When estimating the AudioTrack min buffer size, we must use a PCM
frame of 1 when doing direct playback (passthrough). The code was
passing -1 (C.LENGTH_UNSET).
PiperOrigin-RevId: 489238392
(cherry picked from commit d9d716869b7df2cd95704e9ac24a5f9a376afa2c)
I considered moving this enforcement inside the ExoPlayerImpl
implementation, but it might lead to app crashes in cases that apps
(incorrectly) call a released player, but it wasn't actually causing a
problem.
PiperOrigin-RevId: 489233917
(cherry picked from commit d4c9199a6175a6c89f7a3d3900af1b9dfcb075f0)
They are called from the system only and don't need to be exported
to be visible to other apps.
PiperOrigin-RevId: 489210264
(cherry picked from commit c1fd03df7403019143f503bbe208d73cabc11243)
The method allows clients to specify a pre-existing thread
to use for playback. This can be used to run multiple ExoPlayer
instances on the same playback thread.
PiperOrigin-RevId: 488980749
(cherry picked from commit 79b809b5563260cd20541afe607e36ae351fbbed)
Added new method to check if codec just functionally supports a format. Changed getDecoderInfosSortedByFormatSupport to use new function to order by functional support. This allows decoders that only support functionally and are more preferred by the MediaCodecSelector to keep their preferred position in the sorted list.
UnitTests included
-Two MediaCodecVideoRenderer tests that verify hw vs sw does not have an effect on sort of the decoder list, it is only based on functional support
Issue: google/ExoPlayer#10604
PiperOrigin-RevId: 487779284
(cherry picked from commit 1eb8a6b36ed98bd66a5bad5273526a918511a39f)
This tag is only understood by Dackka, which is used to generate the media3 javadoc.
PiperOrigin-RevId: 489233200
(cherry picked from commit 058cba95d40db83c77f4dd0aa8cb561221463b3a)
This better matches the callback name (onSurfaceSizeChanged) and
probably cause less confusion with getVideoSize.
PiperOrigin-RevId: 488669786
(cherry picked from commit 8438daca1f8cbea24aa414ee500044baf1a28367)
Util.getAudioTrackChannelConfig() maps a channel count to a
channel mask that is passed to AudioTrack. The method expected that
playback of 8-channel audio is possible from Android 5.1 and playback of
12-channel audio is only possible from Android 12L. However, there is no
restriction on the upper number of channels that can be passed to the
AudioTrack. google/ExoPlayer#10701 is an example where the audio decoder
outputs 12 channels on an Android 10.
This change removes the restrictions for 8 and 12 channels. Note, we still
do not support playback of arbitrary number of channels as it would require
further changes to DefaultAudioSink.
#minor-release
Issue: google/ExoPlayer#10701
PiperOrigin-RevId: 488659831
(cherry picked from commit 1b24e6fd14774d5c8b3e7b8084b30204581d13ce)
When we currently trigger the iteration finished event during the
release, we don't mark the event as triggered. This means that
someone can trigger another release from within the callback,
which then tries to resend the event.
Issue: google/ExoPlayer#10758
#minor-release
PiperOrigin-RevId: 488645089
(cherry picked from commit 3e5103a3e55e2dc545b4a2442beaa49f307ecfdc)
We currently skip this calculation entirely, but it can be added by
calculating the window duration using the wrapped window's duration
and the provided AdPlaybackState.
Issue: google/ExoPlayer#10764
PiperOrigin-RevId: 488614767
(cherry picked from commit 20151b9930f1b54d7894ec41465a2e85df92462d)
It's not clear to me why presubmit didn't catch this, I briefly
investigated but couldn't work it out - so I'm just going to fix
it and move on.
#minor-release
PiperOrigin-RevId: 487497827
(cherry picked from commit 13ee34faccaf98ff90850601f40c95622b1e878b)
These aren't caught by presubmit because the classes are annotated
`@VisibleForTesting` and are therefore stripped out by Metalava. However
Metalava doesn't run when we're generating javadoc for real.
#minor-release
PiperOrigin-RevId: 487476260
(cherry picked from commit 14e23d34e17afa076cd912264bb60685961c58e3)
This involves reducing the visibility of methods/constructors that
are already unusable outside the `androidx.media3.test.utils` package.
#minor-release
PiperOrigin-RevId: 487473005
(cherry picked from commit ca8de0e49b03339b03a6e048e7e7a27df86e5ff5)
This uses `@hide` on `protected final` methods to hide them from Dackka
javadoc generation, since these methods are inaccessible to developers
anyway. These symbols will still (currently) be included in artefacts
distributed on Maven (because we don't run Metalava as part of
generating these artefacts).
In some cases I had to change the visibility/finality of methods to make
them `protected final` before adding the `@hide` annotation (but
the impact should be very low, since most of these methods were either
already unusable by app developers, or they shouldn't have been used).
#minor-release
PiperOrigin-RevId: 487472907
(cherry picked from commit 1cd488ac9a1e838139495f55aa32eb951f1e8d8f)
This makes two fixes:
1. Remove `HlsSampleStreamWrapper.Callback` (package-private) from the
list of interfaces implemented by `HlsMediaPeriod` (`public`) and
move the implementation to a private inner class instead. This avoids
Metalava complaining about a public class that inherits from a
package-private type.
2. Reduce the visibility of
`RtpPayloadFormat.isFormatSupported(MediaDescription)` from `public`
to package-private. The `MediaDescription` type is already
package-private, so this method was already unusable outside the
package.
