The isLastBuffer flag passed to MediaCodecRenderer.processOutputBuffer()
is unreliable. It is true for the last frames (plural) of a video
stream, which makes it possible for the video renderer to end before all
the frames have been rendered to the output surface.
PiperOrigin-RevId: 700941685
This aligns the behavior with `MediaExtractor`, which sets this key in its [`MediaFormat`](https://developer.android.com/reference/android/media/MediaExtractor#getTrackFormat(int)) output.
Additionally, unnecessary `selectTrack` calls have been removed from the existing `getTrackFormat_...` tests, as they are not required to fetch the track format.
PiperOrigin-RevId: 700741326
Currently, `size` is cached in `SampleMetadata` when samples are committed in `SampleQueue`, which worked because we didn't support reading files with `CryptoInfo`. This change prepares for future `CryptoInfo` support, as the size changes when reading samples with encrypted data. Caching the size at commit time could lead to incorrect buffer `size` calculations.
Existing tests verify the correctness of the `getSampleSize()` API.
PiperOrigin-RevId: 700714178
The Android SKD version 30 is stable, allowing us to remove the `@SuppressLint("Override")` annotations from `InputReaderAdapterV30` and `OutputConsumerAdapterV30`.
PiperOrigin-RevId: 700681115
The previous code assumed that the `VBRI` Table of Contents (ToC)
covers all the MP3 data in the file. In a file with an invalid VBRI ToC
where this isn't the case, this results in playback silently stopping
mid-playback (and either advancing to the next item, or continuing to
count up the playback clock forever). This change considers the `bytes`
field to determine the end of the MP3 data, in addition to deriving it
from the ToC. If they disagree we log a warning and take the max value.
This is because we handle accidentally reading non-MP3 data at the end
(or hitting EoF) better than stopping reading valid MP3 data partway
through.
Issue: androidx/media#1904
#cherrypick
PiperOrigin-RevId: 700319250
Adds a configuration option to Frame Extractor to choose MediaCodecSelector.
Add a MediaCodecSelector that lists software decoders first
PiperOrigin-RevId: 699962365
Renderers join when they can start processing but not produce output.
In CompositionPlayer before this change, the VideoSink will be flushed when
position is reset regardless. This is useful for seeking, as seeking triggers position reset. But with pre-warming, a video renderer can be joining - it is
enabled (which also triggers position reset) before the previous image renderer
is disabled. At this moment, as the image renderer is still producing frames,
we should not flush the video sink just now.
PiperOrigin-RevId: 699943882
We are currently waiting until stream 1 is completely rendered on screen
before registering stream 2 but that is not necessary anymore because
VideoFrameProcessor supports registering stream 2 before stream 1 is
fully processed.
PiperOrigin-RevId: 699929943
This call does 2 things:
1. It's used to estimate the frame rate early but this is incorrect as
the frame rate can be changed by effects. I have tested playback on
my device and I don't see any difference.
2. For ExoPlayer.setVideoEffects() only: it allows dropping all the
frames until the next key frame in case the input frame is "very
late". This doesn't seem really necessary though because because we
do the same thing after applying effects.
This change is also a step towards moving all the calls to
VideoFrameRelease/RenderControl to DefaultVideoSink.
PiperOrigin-RevId: 698732161
3P app reported it's slow to see the first image frame on the screen, when
`playWhenReady` is false. This is because the player would wake up less
frequently when `playWhenReady` is false. This CL adds a runnable to wake up
the player any time a new frame is made available.
PiperOrigin-RevId: 698066887
The following was happening:
- VideoFrameRenderControl.render() was calling
VideoFrameReleaseAction.getFrameReleaseAction().
- VideoFrameReleaseAction.getFrameReleaseAction() was calling
FrameTimingEvaluator.shouldIgnoreFrame(), which is implemented in
MediaCodecVideoRenderer.
- MediaCodecVideoRenderer.shouldIgnoreFrame(), when returning true,
was also flushing the VideoSink, which was flushing the
VideoFrameRenderControl.
- VideoFrameRenderControl.render() was removing the frame from the
presentationTimestampsUs queue, but the frame had already been removed
by the flush operation, causing an exception to be thrown.
This fix removes the last step.
