Re: [payload] Adam Roach's Discuss on draft-ietf-payload-rtp-ancillary-10: (with DISCUSS and COMMENT)

[Colin Perkins <[e-mail]>]

> On 29 Sep 2017, at 22:00, Adam Roach <[e-mail]> wrote:
> On 9/26/17 8:13 PM, Thomas Edwards wrote:
>> Regarding draft-ietf-payload-rtp-ancillary-10, on further thought I believe that FID grouping is NOT appropriate for Section 4.1 “Grouping ANC Streams with other Media Streams”.
>> 
>> RFC 5888 Section 8.4. (“FID Semantics”) states:
>> 
>>        “It is assumed that the application uses only one codec at a time to encode the media produced.  This codec MAY change dynamically during the session, but at any particular moment, only one codec is in use.”
>> 
>> RFC 5888 Section 8.5.1 states:
>> 
>> 	“FID semantics are useful when the application only uses one codec at a time.  An application that encodes the same media using different codecs simultaneously MUST NOT use FID to group those media lines.”
>> 
>> When we have one video stream and one ancillary data stream, senders are not switching between codecs.  FID really sounds to me like a situation where you have one media stream, but senders could send that stream one of multiple codecs, but not at the same time.
> 
> In this case, where you have data and meta-data, it's arguably two different codecs; and, since they don't represent the same data, you're encoding with one codec or the other, based on whether you're sending video data or meta-data. From that perspective, I don't see the use of FID as being outside the spirit of what is intended.
> 
> That said, if the WG agrees with your evaluation that FID isn't quite the right fit, I'm happy for y'all to come up with some alternate solution...

I tend to agree with Adam that FID is appropriate to group the ancillary data with its associated media stream. The metadata in the ancillary stream is conceptually part of the media stream, which matches the definition of FID. They MUST be sent with the same CNAME, to ensure they can be synchronised, but that doesn’t require any explicit signalling.

You can use LS signalling in addition to FID if two streams, with their associated ancillary streams, need to be synchronised. 

-- 
Colin Perkins
https://csperkins.org/