Re: [AVT] Re: VMR-WB RTP Payload and Storage Formats

On 7 Oct 2004, at 03:20, Qiaobing Xie wrote:
> Colin Perkins wrote:
>> I would have thought the conference bridge could act as an RTP 
>> translator, and resample as appropriate for some participants?
>
> Use this as an example - here we have the advanced codec vmr-wb saying 
> "you no longer need to worry about resampling at the bridge. Just send 
> me the frames and I can handle them at either rates (8k or 16k)." But 
> RTP payload says: "No, you cann't do that, even I know you are capable 
> of it. You still have to stick with the tradition and pay the cost of 
> signaling the rate change and building the resampling function into 
> the bridge." I think this as a failure of taking full advantage of the 
> codec capability.

I don't disagree. However, we need to develop a solution that retains 
the utility and extensibility of RTP and RTP implementations.

My concern with a fixed 16 kHz rate for VMR-WB is that it doesn't allow 
payload format independent implementations, since it offers a differing 
interpretation of the RTP timestamp than do other audio payload 
formats. To some extent that's unavoidable when we move to variable 
rate codecs, but it would greatly benefit implementers if we could 
develop a single approach to supporting variable rate codecs.

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

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