requested_data;
     bool paused;
 
     // Whether the current buffer contains the complete audio.
@@ -87,22 +84,21 @@ static int control(struct ao *ao, enum aocontrol cmd, void *arg)
     return r;
 }
 
+static double unlocked_get_delay(struct ao *ao)
+{
+    struct ao_push_state *p = ao->api_priv;
+    double driver_delay = 0;
+    if (ao->driver->get_delay)
+        driver_delay = ao->driver->get_delay(ao);
+    return driver_delay + mp_audio_buffer_seconds(p->buffer);
+}
+
 static float get_delay(struct ao *ao)
 {
     struct ao_push_state *p = ao->api_priv;
     pthread_mutex_lock(&p->lock);
-    double driver_delay = 0;
-    if (ao->driver->get_delay)
-        driver_delay = ao->driver->get_delay(ao);
-    double delay = driver_delay + mp_audio_buffer_seconds(p->buffer);
+    float delay = unlocked_get_delay(ao);
     pthread_mutex_unlock(&p->lock);
-    if (delay >= AO_EOF_DELAY && p->expected_end_time) {
-        if (mp_time_sec() > p->expected_end_time) {
-            MP_ERR(ao, "Audio device EOF reporting is broken!\n");
-            MP_ERR(ao, "Please report this problem.\n");
-            delay = 0;
-        }
-    }
     return delay;
 }
 
@@ -114,7 +110,9 @@ static void reset(struct ao *ao)
         ao->driver->reset(ao);
     mp_audio_buffer_clear(p->buffer);
     p->paused = false;
-    wakeup_playthread(ao);
+    if (p->still_playing)
+        wakeup_playthread(ao);
+    p->still_playing = false;
     pthread_mutex_unlock(&p->lock);
 }
 
@@ -145,16 +143,28 @@ static void drain(struct ao *ao)
 {
     struct ao_push_state *p = ao->api_priv;
 
+    MP_VERBOSE(ao, "draining...\n");
+
     pthread_mutex_lock(&p->lock);
+    if (p->paused)
+        goto done;
+
     p->final_chunk = true;
-    p->drain = true;
     wakeup_playthread(ao);
-    while (p->drain)
-        pthread_cond_wait(&p->wakeup_drain, &p->lock);
+    while (p->still_playing && mp_audio_buffer_samples(p->buffer) > 0)
+        pthread_cond_wait(&p->wakeup, &p->lock);
+
+    if (ao->driver->drain) {
+        ao->driver->drain(ao);
+    } else {
+        double time = unlocked_get_delay(ao);
+        mp_sleep_us(MPMIN(time, ao->buffer / (double)ao->samplerate + 1) * 1e6);
+    }
+
+done:
     pthread_mutex_unlock(&p->lock);
 
-    if (!ao->driver->drain)
-        ao_wait_drain(ao);
+    reset(ao);
 }
 
 static int unlocked_get_space(struct ao *ao)
@@ -163,13 +173,17 @@ static int unlocked_get_space(struct ao *ao)
     int space = mp_audio_buffer_get_write_available(p->buffer);
     if (ao->driver->get_space) {
         // The following code attempts to keep the total buffered audio to
-        // MIN_BUFFER/2+device_buffer in order to improve latency.
+        // ao->buffer in order to improve latency.
         int device_space = ao->driver->get_space(ao);
         int device_buffered = ao->device_buffer - device_space;
         int soft_buffered = mp_audio_buffer_samples(p->buffer);
-        int min_buffer = MIN_BUFFER / 2 * ao->samplerate + ao->device_buffer;
-        int total_buffer = device_buffered + soft_buffered;
-        int missing = min_buffer - total_buffer;
+        // The extra margin helps avoiding too many wakeups if the AO is fully
+        // byte based and doesn't do proper chunked processing.
+        int min_buffer = ao->buffer + 64;
+        int missing = min_buffer - device_buffered - soft_buffered;
+        // But always keep the device's buffer filled as much as we can.
+        int device_missing = device_space - soft_buffered;
+        missing = MPMAX(missing, device_missing);
         space = MPMIN(space, missing);
         space = MPMAX(0, space);
     }
@@ -185,6 +199,15 @@ static int get_space(struct ao *ao)
     return space;
 }
 
