Bug 174508 – Increase CoreAudio render audio buffer sizes for WebRTC

• Status: RESOLVED
• Resolution: FIXED
• Component: Media

[youenn fablet]

This will increase a bit latency but will make audio rendering more stable

[youenn fablet]

Created attachment 315437 [details]
Patch

[Radar WebKit Bug Importer]

<rdar://problem/33318418>

[Eric Carlson]

Comment on attachment 315437 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=315437&action=review

> Source/WebCore/platform/mediastream/mac/CoreAudioCaptureSource.cpp:226
> +    return AudioSession::sharedSession().sampleRate() / 100;

This will result in a buffer size of 80 for 8K which is definitely too small. 50 might be better.

> Source/WebCore/platform/mediastream/mac/CoreAudioCaptureSource.cpp:304
> +    // For the moment we do not need to configure speaker proc as we ar not providing any reference data.

Nit: "we ar not" => "we are not"

[youenn fablet]

Thanks for the review.

(In reply to Eric Carlson from comment #3)
> Comment on attachment 315437 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=315437&action=review
> 
> > Source/WebCore/platform/mediastream/mac/CoreAudioCaptureSource.cpp:226
> > +    return AudioSession::sharedSession().sampleRate() / 100;
> 
> This will result in a buffer size of 80 for 8K which is definitely too
> small. 50 might be better.

How about picking the max of AudioSession::sharedSession().bufferSize() and 10ms?

> 
> > Source/WebCore/platform/mediastream/mac/CoreAudioCaptureSource.cpp:304
> > +    // For the moment we do not need to configure speaker proc as we ar not providing any reference data.
> 
> Nit: "we ar not" => "we are not"

ok

[youenn fablet]

Created attachment 315447 [details]
Patch for landing

[youenn fablet]

(In reply to youenn fablet from comment #5)
> Created attachment 315447 [details]
> Patch for landing

I removed the CoreAudioCaptureSource changes.
I'll upload a separate patch removing speaker configuration

[youenn fablet]

Created attachment 315452 [details]
Patch for landing

[WebKit Commit Bot]

The commit-queue encountered the following flaky tests while processing attachment 315452 [details]:

storage/websql/database-lock-after-reload.html bug 174447 (authors: [e-mail] and [e-mail])
The commit-queue is continuing to process your patch.

[WebKit Commit Bot]

The commit-queue encountered the following flaky tests while processing attachment 315452 [details]:

storage/indexeddb/modern/new-database-after-user-delete.html bug 174520 (author: [e-mail])
The commit-queue is continuing to process your patch.

[WebKit Commit Bot]

Comment on attachment 315452 [details]
Patch for landing

Clearing flags on attachment: 315452

Committed r219517: <http://trac.webkit.org/changeset/219517>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.