Use the PQ OETF and EOTF to ensure that intermediate fragment shader operations
using PQ are in linear BT.2020 rather than PQ and HLG-1 BT.2020.
Also, swap the OETF and EOTF in shaders, as they were used incorrectly before
Manually tested by verifying transformer demo HLG and PQ videos look the same with and without this CL, including with a BitmapOverlayProcessor enabled to test flows both with one MatrixTransformationProcessor that skips HDR TFs, and with one that doesn't.
PiperOrigin-RevId: 469736067
(cherry picked from commit 2ad07e88ee)
Upstream timestamps from the decoder are also in microseconds,
so using microseconds here is consistent with that.
PiperOrigin-RevId: 468659099
(cherry picked from commit 0b1c540ff9)
Remove the manual overwriting of Note ON events that have 0 velocity with Note OFF. JSyn handles this already.
- The implementation of "running status" means that the amount of bytes read from the file differ from the size of the sample that ends up in the decoder. The decoder sample contains the applied running status (status of previous event), which the file bytes don't contain.
PiperOrigin-RevId: 468537659
(cherry picked from commit 30257c767b)
Adds a method to FrameProcessor.Listener to be called when an
output frame is available and a method releaseOutputFrame in
FrameProcessor allowing the caller to trigger release of the
oldest available output frame at a given timestamp. Late frames
or frames with unset release times are dropped in the
FinalMatrixTransformationProcessorWrapper.
More than one output frame can become available before they are
released if the penultimate GlTextureProcessor is capable of producing
multiple output frames. Processing continues while waiting for
releaseOutputFrame to be called. Frame release tasks are prioritized
over other tasks.
PiperOrigin-RevId: 468473072
(cherry picked from commit a5d7fdcab5)
Manually tested using transformer demo HLG videos. Before this CL, RGB values after the YUV to RGB conversion reached up to 1.025. After this CL, RGB values correctly clamp at 1.0.
PiperOrigin-RevId: 468426092
(cherry picked from commit 244d38cf0e)
FrameProcessingTaskExecutor should be released on error.
There can be a delay until this happens, so
FrameProcessingTaskExecutor will cancel any pending tasks
and drop new tasks until it is released.
PiperOrigin-RevId: 468171820
(cherry picked from commit 0c961a7abd)
We create an empty CueGroup in many places as default or
where none is needed. Instead, we can define a constant
for this purpose and reuse it.
PiperOrigin-RevId: 467944841
(cherry picked from commit 59895646e0)
This is needed as a pre-requisite for allowing MCVR to control
FrameProcessor frame release for previewing.
Submitting a high-priority task is conceptually different from
posting at the front of a single queue of tasks, as the high-priority
tasks are executed in FIFO order among themselves. This will ensure
that frame release tasks submitted in close succession are executed
in the order they are submitted but before any lower priority tasks.
PiperOrigin-RevId: 467675137
(cherry picked from commit 59be732230)
Increase the estimated max sample size for HEVC by 2x, and set a minimum
size of 2MB. The 2MB will be applied for resolutions up to 1080p, after
which the new calculation takes effect. This is in par with the
platform's HEVC software decoder.
PiperOrigin-RevId: 467641494
(cherry picked from commit b83b16eba7)
The specified CMake version doesn't work with the latest
Android Studio releases. Updating to a more recent version
fixes the problem.
Issue: google/ExoPlayer#9933
PiperOrigin-RevId: 467634063
(cherry picked from commit 9e1d4f6fd6)
This base class will simplify the implementation of custom
Player classes. The current version only supports
available commands and playWhenReady handling.
PiperOrigin-RevId: 467618021
(cherry picked from commit 43ede97d2f)
When the player finishes playback and reaches the STATE_ENDED,
the notification remains visible with a pause button and the
service is kept in the foreground. This is a bug.
With this change, when the player reaches the STATE_ENDED, the
service is stopped from the foreground and a notification is shown
with a play button. If the play icon is tapped, the player will restart
playback of the last played item. Playing the last played item again
is the existing behavior when play/pause commands are received from
the legacy MediaSession (e.g. BT headset buttons).
#minor-release
Issue: google/ExoPlayer#112
PiperOrigin-RevId: 467231509
(cherry picked from commit 7a7e1eb23b)
* Rename all Rgba instances to Rgb.
* Remove alpha value from the RGBA Matrices and apply the 4x4 matrix
only to the R, G, B channels.
* Restore the alpha from the input unchanged.
PiperOrigin-RevId: 467208888
(cherry picked from commit 8760ee48b9)
This change adds ExternalTextureManager which implements
InputListener to only queue input frames to the
ExternalTextureProcessor when it is ready to accept an input
frame. This replaces the old retry-logic in GlEffectsFrameProcessor.
Before this change, the retrying in GlEffectFrameProcessor wasted
CPU time if input becomes available faster than the
ExternalTextureProcessor can process it.