+static bool get_eof(struct ao *ao)
+{
+    struct ao_push_state *p = ao->api_priv;
+    pthread_mutex_lock(&p->lock);
+    bool eof = !p->still_playing;
+    pthread_mutex_unlock(&p->lock);
+    return eof;
+}
+
 static int play(struct ao *ao, void **data, int samples, int flags)
 {
     struct ao_push_state *p = ao->api_priv;
@@ -194,6 +217,8 @@ static int play(struct ao *ao, void **data, int samples, int flags)
     int write_samples = mp_audio_buffer_get_write_available(p->buffer);
     write_samples = MPMIN(write_samples, samples);
 
+    MP_TRACE(ao, "samples=%d flags=%d r=%d\n", samples, flags, write_samples);
+
     if (write_samples < samples)
         flags = flags & ~AOPLAY_FINAL_CHUNK;
     bool is_final = flags & AOPLAY_FINAL_CHUNK;
@@ -207,16 +232,14 @@ static int play(struct ao *ao, void **data, int samples, int flags)
 
     bool got_data = write_samples > 0 || p->paused || p->final_chunk != is_final;
 
-    p->expected_end_time = 0;
     p->final_chunk = is_final;
     p->paused = false;
+    p->still_playing |= write_samples > 0;
 
     // If we don't have new data, the decoder thread basically promises it
     // will send new data as soon as it's available.
-    if (got_data) {
-        p->requested_data = false;
+    if (got_data)
         wakeup_playthread(ao);
-    }
     pthread_mutex_unlock(&p->lock);
     return write_samples;
 }
@@ -246,37 +269,31 @@ static void ao_play_data(struct ao *ao)
     }
     r = MPMAX(r, 0);
     // Probably can't copy the rest of the buffer due to period alignment.
-    bool stuck = r <= 0 && space >= max && data.samples > 0;
-    if ((flags & AOPLAY_FINAL_CHUNK) && stuck) {
+    bool stuck_eof = r <= 0 && space >= max && data.samples > 0;
+    if ((flags & AOPLAY_FINAL_CHUNK) && stuck_eof) {
         MP_ERR(ao, "Audio output driver seems to ignore AOPLAY_FINAL_CHUNK.\n");
         r = max;
     }
     mp_audio_buffer_skip(p->buffer, r);
-    if (p->final_chunk && mp_audio_buffer_samples(p->buffer) == 0) {
-        p->expected_end_time = mp_time_sec() + AO_EOF_DELAY + 0.25; // + margin
-        if (ao->driver->get_delay)
-            p->expected_end_time += ao->driver->get_delay(ao);
-    }
-    // In both cases, we have to account for space!=0, but the AO not accepting
-    // any new data (due to rounding to period boundaries).
-    p->buffers_full = max >= space && r <= 0;
-    p->avoid_ao_wait = (max == 0 && space > 0) || p->paused || stuck;
-    MP_TRACE(ao, "in=%d, space=%d r=%d flags=%d aw=%d full=%d f=%d\n", max,
-             space, r, flags, p->avoid_ao_wait, p->buffers_full, p->final_chunk);
-}
-
-// Estimate when the AO needs data again.
-static double ao_estimate_timeout(struct ao *ao)
-{
-    struct ao_push_state *p = ao->api_priv;
-    double timeout = 0;
-    if (p->buffers_full && ao->driver->get_delay) {
-        timeout = ao->driver->get_delay(ao) - 0.050;
-        // Keep extra safety margin if the buffers are large
-        if (timeout > 0.100)
-            timeout = MPMAX(timeout - 0.200, 0.100);
-    }
-    return MPMAX(timeout, ao->device_buffer * 0.75 / ao->samplerate);
+    if (r > 0)
+        p->expected_end_time = 0;
+    // Nothing written, but more input data than space - this must mean the
+    // AO's get_space() doesn't do period alignment correctly.
+    bool stuck = r == 0 && max >= space && space > 0;
+    if (stuck)
+        MP_ERR(ao, "Audio output is reporting incorrect buffer status.\n");
+    // Wait until space becomes available. Also wait if we actually wrote data,
+    // so the AO wakes us up properly if it needs more data.
+    p->wait_on_ao = space == 0 || r > 0 || stuck;
+    p->still_playing |= r > 0;
+    // If we just filled the AO completely (r == space), don't refill for a
+    // while. Prevents wakeup feedback with byte-granular AOs.
+    int needed = unlocked_get_space(ao);
+    bool more = needed >= (r == space ? ao->device_buffer / 4 : 1) && !stuck;
+    if (more)
+        mp_input_wakeup(ao->input_ctx); // request more data
+    MP_TRACE(ao, "in=%d flags=%d space=%d r=%d wa=%d needed=%d more=%d\n",
+             max, flags, space, r, p->wait_on_ao, needed, more);
 }
 