#minor-release
PiperOrigin-RevId: 487472781
(cherry picked from commit 9041d7b979f3b998758d74279f5d281088d879a8)
This makes two types of fix:
1. Align parameter names on overridden methods where the superclass
has `@param` javadoc.
2. Use `@hide` on `protected final` methods that refer to package-private
types. This will hide these symbols from Dackka javadoc generation
but not (currently) from the artefacts distributed on Maven. These
methods are currently unusable outside their package anyway (e.g. by
external developers) because of the dependency on a package-private
type.
This also changes some HLS, SmoothStreaming, and IMA code where I've renamed
parameters of overridden methods to be consistent across the type
hierarchy.
#minor-release
PiperOrigin-RevId: 487472665
(cherry picked from commit 10c4a4dfc172cfac77528c7cb746fe827ca6f78c)
Also, document that we tone map when no HDR features are explicitly set
PiperOrigin-RevId: 487310971
(cherry picked from commit 8bdd2784d330c0c9f493f092825d08f8a5cc4305)
Not setting the color info results in a missing "colr" box in the produced
container, under file/moov/trak/mdia/minf/stbl/stsd/hvc1. This means extractors
will not be able to find out the transcoded file is HDR.
In `Transformer`, this means it can't transcode this transcoded file, because
it currently relies on the container bearing HDR info to construct the
transcoding sample pipeline.
PiperOrigin-RevId: 487276712
(cherry picked from commit d6c8e3a8aff07be5c184473826753047bf4b2387)
In startTransformation method we were throwing UnsupportedEncodingException (IOException) when mediaItem with unsupported arguments is passed.
Changed this to IllegalArgumentException which seems more logical here.
PiperOrigin-RevId: 487259296
(cherry picked from commit 4598cc92485c149f5c613d3c926ae4493a457668)
Just move some code around for now, to start setting up the overall
structure.
PiperOrigin-RevId: 487229329
(cherry picked from commit 5d1cab0cf50e31e5886310dd55e6de7c4b3523ba)
Imported from GitHub PR Issue: google/ExoPlayer#10762
This ensure that ffmpeg error code are properly translated to values that the ExoPlayer decoder understand.
The main gain is that it allows the decoder to properly ignore more cases of invalid data and recover.
The second gain is that the other errors are now proper ExoPlayer errors and no more obscure buffer ones.
Fixes: Issue: google/ExoPlayer#10760
Merge 82ceeb77d6df71f5ffb0474db66a36fd6eb8e51a into 972e169bd85b14848dcae75e34f9e95fe87e1f4b
COPYBARA_INTEGRATE_REVIEW=go/exoghi/10762 from Tolriq:ffmpeg_error_code 82ceeb77d6df71f5ffb0474db66a36fd6eb8e51a
PiperOrigin-RevId: 487189910
(cherry picked from commit 6d2e7a1b5760a05eade18b1a028b3d27f714715c)
This is to avoid having this logic in TransformerInternal once it is
added.
PiperOrigin-RevId: 487159941
(cherry picked from commit b59fdf5e98bf5ec3cd0800edb5fb2e6ebd9da5c3)
This logic is currently in the player renderers. With multi-asset, the
renderers will go into the AssetLoader, which shouldn't be responsible
for muxing.
PiperOrigin-RevId: 486860502
(cherry picked from commit d8754b6642d0a99e2705f3e21ff8b83d50472bbd)
Problem: We are initialising muxer as soon as we start the transformation. Now the startTransformation() method can be called from main thread, but muxer creation is an I/O operation and should be not be done on main thread.
Solution: Added lazy initialisation of muxer object. The actual transformation happens on background thread so the muxer will be initialised lazily from background thread only.
Another way was to provide an initialize() method on MuxerWrapper which will explicitly initialise muxer object but with this approach the caller need to call the initialise method before calling anything else. With current implementation the renderers are calling MuxerWrapper methods on various callbacks (Not sequentially) and also we are sharing same muxer with multiple renderers so It might become confusing for the caller on when to call the initialise() method. Also there are few methods on MuxerWrapper which dont really need muxer object. So in short it might make MuxerWrapper APIs more confusing.
Validation: Verified the transformation from demo app.
PiperOrigin-RevId: 486735787
(cherry picked from commit b10b4e6d4694c7240073b81c3a09227042b58c21)
This should be necessary to ensure decoders see fewer errors.
Setting this resulted in removing native_dequeueOutputBuffer errors on OMX.MTK decoders for in-app tone mapping prototyping.
PiperOrigin-RevId: 486715941
(cherry picked from commit 0b7e5bbad287053d56720ad682c1a5015eedb9cf)
If there's an @param javadoc tag in a supertype then all overrides
of this method that don't also override the javadoc must use the same
parameter name.
PiperOrigin-RevId: 485857711
(cherry picked from commit b8e5d708fc77a06e831abd7f2c2e57f68bb3720f)
Public methods may only refer to public types in their signature. This
change ensures that by switching to a public supertype everywhere.
PiperOrigin-RevId: 485568625
(cherry picked from commit 3fdfe585ffe927c90f412b892101dbe01119b8df)
Not sure what BUILD.ID the fix will first be applied on though.
PiperOrigin-RevId: 485389618
(cherry picked from commit 2ddd645a3484aed4b9e519f99f4ab1bc989b6a0e)