PiperOrigin-RevId: 467177659
(cherry picked from commit d7bf1ed2d7)
From NDK 23.1.7779620 and above, the arm64-v8a ABI needs additional
build flags to correctly link the ffmpeg libraries.
Issue: google/ExoPlayer#9933
PiperOrigin-RevId: 467161973
(cherry picked from commit 5874327e5d)
In some cases, the IMA SDK fails to call the expected loadAd
event to load the next ad to play. This is (potentially) the
only remaining case where playback can get stuck due to missing
calls from IMA as the player doesn't even have a MediaSource at
this stage and is only waiting for IMA to provide the ad URL.
We can reuse the existing adPreloadTimeoutMs that was added for
a similar purpose (when preloading the first ad in the group).
The JavaDoc matches this purpose as well and the default timeout
is appropriate since we expect to get the loadAd call immediately.
Issue: google/ExoPlayer#10510
PiperOrigin-RevId: 466953617
(cherry picked from commit b125d45a63)
createSupportedTransformationRequest is more accurate than
createFallbackTransformationRequest, as a TransformationRequest will be returned
regardless of whether any fallback is applied.
PiperOrigin-RevId: 466641277
(cherry picked from commit c5a5364673)
This change adds a new method onReadyToAcceptInputFrame to
GlTextureProcesssor.InputListener and changes maybeQueueInputFrame
to queueInputFrame, removing the boolean return value.
This avoids the re-trying in ChainingGlTextureProcessorListener
by allowing it to only feed frames from the producing to the consuming
GlTextureProcessor when there is capacity.
MediaPipeProcessor still needs re-trying when processing isn't 1:1.
PiperOrigin-RevId: 466626369
(cherry picked from commit 97e6a86d2b)
Previously, this feature interpreted SDR signals as HDR when called. Now, only HDR
streams are interpreted as HDR, so the javadoc should be updated. Not yet removing
this method, as there are still some loose ends to finish up (ex. PQ support, e2e
tests).
PiperOrigin-RevId: 466425738
(cherry picked from commit 05fcda510a)
While HDR is most closely tied to the color transfer (ex.
COLOR_TRANSFER_SDR is the only one explicitly mentioning dynamic
range), technically color spaces may be associated with HDR as well,
like BT.2020 commonly being used for HDR rather than BT.709 for SDR.
Therefore, it's more specific to mention just that the transfer is HDR.
PiperOrigin-RevId: 466316960
(cherry picked from commit 35161c7489)
This simplifies ChainingGlTextureProcessor as it now only connects a
consuming and a producing GlTextureProcessor rather than a previous,
current, and next GlTextureProcessor.
Also use default no-op implementations of the listeners in
SingleFrameGlTextureProcessor and MediaPipeProcessor to avoid
null-checks.
PiperOrigin-RevId: 466301642
(cherry picked from commit 7dc05edbab)
This allows the GlEffectsFrameProcessor to later handle HLG and PQ
differently, or limited and full color range differently.
No functional change intended in this CL.
PiperOrigin-RevId: 466070764
(cherry picked from commit 536d42c865)
In the case where this check fails, the downstream frame processor chain won't be able to handle the incoming (SDR) data anyway as we've already set it up for HDR.
PiperOrigin-RevId: 465584814
(cherry picked from commit 2bd1a93d27)
`requestCalculateSsim` more clearly represents the intention of the caller.
Also rephrase the javadoc to simplify it and make it more precise.
PiperOrigin-RevId: 465575578
(cherry picked from commit 2eb59fa1af)
* Add RgbaMatrix interface implementation.
* Add Builder class for easy adjustments.
* Adjust existing RgbaMatrixPixelTests to use new RgbAdjustment class.
PiperOrigin-RevId: 465545429
(cherry picked from commit a76dbfd363)
Adds a new event to AudioOffloadListener to get the offload state of the track, which indicates when software decoding is taking place.
PiperOrigin-RevId: 465264362
(cherry picked from commit 7893531888)
Avoids disabling Offload on a write error, and instead relies on this being disabled on the AudioTrack init. It will no longer recover by disabling offload.
PiperOrigin-RevId: 465248917
(cherry picked from commit a10af8ecda)
Format.NO_VALUE is a placeholder value for an invalid @C.ColorTransfer, used
for example when the decoder doesn't support this transfer function.
When encountering this invalid value, interpret this as COLOR_TRANSFER_SDR.
Confirmed locally that an exception is thrown when transcoding on p4head, and no exception is thrown when transcoding with this CL.
PiperOrigin-RevId: 464135080
(cherry picked from commit 687a50e9b3)
This will allow effects preview in ExoPlayer to use the
Effect and FrameProcessor interface (and the interfaces
they depend on) without depending on transformer or the
future effects module.
PiperOrigin-RevId: 464060047
(cherry picked from commit 480c129c54)
* Sets KEY_HDR_STATIC_INFO from MediaFormat in the DefaultCodec.
* Adds checks in mediaparser to ensure color space, range, and transfer are valid
values.