 static void *playthread(void *arg)
@@ -288,25 +305,9 @@ static void *playthread(void *arg)
         if (!p->paused)
             ao_play_data(ao);
 
-        // Request new data from decoder if buffer goes below "full".
-        // Allow a small margin of missing data for AOs that use timeouts.
-        double margin = ao->driver->wait ? 0 : ao->device_buffer / 8;
-        if (!p->buffers_full && unlocked_get_space(ao) > margin) {
-            if (!p->requested_data)
-                mp_input_wakeup(ao->input_ctx);
-            p->requested_data = true;
-        }
-
-        if (p->drain && (p->avoid_ao_wait || p->paused)) {
-            if (ao->driver->drain)
-                ao->driver->drain(ao);
-            p->drain = false;
-            pthread_cond_signal(&p->wakeup_drain);
-        }
-
         if (!p->need_wakeup) {
             MP_STATS(ao, "start audio wait");
-            if (p->avoid_ao_wait || p->paused) {
+            if (!p->wait_on_ao || p->paused) {
                 // Avoid busy waiting, because the audio API will still report
                 // that it needs new data, even if we're not ready yet, or if
                 // get_space() decides that the amount of audio buffered in the
@@ -314,12 +315,39 @@ static void *playthread(void *arg)
                 // The most important part is that the decoder is woken up, so
                 // that the decoder will wake up us in turn.
                 MP_TRACE(ao, "buffer inactive.\n");
-                mp_input_wakeup(ao->input_ctx);
-                pthread_cond_wait(&p->wakeup, &p->lock);
+
+                bool was_playing = p->still_playing;
+                double timeout = -1;
+                if (p->still_playing && !p->paused && p->final_chunk &&
+                    !mp_audio_buffer_samples(p->buffer))
+                {
+                    double now = mp_time_sec();
+                    if (!p->expected_end_time)
+                        p->expected_end_time = now + unlocked_get_delay(ao);
+                    if (p->expected_end_time < now) {
+                        p->still_playing = false;
+                    } else {
+                        timeout = p->expected_end_time - now;
+                    }
+                }
+
+                if (was_playing && !p->still_playing)
+                    mp_input_wakeup(ao->input_ctx);
+                pthread_cond_signal(&p->wakeup); // for draining
+
+                if (p->still_playing && timeout > 0) {
+                    mpthread_cond_timedwait_rel(&p->wakeup, &p->lock, timeout);
+                } else {
+                    pthread_cond_wait(&p->wakeup, &p->lock);
+                }
             } else {
+                // Wait until the device wants us to write more data to it.
                 if (!ao->driver->wait || ao->driver->wait(ao, &p->lock) < 0) {
                     // Fallback to guessing.
-                    double timeout = ao_estimate_timeout(ao);
+                    double timeout = 0;
+                    if (ao->driver->get_delay)
+                        timeout = ao->driver->get_delay(ao);
+                    timeout *= 0.25; // wake up if 25% played
                     mpthread_cond_timedwait_rel(&p->wakeup, &p->lock, timeout);
                 }
             }
@@ -331,6 +359,19 @@ static void *playthread(void *arg)
     return NULL;
 }
 
+static void destroy_no_thread(struct ao *ao)
+{
+    struct ao_push_state *p = ao->api_priv;
+
+    ao->driver->uninit(ao);
+
+    for (int n = 0; n < 2; n++)
+        close(p->wakeup_pipe[n]);
+
+    pthread_cond_destroy(&p->wakeup);
+    pthread_mutex_destroy(&p->lock);
+}
+
 static void uninit(struct ao *ao)
 {
     struct ao_push_state *p = ao->api_priv;
@@ -342,14 +383,7 @@ static void uninit(struct ao *ao)
 
     pthread_join(p->thread, NULL);
 
