This behavior was previously available as opt-in via
`MediaItem.DrmConfiguration.Builder.setPlayClearContentWithoutKey` and
`DefaultDrmSessionManager.Builder.setPlayClearSamplesWithoutKeys`. This
change flips the default of both these properties to true.
This should speed up the time for playback to start when playing DRM
content with a 'clear lead' of unencrypted samples at the start.
Previously playback would wait until the keys for the later encrypted
samples were ready. The new behaviour could result in mid-playback
stalls/rebuffers if the keys are not ready yet by the transition from
clear to encrypted samples, but this is not really a regression since
previously playback wouldn't have started at all at this point.
PiperOrigin-RevId: 595992727
These methods sound similar, but have different behaviour. This change
tries to make the distinction clearer, and sign-post from one to the
other.
#minor-release
Issue: androidx/media#910
PiperOrigin-RevId: 595701540
While investigating Issue: androidx/media#887 I naively assumed the CEA-608
captions were in a TS file, but they're actually in an MP4 (which is
possibly obvious given DASH only supports MP4). This change includes
container info in the `EventLogger` `tracks` output.
PiperOrigin-RevId: 592192752
Move `inputColorInfo` from `VideoFrameProcessor`'s `Factory.create` to `FrameInfo`,
input via `registerInputStream`.
Also, manually tested on exoplayer demo that setVideoEffects still works.
PiperOrigin-RevId: 591273545
The timestamp adjuster also estimates the number of wraparounds
of the 90Khz TS timestamp. It does that by assuming that a new
timestamp is always close to the previous one (in either direction).
This logic doesn't always work for duration estimates because the
timestamp at the end of the media is not close to the one at the
beginning and it may also never be less than the one at the beginning.
This can be fixed by introducing a new estimation model that assumes
the new timestamp is strictly greater than the previous one without
making the assumption that it has to be close to it.
Issue: androidx/media#855
#minor-release
PiperOrigin-RevId: 590936953
Vertical and horizontal tile counts inside Format used by ImageRenderer
used to be required to equal 1 and were set to 1 regardless of what
the manifest said.
This change removes the above requirement and sets the tile counts to
the values from the manifest.
PiperOrigin-RevId: 590608353
Despite GL 3.0 not being required on API 29+, it is experimentally
determined to always be supported on our testing devices, on API 29+.
That said, still fall back to OpenGL 2.0 if 3.0 is not supported,
just in case.
PiperOrigin-RevId: 590569772
The serializable form is used when we need to serialize the result into
bytes in the sample queue. The binder-based (ultimately
filedescriptor-based) form is used for
session/controller IPC, in order to avoid sending the bitmap bytes over
the IPC.
Issue: androidx/media#836
#minor-release
PiperOrigin-RevId: 588420836
`fromBundle` doesn't distinguish between `FIELD_BITMAP` and `FIELD_TEXT`
being present with a null value, or being absent, so we might as well
avoid including them when the value is null.
I've separated this from a later change to add
`Cue.toSerializableBundle` which will also skip setting a bitmap value
into the `Bundle` if `this.bitmap == null`. This is partly because it
results in changes to a lot of extractor test dump files, and it's
easier to review that as a separate change.
PiperOrigin-RevId: 586626141
In `TextOverlay` and `DrawableOverlay`, treat `Bitmap` as a buffer, where we
allocate it rarely and reuse it as long as possible before making a new one.
In `BitmapOverlay`, avoid allocating GL textures too often as well.
Strongly reduces allocations and memory usage growth (saving ~100-150 MB on 4k60fps
at high end), at the cost of more code complexity and low-end using 70MB more, on
1/1 comparisons.
PiperOrigin-RevId: 585990602
Split CompositingVideoSinkProvider.VideoSinkImpl in two classes:
- VideoSinkImpl now only receives input from MediaCodecVideoRenderer and
forwards frames to its connected VideoFrameProcessor
- VideoFrameRenderControl takes composited frames out of the VideoGraph
and schedules the rendering of those.
- CompositingVideoSinkProvider connects VideoSinkImpl with
VideoFramesRenderer.
PiperOrigin-RevId: 584605078
Currently, ads are only defined by a single URL, which makes it
impossible to define additional fields needed to play ads correctly.
This can be fixed by using a full MediaItem in AdPlaybackState,
replacing the previous Uri field.
