[webrtc-extensions] Mixed-codec simulcast proposal (#43)

Orphis has just created a new issue for https://github.com/w3c/webrtc-extensions:

== Mixed-codec simulcast proposal ==
It would be interesting to give users more control over which codec is used for each simulcast layer.
We used to have `codecPayloadType` that took at payload type (codec id negotiated after O/A) but it was removed as it wasn't well designed (PT not known before O/A) and no implementations supported it.

We have a few use cases in mind where mixed codecs could be used, for example using a lower layer with a higher complexity codec (for example software AV1) to get more efficiency and a higher level with another relatively simpler codec (VP8/9).

I imagine that the each simulcast layer could have an array of codecs, listed in the same way as with `setCodecPreferences()`.

Goals are: 

- Filter out unwanted codecs for specific layers
- Override the priority of the codecs negotiated during the O/A.
- Per-layer control of features like FEC or RTX.
- **CAN'T** add new codecs with this API, we are limited to what is in the O/A.

Let me know if you want a more detailed API proposal with IDL.

Please view or discuss this issue at https://github.com/w3c/webrtc-extensions/issues/43 using your GitHub account