-    ao->driver->uninit(ao);
-
-    for (int n = 0; n < 2; n++)
-        close(p->wakeup_pipe[n]);
-
-    pthread_cond_destroy(&p->wakeup);
-    pthread_cond_destroy(&p->wakeup_drain);
-    pthread_mutex_destroy(&p->lock);
+    destroy_no_thread(ao);
 }
 
 static int init(struct ao *ao)
@@ -358,18 +392,23 @@ static int init(struct ao *ao)
 
     pthread_mutex_init(&p->lock, NULL);
     pthread_cond_init(&p->wakeup, NULL);
-    pthread_cond_init(&p->wakeup_drain, NULL);
     mp_make_wakeup_pipe(p->wakeup_pipe);
 
+    if (ao->device_buffer <= 0) {
+        MP_FATAL(ao, "Couldn't probe device buffer size.\n");
+        goto err;
+    }
+
     p->buffer = mp_audio_buffer_create(ao);
     mp_audio_buffer_reinit_fmt(p->buffer, ao->format,
                                &ao->channels, ao->samplerate);
     mp_audio_buffer_preallocate_min(p->buffer, ao->buffer);
-    if (pthread_create(&p->thread, NULL, playthread, ao)) {
-        ao->driver->uninit(ao);
-        return -1;
-    }
+    if (pthread_create(&p->thread, NULL, playthread, ao))
+        goto err;
     return 0;
+err:
+    destroy_no_thread(ao);
+    return -1;
 }
 
 const struct ao_driver ao_api_push = {
@@ -383,6 +422,7 @@ const struct ao_driver ao_api_push = {
     .pause = audio_pause,
     .resume = resume,
     .drain = drain,
+    .get_eof = get_eof,
     .priv_size = sizeof(struct ao_push_state),
 };
 

common/av_common.c

@@ -19,6 +19,9 @@
 
 #include <libavutil/common.h>
 #include <libavutil/log.h>
+#include <libavutil/dict.h>
+#include <libavutil/opt.h>
+#include <libavutil/error.h>
 #include <libavcodec/avcodec.h>
 
 #include "common/common.h"
@@ -64,7 +67,6 @@ void mp_copy_lav_codec_headers(AVCodecContext *avctx, AVCodecContext *st)
     avctx->channels                 = st->channels;
     avctx->block_align              = st->block_align;
     avctx->channel_layout           = st->channel_layout;
-    avctx->audio_service_type       = st->audio_service_type;
     avctx->bits_per_coded_sample    = st->bits_per_coded_sample;
 }
 
@@ -115,6 +117,8 @@ void mp_set_av_packet(AVPacket *dst, struct demux_packet *mpkt, AVRational *tb)
     if (mpkt && mpkt->avpacket) {
         dst->side_data = mpkt->avpacket->side_data;
         dst->side_data_elems = mpkt->avpacket->side_data_elems;
+        if (dst->data == mpkt->avpacket->data)
+            dst->buf = mpkt->avpacket->buf;
     }
     if (mpkt && tb && tb->num > 0 && tb->den > 0)
         dst->duration = mpkt->duration / av_q2d(*tb);
@@ -181,3 +185,45 @@ const char *mp_codec_from_av_codec_id(int codec_id)
     }
     return name;
 }
+
+// kv is in the format as by OPT_KEYVALUELIST(): kv[0]=key0, kv[1]=val0, ...
+// Copy them to the dict.
+void mp_set_avdict(AVDictionary **dict, char **kv)
+{
+    for (int n = 0; kv && kv[n * 2]; n++)
+        av_dict_set(dict, kv[n * 2 + 0], kv[n * 2 + 1], 0);
+}
+
+// For use with libav* APIs that take AVDictionaries of options.
+// Print options remaining in the dict as unset.
+void mp_avdict_print_unset(struct mp_log *log, int msgl, AVDictionary *dict)
+{
+    AVDictionaryEntry *t = NULL;
+    while ((t = av_dict_get(dict, "", t, AV_DICT_IGNORE_SUFFIX)))
+        mp_msg(log, msgl, "Could not set AVOption %s='%s'\n", t->key, t->value);
+}
+
+// kv is in the format as by OPT_KEYVALUELIST(): kv[0]=key0, kv[1]=val0, ...
+// Set these options on given avobj (using av_opt_set..., meaning avobj must
+// point to a struct that has AVClass as first member).
+// Options which fail to set (error or not found) are printed to log.
+// Returns: >=0 success, <0 failed to set an option
+int mp_set_avopts(struct mp_log *log, void *avobj, char **kv)
+{
+    int success = 0;
+    for (int n = 0; kv && kv[n * 2]; n++) {
+        char *k = kv[n * 2 + 0];
+        char *v = kv[n * 2 + 1];
+        int r = av_opt_set(avobj, k, v, AV_OPT_SEARCH_CHILDREN);
+        if (r == AVERROR_OPTION_NOT_FOUND) {
+            mp_err(log, "AVOption '%s' not found.\n", k);
+            success = -1;
+        } else if (r < 0) {
+            char errstr[80];
+            av_strerror(r, errstr, sizeof(errstr));
+            mp_err(log, "Could not set AVOption %s='%s' (%s)\n", k, v, errstr);
+            success = -1;
+        }
+    }
+    return success;
+}