PiperOrigin-RevId: 697915692
To do this, refactor stream start position handling so that
VideoFrameRenderControl is only passed a start position when it should
be applied, and therefore doesn't need to take a timestamp associated
with each start position anymore
PiperOrigin-RevId: 697544416
From [ the last change in `DefaultPreloadManagerTest`](2b54b1ebbe), the preloadManager began to use a separate `preloadThread` in `release_returnZeroCount_sourcesAndRendererCapabilitiesListReleased`, which unveils a bug in `PreloadMediaSource`. When `PreloadMediaSource.releasePreloadMediaSource` is called, `preloadHandler` will post a `Runnable` on the preload looper to release the internal resources. Before this `Runnable` is executed, it is possible that the [`stopPreloading`](https://github.com/androidx/media/blob/main/libraries/exoplayer/src/main/java/androidx/media3/exoplayer/source/preload/PreloadMediaSource.java#L442) method is executed just as the result of preloading has completed. This is expected to remove the posted `Runnable`s for further preloading, however, the posted `Runnable` for releasing will also be removed from the message queue.
Ideally we should use `postDelayed(runnable, token, delayMillis)` to post the runnables so that the token will be useful to identify which messages to remove in `removeCallbacksAndMessages(token)`, but that `postDelayed` method is only available from API 28. So in this change we are using a separate handler for releasing, and then the call of `preloadHandler.removeCallbacksAndMessages` won't impact the runnable for releasing.
#cherrypick
PiperOrigin-RevId: 696894483
This method returns a `PersistableBundle` containing metrics data for the current media container. The bundle includes attributes and values for the media container, as described in `MediaExtractor.MetricsConstants`.
PiperOrigin-RevId: 696893276
The new callback allows an app to read data from the content
timeline to populate the `AdPlaybackState`. To avoid a deadlock
between the `AdMediaSource` waiting for the `AdPlaybackState` and
the app waiting for the content `Timeline`, a boolean flag
`useLazyContentSourcePreparation` is introduced to tell the
`AdsMediaSource` to prepare the content source immediately
to make the `Timeline` available.
A unit test verifies that in none of the cases (lazy preparation or
immediate preparation) a `Timeline` without ad data is leaked to the
caller. This ensures that in the case of a preroll, the player won't
initially and accidentally read content media data before starting to
load the preroll ad. While the content source is prepared early, no
content media period must be created before the preroll starts.
PiperOrigin-RevId: 696885392
- Rename a few methods/variable to improve readability.
- Refactor how video size changes are tracked. This will simplify the
upcoming logic to refactor the timestamp logic in
VideoFrameRenderControl because we will use the same logic for
outputVideoSize and for outputStreamStartPositionUs.
PiperOrigin-RevId: 696026515
Implemented `getCachedDuration()` to provide an estimate of cached data in memory and `hasCacheReachedEndOfStream()` to indicate if the cache has reached the end of the stream.
Note: The Javadoc for the newly added methods closely follows that of the platform `MediaExtractor`. While the current implementation always uses a cache and therefore never returns `-1` from `getCachedDuration`, this leaves room for future changes should caching behavior or conditions evolve.
PiperOrigin-RevId: 695694460
The `RendererCapabilities` and `TrackSelector` objects are accessed on the preload thread during the preloading, when releasing them, they need to be released on the same thread. Otherwise, it is possible that they have already released on the application thread, while the PreloadMediaSource still tries to access them on the preload thread before the source is released.
#cherrypick
PiperOrigin-RevId: 694173131
Note: `androidx.media3.common.DrmInitData` is returned instead of the platform `android.media.DrmInitData` because the platform class has a package-private constructor, making it impossible to implement the abstract class outside the platform.
PiperOrigin-RevId: 694096542
Implemented `setLogSessionId(LogSessionId)` and `getLogSessionId()` methods.
Note: The `LogSessionId` is currently **not** forwarded to `MediaParser`, but this will be addressed once `MediaParser` can be used to configure `MediaExtractorCompat`.
PiperOrigin-RevId: 692945255
The test has been moved to an instrumentation test as it relies on APIs that vary by SDK version. Robolectric’s emulation lacks sufficient realism in some cases, which impacts test accuracy. By using an instrumentation test, we ensure that the tests run in a real Android environment, providing reliable results for SDK-dependent APIs.
PiperOrigin-RevId: 692933259
This change keeps getSurfaceForCodec and hasSurfaceForCodec in sync.
Before this change when ExoPlayer is configured with setVideoEffects
and no display surface, codecs that need Set Output Surface Workaround
would not be initialized because hasSurfaceForCodec always returns false
PiperOrigin-RevId: 692900899
Restructured sample metadata management in `MediaExtractorCompat` to maintain a queue of `SampleMetadata` objects, each containing `timeUs`, `flags`, `size`, and `trackIndex`, rather than storing only track indices.