PiperOrigin-RevId: 463921325
(cherry picked from commit 65a2a3a0a0)
Despite unregistering the callback and clearing pending Handler
messages, the callback may still receive pending calls if they
are already triggered by the AudioTrack. Instead of asserting
that the track is correct, we should gracefully ignore stale
events.
PiperOrigin-RevId: 463851393
(cherry picked from commit e331bf2c53)
Once a controller bound to a MediaSessionService unbinds, this service
may be destroyed immediately if this was the last bound controller and
the service isn't started in the foreground.
At the time of releasing the controller, there may still be pending
commands that need to be handled by the session service. These commands
may cause the session service to post a foreground notification to
keep it alive. So to avoid the destruction of the service while these
commands are still handled, we need to keep the controller bound.
We also add a timeout in case the session tasks are never completed
by the session implementation. In case the controller is destroyed,
the unbinding happens automatically by the system.
PiperOrigin-RevId: 463802220
(cherry picked from commit ac42b593cc)
TransformationException error codes were previously tied to OpenGL
but other FrameProcessor implementations are possible. So this CL
renames the error codes.
Also, remove GL_INIT_FAILED error code, as FrameProcessor
implemenations may initialize resources on a background thread
after the factory method returns, so it's not obvious how to
distinguish between initialization failures and processing failures.
PiperOrigin-RevId: 463704902
(cherry picked from commit 0f96d8615d)
Both are used in the public FrameProcessor interface, so they
should be public too.
PiperOrigin-RevId: 463454859
(cherry picked from commit b1b9ba825f)
Just a misc nit, since I found the name a bit confusing, and figured
findEncoderWithClosestSupportedFormat might be more descriptive.
PiperOrigin-RevId: 463433646
(cherry picked from commit a7a17dc2bb)
Extract a FrameProcessor.Factory interface from GlEffectsFrameProcessor
and allow it to be customized using a setter on Transformer.Builder.
PiperOrigin-RevId: 463433438
(cherry picked from commit 22822d8e19)
Size requires API 21. Using Pair instead will allow effects to be
used from API 18 during previewing once they are moved out of
transformer.
PiperOrigin-RevId: 463206474
(cherry picked from commit 4adf199398)
This is needed for applying effects to a playlist.
The effects are applied based on the presentation time of the
frame in its corresponding media item and the offset is added
back before encoding.
Each time the offset changes, end of input stream is signalled
to the texture processors. This is needed because the texture
processors can expect monotonically increasing timestamp within
the same input stream but when the offset changes, the timstamps
jump back to 0.
PiperOrigin-RevId: 462714966
(cherry picked from commit 30e5bc9837)
The DefaultMediaNotificationProvider caches the last loaded artwork
bitmap so that the bitmap isn't loaded again when the notification is
updated, e.g., the player is transiting from playing to paused. However,
loading bitmap requests for bitmaps that are already being loaded are
not suppressed. For example, if the notification is updated while the
artwork is still downloading, the same artwork might be downloaded
multiple times.
This change suppresses a bitmap load request if the same artwork is
still being loaded, to avoid additional artwork downloads.
#minor-release
PiperOrigin-RevId: 462572221
(cherry picked from commit dd2c16bc45)
It's always safe to ignore the result of these methods, because the
caller already has a reference to the returned value.
PiperOrigin-RevId: 462388947
(cherry picked from commit 2deb435625)
We currently clear all pending messages, including the one that flushes
pending commands to the MediaSession. To ensure all commands that have
been called before controller.release() are still sent, we can manually
trigger the flush message from the release call.
Related to handling the final flush because disconnecting the controller,
MediaSessionStub didn't post the removal of the controller to the
session thread, creating a race condition between removing the controller
and actually handling the flush.
Issue: androidx/media#99
PiperOrigin-RevId: 462342860
(cherry picked from commit ee209690cb)
These constructors are currently very intertwined, passing `this`
references from the constructor of one to the constructor of another
before the first constructor is complete (and so the `this` reference
isn't really valid yet).
This change uses checker framework `@UnderInitialization` and
`@NotOnlyInitialized` annotations to make it more clear that the
references are not available yet. For the one 'direct' access needed
in the second constructor (calling `getApplicationLooper()`) we now
pass the `applicationLooper` directly alongside (to avoid needing to
dereference the reference 'too early').
This change also ensures that where a class hierarchy has a
'dependent' class hierarchy, the 'subclass' instance is always used
(by both subclass and superclass) without casting or manually hiding
the superclass field, by defining an overridable `getFoo()` method
instead and always using it.
#minor-release
PiperOrigin-RevId: 462335043
(cherry picked from commit 287c757944)
This allows us to use BT.2020 RGB linear for intermediate shaders, which also
allows us to re-enable PeriodicVignetteProcessor, which should work properly in
linear color-spaces.
Manually tested by adding a GlEffectsWrapper, and confirming that HLG HDR editing still looks correct.
PiperOrigin-RevId: 462265821
(cherry picked from commit 2f977eeec9)