common/av_common.h

@@ -26,6 +26,8 @@
 
 struct mp_decoder_list;
 struct demux_packet;
+struct AVDictionary;
+struct mp_log;
 
 int mp_lavc_set_extradata(AVCodecContext *avctx, void *ptr, int size);
 void mp_copy_lav_codec_headers(AVCodecContext *avctx, AVCodecContext *st);
@@ -36,5 +38,8 @@ void mp_set_avcodec_threads(AVCodecContext *avctx, int threads);
 void mp_add_lavc_decoders(struct mp_decoder_list *list, enum AVMediaType type);
 int mp_codec_to_av_codec_id(const char *codec);
 const char *mp_codec_from_av_codec_id(int codec_id);
+void mp_set_avdict(struct AVDictionary **dict, char **kv);
+void mp_avdict_print_unset(struct mp_log *log, int msgl, struct AVDictionary *d);
+int mp_set_avopts(struct mp_log *log, void *avobj, char **kv);
 
 #endif

common/av_opts.c

@@ -1,55 +0,0 @@
-/*
- * AVOption parsing helper
- * Copyright (C) 2008 Michael Niedermayer
- *
- * This file is part of MPlayer.
- *
- * MPlayer is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * MPlayer is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License along
- * with MPlayer; if not, write to the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- */
-
-#include <stdlib.h>
-#include <string.h>
-
-#include <libavutil/opt.h>
-
-#include "av_opts.h"
-
-int parse_avopts(void *v, char *str){
-    char *start;
-
-    if (!str)
-        return 0;
-
-    start= str= strdup(str);
-
-    while(str && *str){
-        char *next_opt, *arg;
-
-        next_opt= strchr(str, ',');
-        if(next_opt) *next_opt++= 0;
-
-        arg     = strchr(str, '=');
-        if(arg)      *arg++= 0;
-
-        if (av_opt_set(v, str, arg, AV_OPT_SEARCH_CHILDREN) < 0) {
-            free(start);
-            return -1;
-        }
-        str= next_opt;
-    }
-
-    free(start);
-    return 0;
-}

common/av_opts.h

@@ -1,30 +0,0 @@
-/*
- * AVOption parsing helper
- * Copyright (C) 2008 Michael Niedermayer
- *
- * This file is part of MPlayer.
- *
- * MPlayer is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * MPlayer is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License along
- * with MPlayer; if not, write to the Free Software Foundation, Inc.,
- * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
- */
-
-#ifndef MPLAYER_AV_OPTS_H
-#define MPLAYER_AV_OPTS_H
-
-/**
- * Parses str and sets AVOptions in v accordingly.
- */
-int parse_avopts(void *v, char *str);
-
-#endif /* MPLAYER_AV_OPTS_H */

common/codecs.c

@@ -17,7 +17,7 @@
 
 #include <assert.h>
 #include "talloc.h"
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 #include "common/msg.h"
 #include "codecs.h"
 

common/common.c

@@ -22,7 +22,7 @@
 #include <libavutil/common.h>
 
 #include "talloc.h"
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 #include "common/common.h"
 
