[webkit-changes] [WebKit/WebKit] cd3d29: [MediaRecorder/WebM] ASSERT(pts >= m_lastMuxedSamp...
Jean-Yves Avenard
noreply at github.com
Mon Oct 14 14:55:24 PDT 2024
Branch: refs/heads/main
Home: https://github.com/WebKit/WebKit
Commit: cd3d290bb0cbc0078f6af58868d427c615ada404
https://github.com/WebKit/WebKit/commit/cd3d290bb0cbc0078f6af58868d427c615ada404
Author: Jean-Yves Avenard <jya at apple.com>
Date: 2024-10-14 (Mon, 14 Oct 2024)
Changed paths:
M Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterWebM.h
M Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterWebM.mm
Log Message:
-----------
[MediaRecorder/WebM] ASSERT(pts >= m_lastMuxedSampleTime);
https://bugs.webkit.org/show_bug.cgi?id=281396
rdar://137830389
Reviewed by Youenn Fablet and Jer Noble.
We were using the start time of the last received audio sample.
Audio is muxed as quickly as possible and will only ever be delayed if
we know that a video frame is incoming (as we don't want to enqueue audio until a video frame,
which may never come is finally available).
On an heavily loaded system, audio could have been received and muxed before
the related video frame got received. To ensure the best A/V sync we will
prefer the clock time at which the video frame got received but handle the
case where audio had already been muxed with a later time.
Manually tested, this is highly timing dependent and I could only reproduce while running in a
debugging session and turning on all the logging.
* Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterWebM.h:
* Source/WebCore/platform/mediarecorder/cocoa/MediaRecorderPrivateWriterWebM.mm:
(WebCore::MediaRecorderPrivateWriterWebM::nextVideoFrameTime const):
(WebCore::MediaRecorderPrivateWriterWebM::appendAudioSampleBuffer):
(WebCore::MediaRecorderPrivateWriterWebM::resume):
Canonical link: https://commits.webkit.org/285147@main
To unsubscribe from these emails, change your notification settings at https://github.com/WebKit/WebKit/settings/notifications
More information about the webkit-changes
mailing list