Introduced a pooling mechanism for `SampleMetadata` to reduce object allocation and improve memory efficiency. The new `SampleMetaDataQueue` centralizes metadata for easier access and management, eliminating the need to allocate a buffer or peek into the `SampleQueue` to retrieve sample metadata.
This change does not change existing behavior, it optimizes the internal structure, and existing tests already verify the relevant APIs.
PiperOrigin-RevId: 692285581
If audio processors report a drifting position, we currently update
the media position parameters to correct this drift. However, this
means we pass in the wrong value to
audioProcessorChain.getMediaDuration, which reuqires the time since
the last flush.
To fix this problem, we can instead save the drift seperately and
apply it where needed.
PiperOrigin-RevId: 692202219
This affects both `AnalyticsListener` and `MediaSourceEventListener`
This was introduced by d051b4b993
Also fix a missing 'load started' event for HLS media playlists (this
was also introduced by d051b4b993).
PiperOrigin-RevId: 691580183
Previously, `getTrackFormat()` in `MediaExtractorCompat` returned a `MediaFormat` without setting `MediaFormat.KEY_DURATION`. With this change:
- `MediaFormat.KEY_DURATION` is set based on the track's duration, if available.
- If the track duration is unset, the duration from the seek map is used as a fallback.
- When neither duration is set, `MediaFormat.KEY_DURATION` remains unset.
This ensures that `MediaFormat.KEY_DURATION` is populated when possible, enhancing duration information availability.
PiperOrigin-RevId: 691395114
This method will likely be removed in the next release, and is currently
only needed from within the `source` package.
#cherrypick
PiperOrigin-RevId: 691351449
In some cases, the streamOffsetUs was passed to
VideoFrameReleaseControl.getFrameReleaseAction() but it should be the
streamStartPositionUs.
PiperOrigin-RevId: 691040172
When transitioning to the next media position parameter checkpoint
we estimate the position because the audio processor chain no longer
provides access to the actual playout duration.
The estimate using the declared speed and the last checkpoint may
have drifted over time, so we currently estimate relative to the
next checkpoint, which is closer and presumably provides a better
estimate. However, this assumes that these checkpoint are perfectly
aligned without any position jumps.
The current approach has two issues:
- The next checkpoint may include a position jump by design, e.g.
if it was set for a new item in the playlist and the duration of
the current item wasn't perfectly accurate.
- The sudden switch between two estimation methods may cause a jump
in the output position, which is visible when we add new media
position checkpoints to the queue, not when we actually reach the
playback position of the checkpoint.
We can fix both issues by taking a slightly different approach:
- Continuously monitor the estimate using the current checkpoint. If
it starts drifting, we can adjust it directly. This way the estimate
is always aligned with the actual position.
- The change above means we can safely switch to using the estimate
based on the previous checkpoint. This way we don't have to make
assumptions about the next checkpoint and any position jumps will
only happen when we actually reach this checkpoint (which is more
what a user expects to see, e.g. at a playlist item transition).
Issue: androidx/media#1698
PiperOrigin-RevId: 690979859
These methods are marked `default` on the `AnalyticsListener` interface
with an empty implementation, so there's no need to override them just
to re-define the empty implementation.
PiperOrigin-RevId: 689416584
Add `DecoderAudioRenderer.getDurationToProgressUs()` so that `ExoPlayer`, if set with `experimentalSetDynamicSchedulingEnabled()`, will dynamically schedule its main work loop to when the `DecoderAudioRenderer` can make progress.
PiperOrigin-RevId: 689377247
- Create `LibiamfAudioRenderer` with `DefaultRenderersFactory` in `ExoPlayerModuleProguard`.
- Remove redundant library availability check from `IamfModuleProguard`.
- Move `proguard-rules.txt` to the root folder.
- Removed unused `cryptoType` parameter from `setLibraries()` method in `IamfLibrary`.
- Added log when `LibiamfAudioRenderer` is loaded in `DefaultRenderersFactory`.
- Annotated missing classes with `@UnstableApi`.
- Check for library availability and throw exception in `IamfDecoder` constructor.
#cherrypick
PiperOrigin-RevId: 689330016
Also add a test for this to avoid missing any others in future. Also
flesh out the existing test for the deprecated builder, to assert the
return type is correctly updated.