 #define appendf(ptr, ...) \
@@ -171,7 +171,7 @@ static bool mp_parse_escape(void *talloc_ctx, bstr *dst, bstr *code)
     if (code->start[0] == 'x' && code->len >= 3) {
         bstr num = bstr_splice(*code, 1, 3);
         char c = bstrtoll(num, &num, 16);
-        if (!num.len)
+        if (num.len)
             return false;
         bstr_xappend(talloc_ctx, dst, (bstr){&c, 1});
         *code = bstr_cut(*code, 3);

common/common.h

@@ -24,7 +24,7 @@
 #include <stdbool.h>
 #include <stdint.h>
 
-#include "compat/compiler.h"
+#include "osdep/compiler.h"
 #include "talloc.h"
 
 // double should be able to represent this exactly

common/encode_lavc.c

@@ -590,11 +590,8 @@ AVStream *encode_lavc_alloc_stream(struct encode_lavc_context *ctx,
         }
         avcodec_get_context_defaults3(stream->codec, ctx->vc);
 
-        // stream->time_base = ctx->timebase;
-        // doing this breaks mpeg2ts in ffmpeg
-        // which doesn't properly force the time base to be 90000
-        // furthermore, ffmpeg.c doesn't do this either and works
-
+        // Using codec->time_base is deprecated, but needed for older lavf.
+        stream->time_base = ctx->timebase;
         stream->codec->time_base = ctx->timebase;
 
         ctx->voptions = NULL;
@@ -626,6 +623,8 @@ AVStream *encode_lavc_alloc_stream(struct encode_lavc_context *ctx,
         }
         avcodec_get_context_defaults3(stream->codec, ctx->ac);
 
+        // Using codec->time_base is deprecated, but needed for older lavf.
+        stream->time_base = ctx->timebase;
         stream->codec->time_base = ctx->timebase;
 
         ctx->aoptions = NULL;

common/global.h

@@ -7,9 +7,6 @@
 struct mpv_global {
     struct MPOpts *opts;
     struct mp_log *log;
-
-    int (*stream_interrupt_cb)(void *ctx);
-    void *stream_interrupt_cb_ctx;
 };
 
 #endif

common/msg.c

@@ -27,8 +27,8 @@
 
 #include "talloc.h"
 
-#include "bstr/bstr.h"
-#include "compat/atomics.h"
+#include "misc/bstr.h"
+#include "osdep/atomics.h"
 #include "common/common.h"
 #include "common/global.h"
 #include "misc/ring.h"
@@ -158,16 +158,10 @@ static void prepare_status_line(struct mp_log_root *root, char *new_status)
     size_t clear_lines = MPMIN(MPMAX(new_lines, old_lines), root->blank_lines);
 
     // clear the status line itself
-    if (terminal_erase_to_end_of_line[0]) {
-        fprintf(f, "\r%s", terminal_erase_to_end_of_line);
-    } else {
-        // This code is for MS windows (no ANSI control sequences)
-        get_screen_size();
-        fprintf(f, "\r%*s\r", screen_width - 1, "");
-    }
+    fprintf(f, "\r\033[K");
     // and clear all previous old lines
     for (size_t n = 1; n < clear_lines; n++)
-        fprintf(f, "%s\r%s", terminal_cursor_up, terminal_erase_to_end_of_line);
+        fprintf(f, "\033[A\r\033[K");
     // skip "unused" blank lines, so that status is aligned to term bottom
     for (size_t n = new_lines; n < clear_lines; n++)
         fprintf(f, "\n");
@@ -200,10 +194,20 @@ bool mp_msg_has_status_line(struct mpv_global *global)
     return r;
 }
 
