Temporal layers are ignored

mediasoup libraries
1

hey everyone,

I am using simulcast and I can control how many spatial layers I produce by specifying different encodings, exactly how it is documented and done in mediasoup demo, i.e.:

const WEBCAM_SIMULCAST_ENCODINGS = [
  { scaleResolutionDownBy: 4, maxBitrate: 500000 },
  { scaleResolutionDownBy: 2, maxBitrate: 1000000 },
  { scaleResolutionDownBy: 1, maxBitrate: 5000000 }
];

With this setup on consumer side I get scalability mode S3T3.

If I remove one encoding I get S2T3.

But I can’t control how many temporal layers are there, it is always 3.

I tried this setup:

const WEBCAM_SIMULCAST_ENCODINGS = [
  { scaleResolutionDownBy: 4, maxBitrate: 500000, scalabilityMode: 'S1T1' },
  { scaleResolutionDownBy: 2, maxBitrate: 1000000, scalabilityMode: 'S1T1' },
  { scaleResolutionDownBy: 1, maxBitrate: 5000000, scalabilityMode: 'S1T1' }
];

But it comes to consumer as S3T3 anyway.

Is there anyway to controls number of temporal layers?

I guess there is no actual real stream for temporal layers if we just drop frames, but still.
Or is it?

Thanks!

2

You don’t have actual control over number of temporal layers it in the browser, so it is what browser wants it to be (mediasoup-client assumes it is 3, but it can’t know for sure).

1 Like
3

Oh, I see. thanks!