Hi Anthony, Thank you very much for your response. Sorry, I meant that we are using the _latest_ git HEAD at time of writing. I have now tried passthru_ptime_mismatch=true, and it fixed the recording of some of the cases of asynchronous ptime, but also broke other previously "working" cases. Kind regards, Colin On 8 June 2012 21:03, Anthony Minessale <anthony.minessale@gmail.com> wrote:
have you tried updating lately? also try global_setvar passthru_ptime_mismatch=true
On Fri, Jun 8, 2012 at 3:29 AM, <colins@succinct.co.za> wrote:
Hi,
Please, I was wondering if there is anything obvious that we are missing when trying to record a G729 call with asynchronous pimes (in: 20ms, out: 60ms).
The problem case call flow is: inbound leg (ptime 20 ms), start recording, playback a short message, bridge to outbound (destination can only handle a ptime of 60 ms). We are using mod_com_g729's codec. The audio while on the call in normally fine (rarely there is one way speech depending on inbound/outbound provider), but the recording has a garbled outbound (Callee) leg (inbound/Caller voice is fine.).
We have tried all the combinations of G729@20i, G729@60i, rtp-autofix-timing, and default-ptimes to no avail.
We are using git HEAD.
I do see this on fs_cli:
2012-06-07 16:56:45.543269 [INFO] mod_com_g729.c:78 ENCODER DESTROYX - 0xa8894898 (nil) 2012-06-07 16:56:45.543269 [INFO] mod_com_g729.c:79 DECODER DESTROYX - 0xa8894898 (nil) 2012-06-07 16:56:45.583270 [WARNING] mod_sofia.c:1158 Asynchronous PTIME not supported, changing our end from 20 to 60 2012-06-07 16:56:45.583270 [DEBUG] sofia_glue.c:2931 Changing Codec from G729@20ms@8000hz to G729@60ms@8000hz 2012-06-07 16:56:45.623271 [INFO] mod_com_g729.c:78 ENCODER DESTROYX - 0xa86fe340 0xa86f1e78
Should this be working, are we missing something obvious?
Thank you,
Kind regards,
Colin