PiperOrigin-RevId: 582331588
MIME types are case-insensitive, but none of the many existing
comparisons across our code base take this into account. The
code can be made more robust by normalizing all MIME types at the
moment they are first set into a class/builder and adding toLowerCase
as part of the normalization.
Most concretely, this fixes an issue with playing HLS streams via
the IMA SDK where the stream MIME type is indicated with all lower
case "application/x-mpegurl", which failed the MIME type comparison
in DefaultMediaSourceFactory.
PiperOrigin-RevId: 582317261
All methods in VideoFrameProcessor are expected to be called by the owning thread,
as far as I understand (vs. 10 threads each queuing frames/textures/streams, which
invalidates blocking done by registerInputStream and flush)
PiperOrigin-RevId: 582295240
Based on on-device testing, this device seems to have the same issue as Moto G (20) where frames are dropped despite configuring the decoder not to drop frames.
PiperOrigin-RevId: 581943805
The dump file diff as part of this change is because using AudioGraph
means the 2nd item is automatically edited to match the AudioFormat of
the 1st item {44.1KHz mono}, rather than {48KHz stereo}.
Manually verified that for the 2nd item, data output:
* Before: 66936 bytes (16734 frames) output = 348_625us of audio.
* After: 30750 bytes (15375 frames) output = 348_639us of audio.
The small final buffer is caused by SonicAudioProcessor outputting all
pending data when EOS queued, and is WAI.
PiperOrigin-RevId: 580494578
Both interfaces are not really needed as the methods can be called
on the respective classes directly. This also avoid error-prone
situations where classes define to/fromBundle methods with parameters
that should be used instead of the parameter-less version.
Also deprecate all existing CREATOR static instances and make the
corresponding fromBundle method public where needed.
PiperOrigin-RevId: 579189766
BundleableUtil is the only class that really depends on the type
inheritence from Bundleable. However, it only needs this as a way
to define Function<T, Bundle> and Function<Bundle, T>, which could
just be passed in as parameters as it's already done for some of
these methods.
Also rename the class to BundleCollectionUtil accordingly.
PiperOrigin-RevId: 578791946
This change applies to standalone TTML files linked directly from the manifest.
As a result, we no longer have the flakiness in the DashPlaybackTest which uses sidecar-loaded (standalone file) TTML subtitles. We experimentally opt into parsing subtitles during extraction and use SubtitleExtractor in hybrid mode.
PiperOrigin-RevId: 577457256
This CL adds an isPrioritizeImageOverVideoEnabled flag to
TrackSelectionParameters and an API to set the flag value.
The flag will be used by DefaultTrackSelector to determine whether
to select an image track if both an image track and a video track are available.
PiperOrigin-RevId: 576175162
A few methods in PlayerInfo and related classes combine filtering
information with bundling in one method. This makes it impossible
to use just the filtering for example and it's also easier to reason
about than two dedicated methods. This change splits these methods
into two parts accordingly.
PiperOrigin-RevId: 572592458
When used within the same process, we don't have to go via the
onTransact method (which includes marshalling and unmarhsalling
the data), but can directly return the list.
#minor-release
PiperOrigin-RevId: 572179846
The flag is no longer used by our components and only set and checked
in a few places to guarantee compatiblity with existing renderers and
decoders that still use it.
The flag will be removed in the future due to its design limitations.
#minor-release
PiperOrigin-RevId: 571291168
location and index are mostly equivalent, except that it's possible that location is more up-to-date than index, so we can just always use location.
PiperOrigin-RevId: 570950407
The per-stream methods are generally preferred and having a reference
to them from the device-wide methods may help with discoverability.
Issue: google/ExoPlayer#11295
PiperOrigin-RevId: 567604785
Adjust the Javadoc to highlight that data of this MIME type just contains a URI, not the actual image content.
And also remove the superfluous "key" part of the MIME type string that doesn't
really add information (and it's also really just an URI, not an URI key).
PiperOrigin-RevId: 567560238
Move audio offload mode related interfaces and definitions from `TrackSelectionParameters` to a new `AudioOffloadModePreferences` class.
PiperOrigin-RevId: 566905017
This method isn't used by the library (since <unknown commit>).
It doesn't really work well (e.g. arbitrarily defaults to `MUSIC` when
`UNKNOWN` would be a better default). There's no suggested replacement.