PiperOrigin-RevId: 688948768
Frame rate change is currently notified to the release control when the
video frame processor input frame rate changes, but it should be when
the video frame processor output frame rate changes.
PiperOrigin-RevId: 688512300
Added start and end time details to the error message for `REASON_START_EXCEEDS_END`, helping to debug cases where the start time exceeds the end time.
PiperOrigin-RevId: 688117440
Before this change:
* With legacy subtitle decoding (at render time), load errors (e.g. HTTP
404) would result playback completely failing, while parse errors
(e.g. invalid WebVTT data) would be silently ignored, so playback
would continue without subtitles.
* With new subtitle decoding (at extraction time), both load and parse
errors would result in playback completely failing.
This change means that now neither load nor parse errors in text or
metadata tracks stop playback from continuing. Instead the error'd track
is disabled until the end of the current period.
With new subtitle decoding, both load and parse errors happen during
loading/extraction, and so are emitted to the app via
`MediaSourceEventListener.onLoadError` and
`AnalyticsListener.onLoadError`. With legacy subtitle decoding, only
load errors are emitted via these listeners and parsing errors continue
to be silently ignored.
Issue: androidx/media#1722
PiperOrigin-RevId: 686902979
These events are always reported with the primary child period ID,
because this is the same ID used in the parent `MergingMediaSource`'s
Timeline.
This ensures that e.g. loading errors from sideloaded subtitles (which
uses `MergingMediaSource`) are now reported via
`AnalyticsListener.onLoadError`.
It results in non-error events being reported from these children too,
which will result in more `onLoadStarted` and `onLoadCompleted` events
being reported (one for each child).
Issue: androidx/media#1722
PiperOrigin-RevId: 686901439
Use this for sideloaded subtitles, so preparation can still complete
despite an error from e.g. `DataSource.open`. In this case, no subtitle
tracks will be emitted.
Issue: androidx/media#1722
PiperOrigin-RevId: 686888588
When a `SampleQueue` is prepared prior to playback, the start position may be less than the timestamp of the first sample in the queue and still be valid. This scenario can come about with specific clipping values and if all samples are sync samples. Currently, with `ClippingMediaPeriods` around `ProgressiveMediaPeriods`, if the `SampleQueue` has already been reset through the seekTo operation in `onPrepared`, then in the aforementioned scenario the seekTo operation in `handleDiscontinuity` will remove all samples and reset the loading periods unnecessarily.
The solution is that if the `ProgressiveMediaPeriod` has already handled a seekTo operation for the same position and the sample queue has not been read yet, then loading does not need to be reset.
The tests in `MergingPlaylistPlaybackTest` were specifically causing this behavior through its setup of `MergingMediaSources` around clipped `FilteringMediaSources`. Since the video content was not 'all sync samples', there would always be a discontinuity to handle and the audio content being 'all sync samples' would start with samples post start time.
These changes also remove the flakiness from the `MergingPlaylistPlaybackTest`.
PiperOrigin-RevId: 686858444
Currently every test in the library passes `null` here, which seems to
end up being passed into non-null places in the library. This change
removes the parameter and instead initializes the field to a
`DefaultAllocator` instance in the same way that `DefaultLoadControl`
creates one.
PiperOrigin-RevId: 686823273
Many of the Android TV platforms do not report accurate `SystemClock.elapsedRealtime()`
In addition this drift is not consistant accross multiple boxes, so the time offset
between any pair of boxes will vary over time.
This variance will lead to a drift in the Live Offset calculation.
Fix is simple, inValidate the NTP server query after a period of time (10 minutes)
and re-issue the call
The `DefaultLoadControl` implementation of onTracksSelected only utilizes the `Renderer[]` parameter for use in stream type, of which it can collect from the `ExoTrackSelection[]` parameter.
PiperOrigin-RevId: 685677726
The `DefaultPreloadManager.Builder` is able to build the `DefaultPreloadManager` and `ExoPlayer` instances with the consistently shared configurations. Apps can:
* Simply setup the `DefaultPreloadManager` and `ExoPlayer` with all default configurations via `build()` and `buildExoPlayer()`;
* Or customize the shared configurations by the setters on `DefaultPreloadManager.Builder` and setup via `build()` and `buildExoPlayer()`;
* Or customize the player-only configurations for `ExoPlayer` via `buildExoPlayer(ExoPlayer.Builder)`.
PiperOrigin-RevId: 684852808
