Hi list, hi Jolly,
I just tested the DTMF support on a real network and discovered that the timeout of 70ms for receiving the Start DTMF ACK is too short. The consequences are many repeated Start DTMF messages, never received ACKs and no tone is played. Using a timer value of 170ms in mnccms.c:759 fixes this for that network. Unfortunately I didn't found - or overlooked - a mandatory timeout value for Ta in TS 23.014 which the MSC has to meet. Is that undefined in the standard and where came the 70ms from?
Regards, Mad
mad wrote:
Hi list, hi Jolly,
I just tested the DTMF support on a real network and discovered that the timeout of 70ms for receiving the Start DTMF ACK is too short. The consequences are many repeated Start DTMF messages, never received ACKs and no tone is played. Using a timer value of 170ms in mnccms.c:759 fixes this for that network. Unfortunately I didn't found - or overlooked - a mandatory timeout value for Ta in TS 23.014 which the MSC has to meet. Is that undefined in the standard and where came the 70ms from?
Regards, Mad
hi mad,
the timeout is set after the reception of the DTMF ACK. this ensures that the time to play the tone is at least 70ms. then the tones is stooped, and after it has been acked, the timer is started again to wait for the next tone. then the next tone, if any, is sent.
see: http://en.wikipedia.org/wiki/Dual-tone_multi-frequency_signaling
can you send me a log of the process when you send the dtmf tones?
regards,
andreas
Hi Andreas!
the timeout is set after the reception of the DTMF ACK. this ensures that the time to play the tone is at least 70ms. then the tones is stooped, and after it has been acked, the timer is started again to wait for the next tone. then the next tone, if any, is sent.
see: http://en.wikipedia.org/wiki/Dual-tone_multi-frequency_signaling
can you send me a log of the process when you send the dtmf tones?
Ok, I misread the dtmf_statemachine, I understand the 70ms are the minimum length a DTMF tone should last according to convention. Actually I re-tested with this value and - of course - now it does work as expected. Unfortunately I don't have the layer23 log anymore, but my first test calls rather seem to had some issue with the MSC not confirming the Start DTMF frame. While changing the tone length value, the MSC might got responsive to that again and after testing, I draw a false conclusion. :)
Regards, Mad
baseband-devel@lists.osmocom.org