PiperOrigin-RevId: 566676744
This change introduces two new types of method to `SubtitleParser`:
1. `parse()` methods that take a `Consumer<CuesWithTiming>` and return `void`
2. `parseToLegacySubtitle` method that returns `Subtitle`
(1) ensures that in the new 'parse before SampleQueue' world we can
write cues to the `SampleQueue` as soon as they're ready - this is
especially important when parsing monolithic text files, e.g. for a
whole movie.
(2) ensures that during the transition, the legacy 'parse after
SampleQueue' behaviour doesn't see any regressions in 'time to first
cue being shown'. Previously we had a single implementation to convert
from `List<CuesWithTiming>` to `Subtitle`, but this relies on the
complete list of cues being available, which can take a long time for
large files in some formats (with ExoPlayer's current parsing logic).
By allowing implementations to customise the way they create a
`Subtitle`, we can directly re-use the existing logic, so that the
'time to first cue being shown' should stay the same.
This change migrates all **usages** to the new methods, but doesn't
migrate any **implementations**. I will migrate the implementations in
follow-up CLs before deleting the old list-returning `parse()` methods.
PiperOrigin-RevId: 565057945
Previously any `CuesWithTiming.durationUs` could be `TIME_UNSET`,
meaning it should be replaced by the next `CuesWithTiming` instance
(instead of being merged if the durations overlap, which is currently
the expected behavior for all `CuesWithTiming` with a 'real' duration).
This technically allowed a single subtitle track to include a mixture of
`CuesWithTiming` that should be merged, and some that should be
replaced. This is not actually needed for any of the subtitle formats
currently supported by ExoPlayer - in all cases a format expects either
all cues to be merged, or each cue to replace the previous one.
Supporting this mixture of merging and replacing in `TextRenderer` ended
up being very complicated, and it seemed a bit pointless since it's not
actually needed. This change means a given subtitle track either merges
**all** cues (meaning `CuesWithTiming.durationUs = C.TIME_UNSET` is not
allowed), or **every** cue is replaced by the next one (meaning
`CuesWithTiming.durationUs` may be set (to allow for cues to 'time out',
needed for CEA-608), or may be `TIME_UNSET`).
This value will be used in a subsequent change that adds cue-merging
support to `TextRenderer`.
PiperOrigin-RevId: 565028066
This is equivalent to the existing `scaleLargeTimestamp` method with the
following changes/improvements:
* No longer specific to timestamps (there was nothing inherently
time-specific about the logic in `scaleLargeTimestamp`, but the name
and docs suggested it shouldn't be used for non-timestamp use-cases).
* Additional 'perfect division' checks between `value` and `divisor`.
* The caller can now provide a `RoundingMode`.
* Robust against `multiplier == 0`.
* Some extra branches before falling through to (potentially lossy)
floating-point math, including trying to simplify the fraction with
greatest common divisor to reduce the chance of overflowing `long`.
This was discussed during review of 6e91f0d4c5
This change also includes some golden test file updates - these
represent a bug fix where floating-point maths had previously resulted
in a timestamp being incorrectly rounded down to the previous
microsecond. These changes are due to the 'some more branches' mentioned
above.
PiperOrigin-RevId: 564760748
Instead of invoking "undefined behavior" may occur when a class is used
incorrectly, which is clear but not very helpful to callers, specify what Apps
or users of a class should do when issues occur.
Apply this to Compositor as well, where this otherwise might seem unclear.
PiperOrigin-RevId: 563406219
This wasn't properly documented yet (unlike the longer setMediaItems
method that already includes this documentation).
Issue: androidx/media#607
PiperOrigin-RevId: 561910073
This will help with debugging subtitle sync issues as it will show up in
track selection details emitted to logcat by `EventLogger`.
PiperOrigin-RevId: 561355836
Also use `@linkplain` in more places: If I was already touching a javadoc
block, I switched to `@linkplain` throughout the whole block.
PiperOrigin-RevId: 561320273
The first call to method `registerInputStream` doesn't block.
Later successive calls to the method blocks until the previous register call
finishes.
PiperOrigin-RevId: 559694490
Replace Queue<Long> with LongArrayQueue which provides queue semantics
for long primitives. LongArrayQueue is forked from IntArrayQueue which
in turn was forked from Androidx CircularIntArray.
IntArrayQueue is deleted and we now use CircularIntArray directly from
Androidx Collection.
PiperOrigin-RevId: 559129744
