- From: Peter Thatcher <pthatcher@google.com>
- Date: Thu, 30 Jul 2015 21:06:02 -0700
- To: Philipp Hancke <fippo@andyet.net>
- Cc: "public-ortc@w3.org" <public-ortc@w3.org>
- Message-ID: <CAJrXDUEZqM8HK9xeqEzxcPhdAFjw0kY3bP6jTFXiB+1OcN387Q@mail.gmail.com>
On Thu, Jul 30, 2015 at 9:01 PM, Philipp Hancke <fippo@andyet.net> wrote: > On Thu, Jul 30, 2015 at 9:42 AM, Peter Thatcher <pthatcher@google.com> > wrote: > >> You mean that you want the RtpReceiver.track created by the RtpReceiver >> to have a particular track.id instead of being some random thing? >> > > Yes. For proper mapping of the msid behaviour of the PeerConnection API > there is the additional problem of getting the stream id right.... > > I hadn't thought of that, but now that you mention it, it does sound >> reasonable. We can't add track.setId(...). That's outside the scope of >> ORTC. But maybe we could have a RtpReceiver.setTrackId(...) that can only >> be called before you call RtpReceiver.receive() (calling it after would >> either blow up or be a no-op). >> > > How about passing an attribute to .receive()? (I am assuming the track > property does not exist before this is called) > > Then we'd either need to dirty the parameter list or add it to RtpParameters, neither of which seem great. > On the other hand, I think that apps shouldn't use track IDs for anything >> anyway. They should use MIDs instead. So 'd prefer that this weren't >> needed in the first place. >> > > The usecase I have in mind multiparty. In the PeerConnection API I am > using the stream id to associate the onaddstream calls with participants > (think of calling SRD with a bunch of new msid lines). > At a higher level, my code expect to correlate the stream id to a > participant. > > (that said, I don't see a way to control the _stream_ id so I might have > to go back to the drawing board; ORTC makes this simpler since I don't get > callbacks randomly but am in control so it might work) > Why not correlate the MID to the participant? > >> >> On Sun, Jul 26, 2015 at 2:45 PM, Philipp Hancke <fippo@andyet.net> wrote: >> >>> i am thinking about mapping >>> https://tools.ietf.org/html/draft-ietf-mmusic-msid-10 >>> >>> At the javascript level, the gist of this is (as I understand it) to >>> allow a correlation between SDP operations and onaddstream. For multiparty >>> applications (using plan B) I add a new msid in the SDP and sometime later >>> get an onaddstream and need some way to correlate that, in particular when >>> adding multiple new streams at the same time. >>> >>> Now ignoring that with ORTC i'm in charge and don't get streams/tracks >>> popping up at random... >>> is there a way I can control the id of the Receiver's track? The track's >>> id is readonly so I can't change it. >>> >> >> >
Received on Friday, 31 July 2015 04:07:10 UTC