Re: "a=extmap" with BUNDLE
Iñaki Baz Castillo <[email protected]>
| Newsgroups | gmane.ietf.mmusic |
|---|---|
| Message-ID | <CALiegfkfq5Li9hnKNNNrBpzgi7zacgBv4EM4Pa2c0tXoWXkFXQ@mail.gmail.com> |
2017-03-27 19:44 GMT+02:00 Taylor Brandstetter <[email protected]>: > If extension IDs aren't unique across bundled m= sections, then on receiving > a packet, an implementation may need to determine which m= section it's > "associated" with before knowing how to interpret the extension IDs. Yes. > To do > this, it may need to use the MID, which means that at a minimum, the MID > extension must be using a unique ID. Oh no, not at all. It's just about getting the appropriate RtpReceiver associated to the current RTP packet (this can be done by MID, RID, SSRC or even PT, as the ORTC and JSEP "RTP Matching Rules" define). Once we have the proper RtpReceiver, pass the packet to it so it can set the corresponding RTP extension ID mappings (because those params belong to each RtpReceiver), and later, let the transport get transport related IDs (such as REMB or transport-cc) from the RTP packet. Well, I implemented it yesterday so... :) https://github.com/ibc/mediasoup/blob/master/worker/src/RTC/Transport.cpp https://github.com/ibc/mediasoup/blob/master/worker/src/RTC/RtpStreamRecv.cpp#L57 > But this increases implementation complexity without much benefit (the > benefit being that by sharing IDs, you could more easily avoid running out > of spare IDs). So I'd be in favor of just requiring them to be unique. Me too. -- Iñaki Baz Castillo <[email protected]> _______________________________________________ mmusic mailing list [email protected] https://www.ietf.org/mailman/listinfo/mmusic