Decoder Capabilities

[aboba]

As noted in w3c/media-capabilities#182, there are decoder implementations which do not support reference scaling.

Reference scaling support is not purely an SVC issue, but a decoder that does not support reference scaling also will not be able to support spatial scalability modes.

So one potential approach would be for such a decoder to return the scalabilityModes that it does support.

Is this sufficient?

[chcunningham]

For MC, I'm hoping to meet w/ Danil soon to find a path we can all agree on. If getCapabilities() and MC both come to describe scalability, my preference would be that we more or less match in approach.

[aboba]

@chcunningham @alvestrand Proposed resolution is to remove RTCRtpReceiver.getCapabilities(). Instead, the spec will recommend the following:

1. Use MC to determine referenceScaling for the codec.
2. If referenceScaling is true, then the decoder can decode any scalabilityMode that the encoder can send.
3. If referenceScaling is false, then the decoder can decode any scalabilityMode that the encoder can send with the exception of spatial scalability modes (e.g. only temporal modes and S modes, if supported in the codec).

Does this make sense?

[aboba]

Closing, based on merger of PR #54