Speex dev - Aug 2005 - Re: Speex-dev Digest, Vol 15, Issue 15

Seems to me that they are using speex already.
Ethereal shows that the voice RTP stream's payload type is 103. According to
this page [ 
http://www.openh323.org/pipermail/openh323/2004-June/068705.html], 103 is
SpeexNarrow-8k, although this email from Craig might not be correct.

- Cheng
> Message: 2
> Date: Thu, 25 Aug 2005 04:15:44 -0500 (CDT)
> From: Ashhar Farhan <farhan@phonestack.com>
> Subject: [Speex-dev] speex in google talk
> To: speex-dev@xiph.org
> Message-ID: <Pine.LNX.4.62.0508250413360.1266@ubgsbba.pbz>
> Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
> 
> hi,
> google talk is evaluating speex. congratulations to everybody.
> check http://www.google.com/talk/developer.html#protocols
> 
> - farhan

At 03:07 AM 8/26/2005, you wrote:
>Seems to me that they are using speex already.
>Ethereal shows that the voice RTP stream's payload type is 103.
According to
>this page [
>http://www.openh323.org/pipermail/openh323/2004-June/068705.html], 103 is
>SpeexNarrow-8k
This is incorrect.  103 (being >= 96) is a dynamic payload type code, 
and is not (and cannot be) 'registered' to any one 
codec.  "SpeexNarrow-8k" can use any dynamic payload type code, and 
any other codec can use dynamic payload type 103, as specified by a 
"a=rtpmap:" line in the stream's SDP description.

The actual list of registered RTP payload format codes can be found 
here: <http://www.iana.org/assignments/rtp-parameters>.  Note that 
this registry of static payload format codes is now closed; no more 
static payload format codes will be registered.


	Ross Finlayson
	LIVE.COM
	<http://www.live.com/>