fix(codec-selection): Select VP8 as default codec.

[jallamsetty1]

When RTCRtpReceiver.getCapabilities() is not supported by browser or if it returns an empty list. Fixes an issue where call doesn't get established in Firefox ESR 102. Fixes jitsi/jitsi-meet#13839.

[mir07]

When will this commit hit meet.jit.si?
Video is completely broken in all stable versions of Firefox. Sound ok (opus) but video is falling since video codec cannot be chosen. 122.0 (64-bit) (opus, NA)
Working 115.7.0esr (opus, VP8)

[damencho]

This had been on meet.jit.si for months.

[mir07]

Maybe, but the problem persists.

[jallamsetty1]

Are you testing meet.jit.si? We applied few Firefox related fixes in the latest release.

[mir07]

I have found the problem ;-)
Firefox 122.0 from Debian Sid: No video codec found so video not working
Firefox 122.0.1 from official Mozilla repo: Video codec found so video working

It could be nice to find out the difference between 122.0 and 122.0.1?

[damencho]

Probably you can narrow down the change by using https://mozilla.github.io/mozregression/documentation/usage.html

[jallamsetty1]

Firefox 122.0 from Debian Sid: No video codec found so video not working

When you say no video codec found, are you looking at the browser capabilities? Can you check if the following command when executed on the browser console returns any codecs?

window.RTCRtpReceiver?.getCapabilities('video')?.codecs

[mir07]

Seems pretty identical:
Firefox 122.0.1
Array(7) [ {…}, {…}, {…}, {…}, {…}, {…}, {…} ]
​0: Object { clockRate: 90000, mimeType: "video/VP8", sdpFmtpLine: "max-fs=12288;max-fr=60" }
​1: Object { clockRate: 90000, mimeType: "video/VP9", sdpFmtpLine: "max-fs=12288;max-fr=60" }
​2: Object { clockRate: 90000, mimeType: "video/H264", sdpFmtpLine: "profile-level-id=42e01f;level-asymmetry-allowed=1;packetization-mode=1" }
​3: Object { clockRate: 90000, mimeType: "video/H264", sdpFmtpLine: "profile-level-id=42e01f;level-asymmetry-allowed=1" }
​4: Object { clockRate: 90000, mimeType: "video/ulpfec" }
​5: Object { clockRate: 90000, mimeType: "video/red" }
​6: Object { clockRate: 90000, mimeType: "video/rtx" }
​length: 7

Firefox 122.0
Array(7) [ {…}, {…}, {…}, {…}, {…}, {…}, {…} ]
​0: Object { clockRate: 90000, mimeType: "video/VP8", sdpFmtpLine: "max-fs=12288;max-fr=60" }
​1: Object { clockRate: 90000, mimeType: "video/VP9", sdpFmtpLine: "max-fs=12288;max-fr=60" }
​2: Object { clockRate: 90000, mimeType: "video/H264", sdpFmtpLine: "profile-level-id=42e01f;level-asymmetry-allowed=1;packetization-mode=1" }
​3: Object { clockRate: 90000, mimeType: "video/H264", sdpFmtpLine: "profile-level-id=42e01f;level-asymmetry-allowed=1" }
​4: Object { clockRate: 90000, mimeType: "video/ulpfec" }
​5: Object { clockRate: 90000, mimeType: "video/red" }
​6: Object { clockRate: 90000, mimeType: "video/rtx" }
​length: 7

[jallamsetty1]

Thanks for the codec info, they are identical like you said. Do you see any errors? Will you be able to try the call on beta.meet.jit.si and share the conference details with me so that I can see the client side and backend logs?

[mir07]

The meeting: https://beta.meet.jit.si/RespectivePoetsTelephoneCleverly
Debian has even updated their Sid repo so the version is now 122.0.1 (64-bit) same version as latest Mozilla build but it still does not work with video. To be absolutely certain I connected in a private browser window but no dice. I will stop using the debian build for now and stick to the mozilla build since obviously the debian maintainers brakes firefox.

[mir07]

I have filed a bug to Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1063420
Another bug report showing the same problem: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1061495 which points to this bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1875201

[mir07]

A working work around in Debian is the following:
media.encoder.webm.enabled true
media.ffmpeg.encoder.enabled true
media.webrtc.platformencoder true
media.webrtc.platformencoder.sw_only false
media.webrtc.software_encoder.fallback true

[jallamsetty1]

Thank you @mir07 for debugging this. https://bugzilla.mozilla.org/show_bug.cgi?id=1875201 does seem like the root cause.