+static void set_term_color(FILE *stream, int c)
+{
+    if (c == -1) {
+        fprintf(stream, "\033[0m");
+    } else {
+        fprintf(stream, "\033[%d;3%dm", c >> 3, c & 7);
+    }
+}
+
+
 static void set_msg_color(FILE* stream, int lev)
 {
     static const int v_colors[] = {9, 1, 3, -1, -1, 2, 8, 8, 8, -1};
-    terminal_set_foreground_color(stream, v_colors[lev]);
+    set_term_color(stream, v_colors[lev]);
 }
 
 static void pretty_print_module(FILE* stream, const char *prefix, bool use_color, int lev)
@@ -214,12 +218,12 @@ static void pretty_print_module(FILE* stream, const char *prefix, bool use_color
         unsigned int mod = 0;
         for (int i = 0; i < prefix_len; ++i)
             mod = mod * 33 + prefix[i];
-        terminal_set_foreground_color(stream, (mod + 1) % 15 + 1);
+        set_term_color(stream, (mod + 1) % 15 + 1);
     }
 
     fprintf(stream, "%10s", prefix);
     if (use_color)
-        terminal_set_foreground_color(stream, -1);
+        set_term_color(stream, -1);
     fprintf(stream, ": ");
     if (use_color)
         set_msg_color(stream, lev);
@@ -291,7 +295,7 @@ static void print_msg_on_terminal(struct mp_log *log, int lev, char *text)
         fprintf(stream, "%s", terminate);
 
     if (root->color)
-        terminal_set_foreground_color(stream, -1);
+        set_term_color(stream, -1);
     fflush(stream);
 }
 
@@ -363,34 +367,39 @@ void mp_msg_va(struct mp_log *log, int lev, const char *format, va_list va)
 // parent's name. If the name starts with "/", the parent's name is not
 // prefixed (except in verbose mode), and if it starts with "!", the name is
 // not printed at all (except in verbose mode).
+// If name is NULL, the parent's name/prefix is used.
 // Thread-safety: fully thread-safe, but keep in mind that talloc is not (so
 //                talloc_ctx must be owned by the current thread).
 struct mp_log *mp_log_new(void *talloc_ctx, struct mp_log *parent,
                           const char *name)
 {
     assert(parent);
-    assert(name);
     struct mp_log *log = talloc_zero(talloc_ctx, struct mp_log);
     if (!parent->root)
         return log; // same as null_log
     log->root = parent->root;
-    if (name[0] == '!') {
-        name = &name[1];
-    } else if (name[0] == '/') {
-        name = &name[1];
-        log->prefix = talloc_strdup(log, name);
-    } else {
-        log->prefix = parent->prefix
+    if (name) {
+        if (name[0] == '!') {
+            name = &name[1];
+        } else if (name[0] == '/') {
+            name = &name[1];
+            log->prefix = talloc_strdup(log, name);
+        } else {
+            log->prefix = parent->prefix
+                    ? talloc_asprintf(log, "%s/%s", parent->prefix, name)
+                    : talloc_strdup(log, name);
+        }
+        log->verbose_prefix = parent->prefix
                 ? talloc_asprintf(log, "%s/%s", parent->prefix, name)
                 : talloc_strdup(log, name);
+        if (log->prefix && !log->prefix[0])
+            log->prefix = NULL;
+        if (!log->verbose_prefix[0])
+            log->verbose_prefix = "global";
+    } else {
+        log->prefix = talloc_strdup(log, parent->prefix);
+        log->verbose_prefix = talloc_strdup(log, parent->verbose_prefix);
     }
-    log->verbose_prefix = parent->prefix
-            ? talloc_asprintf(log, "%s/%s", parent->prefix, name)
-            : talloc_strdup(log, name);
-    if (log->prefix && !log->prefix[0])
-        log->prefix = NULL;
-    if (!log->verbose_prefix[0])
-        log->verbose_prefix = "global";
     return log;
 }
 
@@ -428,8 +437,8 @@ void mp_msg_update_msglevels(struct mpv_global *global)
     root->use_terminal = opts->use_terminal;
     root->show_time = opts->msg_time;
     if (root->use_terminal) {
-        root->color = opts->msg_color && isatty(fileno(stdout));
-        root->termosd = !opts->slave_mode && isatty(fileno(stderr));
+        root->color = opts->msg_color && isatty(STDOUT_FILENO);
+        root->termosd = isatty(STDERR_FILENO);
     }
 
     talloc_free(root->msglevels);
@@ -469,6 +478,10 @@ struct mp_log_buffer *mp_msg_log_buffer_new(struct mpv_global *global,
 {
     struct mp_log_root *root = global->log->root;
 
