Re: Mapping RTX in SDP to ORTC RtpParameters

2016-03-04 16:55 GMT+01:00 Philipp Hancke <fippo@goodadvice.pages.de>:
> I just tried to parse the chrome SDP I showed earlier and this seemed quite
> messy.

I don't think it's so hard:

- Get all the SSRC values (typically 2 if it's a=sendrecv or
a=sendonly). Assume first one is for real media codecs and second one
for RTX and other stuff.

- Get all the "a=rtpmap:CC" lines that contain real audio/video codecs.

- Create a RtpEncodingParameters (encoding) object for each one and
set encoding.codecPayloadType=CC and .ssrc=SSRC1 according.

- Parse all the "a=fmtp:RR apt CC" lines and keep a map with CC as key
and RR as value.

- Iterate such a map. For each CC key take the encoding with
codecPayloadType=CC and add fill its rtx object by setting
payloadType=RR and .ssrc=SSRC2.

- If a codec for RTX is still needed (what for?) just create a fake
codec entry for them. But again: not needed at all given the above
information.


> I shrugged but came up with a question... how do you determine your
> local FEC capabilities if rtx is not a codec?
>
> Does
>
> http://ortc.org/wp-content/uploads/2015/11/ortc.html#dictionary-rtcrtpcodeccapability-members
> need a rtx member as well?

I've not investigated the ORTC capabilities in depth yet, but I expect
a simple boolean .rtx should be enough.


-- 
Iñaki Baz Castillo
<ibc@aliax.net>

Received on Friday, 4 March 2016 16:19:02 UTC