[Asterisk-Dev] Re: G.729

[Daniel Pocock]

It sounds like the problems you are having are because your phone is 
sending VAD or CNG frames - these frames are smaller than normal frames 
and used to cover silence periods.

I don't know whether Asterisk knows how to forward such variable sized 
frames to my codec implementation, although the underlying Intel code 
can certainly support this.

Maybe someone else on the list knows more about how Asterisk deals with 
variable size G.729B RTP packets.  Otherwise, try disabling VAD/CNG in 
your phone or telling it to use G.729A instead of G.729B.  Every phone 
is slightly different, so you might not be able to change this stuff.

If anyone wants to know more, please see the G.729B spec and at the ITU 
web site and RFC 3389.

Regards,

Daniel


The Traveller wrote:

>Hey Daniel,
>
>Great work!  Now I can finally test-drive this with asterisk, without
>having to purchase a license through Digium and having to cope with the
>limitations in how Digium's version enforces the licensing and locks it
>to a single machine.
>
>After Intel cleared my request for a non-commercial IPP-license, I
>installed it and gave it a spin with a Leadtek BVA8051 SIP to phone-
>box.  Audio seemed good, but on several occasions, I got a flood of these
>on the Asterisk-console:
>
>Sep 24 23:09:53 WARNING[360466]: samples/codec_g729.c:217 g729tolin_framein: Received a G.729 frame that was 2 bytes from RTP
>
>Audio seems to remain good in those cases, but I did get a few spontanious
>disconnects.
>
>Also, when calling app_echo locally, I got choppy audio and some of these:
>
>Sep 24 23:18:41 NOTICE[376850]: frame.c:120 ast_smoother_feed: Dropping extra frame of G.729 since we already have a VAD frame at the end
>
>
>That concludes my observations during the few tests I have done so far.
>
>I haven't tried any other SIP-clients yet.  If you need more specific
>debugging-info, let me know what you want and how to generate it.
>
>
>
>   Grtz,
>
>     Oliver
>  
>