Re: [AVT] I-D ACTION:draft-ietf-avt-rtp-h264-05.txt

[Magnus Westerlund <magnus.westerlund at ericsson.com>]

Hi Philippe,

Thanks for the review. Below I try to answer your questions.

philippe.gentric@philips.com wrote:
> AVTers, authors,
>
> I have found this draft pretty advanced, I have a few questions
>
> there is the text related to parameter sets that I do not fully understand,
> it is distributed across 1.2 (which is fine) and 6.1 at 8.4
> and I suspect section 6.1 needs to be revisited/merged
> due to the new dedicated section 8.4 ?
I can agree that at least the  following bullet:

"   o Sequence and picture parameter set NAL units MUST NOT be sent in
      an RTP session whose parameter sets were already changed by
      control protocol messages during the lifetime of the RTP session.

      Informative note: In many applications, parameter sets can be
      conveyed before establishing an RTP session by using the optional
      parameter-sets MIME parameter in an SDP description.  Thus, when
      no control protocol is used to convey parameter sets during the
      RTP session, the use of sequence and picture parameter set NAL
      units within the RTP packet stream is allowed in these
      applications."

should be removed in favour of the new section 8.4. I think we missed to 
review this fully for consistency. I do think that the next bullet 
concerning handling of NALUs in gateways belong here, although it talks 
a little bit about parameter sets.

> both in section 6.1 and section 8.4 there is the concept of:
>
> "an RTP session whose parameter sets were already changed by control 
> protocol messages during the lifetime of the RTP session"
>
> Now apart from the fact that there seem to be a duplication here between 6.1 and 8.4,
> I have to admit this got me a headache to understand (and actually I am not sure I got it ?)
>
> Can we strcuture this a bit by explaining that there are 3 ways to send parameters sets:
>
> (1) out of band before RTP session starts
>
> (2) out of band during RTP session
>
> (3) in band (of course, during RTP session)
>
> If yes , Would it be possible that what you want to express is that:
>
> (A) parameter sets are transmitted using either (1) or (3) or (1+2) or (1+3),
>
> (B) (1) and (1+2) are the RECOMMENDED way to do it
>
> (C) (3) and (1+3) are only for some specific applications and are NOT RECOMMENDED
>
> (D) in all cases (2) and (3) MUST NOT be used together in the same session
>
> Is this is what you wanted to express ?
Yes, that is what we try to express in section 8.4. Plus a few things 
other in regards to how to add parameter sets to avoid problems, and the 
usage of reliable protocols. I will take your input that it is hard to 
read and try to structure the section better.



> regards,
>
> PS: in section 8.4 there is a typo in the word "specificaion"
I will take care of this.

If more people has comments they are appreciated if they can be sent no 
later than monday. I will try to do a update in the beginning of the 
next week, and submit that as soon as possible.

Cheers

Magnus Westerlund

Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB                | Phone +46 8 4048287
Torshamsgatan 23           | Fax   +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@ericsson.com


_______________________________________________
Audio/Video Transport Working Group
avt@ietf.org
https://www1.ietf.org/mailman/listinfo/avt