[Webkit-unassigned] [Bug 237017] New: STP140 audio/video issues

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Feb 21 21:02:32 PST 2022


https://bugs.webkit.org/show_bug.cgi?id=237017

            Bug ID: 237017
           Summary: STP140 audio/video issues
           Product: WebKit
           Version: Safari Technology Preview
          Hardware: Mac (Intel)
                OS: macOS 12
            Status: NEW
          Severity: Major
          Priority: P2
         Component: WebRTC
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: glefebvr at epic.com
                CC: youennf at gmail.com

WebRTC audio (and some video) is unusable in Safari Technology Preview 140:

- Local audio MediaStreamTrack from Intel MacBook Pro Microphone is generally not delivered to remote peer, only occasional, discontinuous chunks of audio make it through (unintelligible)
- Also local microphone audio level is generally NOT detected by WebAudio AnalyserNode in STP
- However using Mock Capture Device setting in STP DOES deliver audio (and is detected by AnalyserNode)
- Under this setting, Mock Capture Video is only able to deliver 5-6 fps, even when just viewed locally (???)
- Remote audio MediaStreamTrack OCCASIONALLY is detected by AnalyserNode in STP, but RARELY creates any output (and when it does, it is essentially just "clicks")

Looks like buffer under-runs in console (verbose WebRTC Logs enabled)

LocalAudio (consistently 7 samples short):

[Error] AudioSampleDataSource::pullSamples(61979E3D82774AFC) need more data, sample 1429952 with offset 1034944, trying to get 128 samples, but not completely in range [1430455 .. 1495991]
[Log] AudioSampleDataSource::pushSamplesInternal(61979E3D82774AFC) needed more data, pushing 121 samples

RemoteAudio (consistently 24 samples short):

[Error] AudioSampleDataSource::pullSamples(14665F5E17826BDE) need more data, sample 1258664 with offset 968744, trying to get 480 samples, but not completely in range [1260216 .. 1292984]
[Log] AudioSampleDataSource::pushSamplesInternal(14665F5E17826BDE) needed more data, pushing 456 samples
[Info] RealtimeIncomingVideoSource::videoSampleAvailable(8DF9FB82FB5AD353) 24 frames sent in 1.0015237659999912 seconds
[Info] RealtimeVideoCaptureSource::videoSampleAvailable(F9D25F3356E64DF1) 26 frames sent in 1.0576630170003227 seconds
[Info] RealtimeMediaSource::videoSampleAvailable(14A82F22ADAE606E) 26 frames sent in 1.0577196350000122 seconds
[Info] RealtimeOutgoingVideoSource::sendFrame(F9D25F3356E64DF1) 26 frames sent in 1.0577254170002561 seconds
[Error] AudioSampleDataSource::pullSamples(14665F5E17826BDE) need more data, sample 1301984 with offset 1002464, trying to get 480 samples, but not completely in range [1303536 .. 1336304]
[Log] AudioSampleDataSource::pushSamplesInternal(14665F5E17826BDE) needed more data, pushing 456 samples

I have not such issues on the same machine / hardware (even using the Mock Capture Devices) on Safari Version 15.3 (17612.4.9.1.8)

We've had reports from a particular customer of similar behavior where one party was connected via Safari (though doubtful STP). Is this a chipset incompatibility (i.e. 15.3 works on Intel but not Apple, and STP140/15.4 works on Apple but not Intel)?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20220222/8dfa1dc7/attachment.htm>


More information about the webkit-unassigned mailing list