+#if !HAVE_ATOMICS
+    return NULL;
+#endif
+
     pthread_mutex_lock(&mp_msg_lock);
 
     struct mp_log_buffer *buffer = talloc_ptrtype(NULL, buffer);

common/msg.h

@@ -24,7 +24,7 @@
 #include <stdlib.h>
 #include <stdint.h>
 
-#include "compat/compiler.h"
+#include "osdep/compiler.h"
 
 struct mp_log;
 

common/playlist.c

@@ -30,7 +30,8 @@
 struct playlist_entry *playlist_entry_new(const char *filename)
 {
     struct playlist_entry *e = talloc_zero(NULL, struct playlist_entry);
-    e->filename = talloc_strdup(e, filename);
+    char *local_filename = mp_file_url_to_filename(e, bstr0(filename));
+    e->filename = local_filename ? local_filename : talloc_strdup(e, filename);
     return e;
 }
 
@@ -105,10 +106,18 @@ static void playlist_unlink(struct playlist *pl, struct playlist_entry *entry)
     entry->pl = NULL;
 }
 
+void playlist_entry_unref(struct playlist_entry *e)
+{
+    e->reserved--;
+    if (e->reserved < 0)
+        talloc_free(e);
+}
+
 void playlist_remove(struct playlist *pl, struct playlist_entry *entry)
 {
     playlist_unlink(pl, entry);
-    talloc_free(entry);
+    entry->removed = true;
+    playlist_entry_unref(entry);
 }
 
 void playlist_clear(struct playlist *pl)
@@ -215,6 +224,15 @@ void playlist_transfer_entries(struct playlist *pl, struct playlist *source_pl)
     }
 }
 
+void playlist_append_entries(struct playlist *pl, struct playlist *source_pl)
+{
+    while (source_pl->first) {
+        struct playlist_entry *e = source_pl->first;
+        playlist_unlink(source_pl, e);
+        playlist_add(pl, e);
+    }
+}
+
 // Return number of entries between list start and e.
 // Return -1 if e is not on the list, or if e is NULL.
 int playlist_entry_to_index(struct playlist *pl, struct playlist_entry *e)
@@ -277,9 +295,6 @@ struct playlist *playlist_parse_file(const char *file, struct mpv_global *global
     if (ret && !ret->first)
         mp_warn(log, "Warning: empty playlist\n");
 
-    if (ret)
-        playlist_add_base_path(ret, mp_dirname(file));
-
     talloc_free(log);
     return ret;
 }

common/playlist.h

@@ -19,7 +19,7 @@
 #define MPLAYER_PLAYLIST_H
 
 #include <stdbool.h>
-#include "bstr/bstr.h"
+#include "misc/bstr.h"
 
 struct playlist_param {
     bstr name, value;
@@ -41,6 +41,17 @@ struct playlist_entry {
     bool playback_short : 1;
     // Set to true if not at least 1 frame (audio or video) could be played.
     bool init_failed : 1;
+    // Entry was removed with playlist_remove (etc.), but not deallocated.
+    bool removed : 1;
+    // Additional refcount. Normally (reserved==0), the entry is owned by the
+    // playlist, and this can be used to keep the entry alive.
+    int reserved;
+    // Used to reject loading of unsafe entries from external playlists.
+    // Can have any of the following bit flags set:
+    //  STREAM_SAFE_ONLY: only allow streams marked with is_safe
+    //  STREAM_NETWORK_ONLY: only allow streams marked with is_network
+    // The value 0 allows everything.
+    int stream_flags;
 };
 
 struct playlist {
@@ -74,6 +85,7 @@ void playlist_shuffle(struct playlist *pl);
 struct playlist_entry *playlist_get_next(struct playlist *pl, int direction);
 void playlist_add_base_path(struct playlist *pl, bstr base_path);
 void playlist_transfer_entries(struct playlist *pl, struct playlist *source_pl);
+void playlist_append_entries(struct playlist *pl, struct playlist *source_pl);
 
 int playlist_entry_to_index(struct playlist *pl, struct playlist_entry *e);
 int playlist_entry_count(struct playlist *pl);
@@ -82,4 +94,6 @@ struct playlist_entry *playlist_entry_from_index(struct playlist *pl, int index)
 struct mpv_global;
 struct playlist *playlist_parse_file(const char *file, struct mpv_global *global);
 
+void playlist_entry_unref(struct playlist_entry *e);
+
 #endif