Re: [AVTCORE] [Technical Errata Reported] RFC4733 (3451)

[Xavier Marjou <[e-mail]>]

Hi,

I also agree that this text proposal is better.

I would even suggest to remove the end of the sentence (i.e. "to
simplify the generation of audio waveforms at a gateway") in order to
make it easier to understand (as we would add " if they use the same
SSRC ..." before).

In other words, I would suggest :
 Named telephone events are carried as part of the audio stream and
 if they use the same SSRC (therefore the same timing and sequence
 number space), they MUST use the same timestamp clock rate as the
 regular audio channel.

Cheers,
Xavier

> On 2013-01-16 17:45, Marc Petit-Huguenin wrote:
>> On 01/15/2013 08:54 AM, Magnus Westerlund wrote:
>>> Hi,
>>
>>> (As Individual)
>>
>>> Are we certain that the RTP payload format will never be sent
>>> over two different but linked SSRCs to the original audio. There
>>> is such a use case (See RFC 3388) where DTMF and original audio
>>> are sent in two different RTP sessions and thus with different
>>> SSRCs (although, they could have the same number for binding
>>> reasons). I fear the proposed corrected text is to restrictive.
>>
>> Right.  I propose this:
>>
>> Named telephone events are carried as part of the audio stream and
>> if they use the same SSRC (therefore the same timing and sequence
>> number space), they MUST use the same timestamp clock rate as the
>> regular audio channel to simplify the generation of audio waveforms
>> at a gateway.
>>
>
> I think this is much better.
>
> Cheers
>
> Magnus Westerlund