[Asterisk-Users] Detecting DTMF tones

[Richard Scobie]

matt.riddell at sineapps.com wrote:
> On 15 Sep 2004 at 1:52, San Singhania wrote:
> 
> 
>>Hello everyone,
>>
>>I am having big problems trying to detect dtmf tones while a IVR
>>prompt is playing on zap channels. Sometimes the detection only starts
>>4-5 seconds into the prompts. Other times it works very well for the
>>1st few calls and then starts having problems. And most times it also
>>does not detect all digits, eg when 123 is keyed in, it may detect 23
>>or 2 or 3, but never the complete string. 
>>
>>Can someone help me with this? I am in Singapore so I dont know if its
>>a localisation issue. Also I have relexdtmf on and my rxgain and
>>txgain on zap channels are set to 0 so i dont think thats the problem.
>>
>>Help is most appreciated as I cannot go on with Asterisk if this
>>function does not work :(
>>
> 
> Ok, I'm based here in New Zealand and found that the volumes for PSTN 
> were quite a way off.

I do not think this is a localisation issue - Mark has measured this 
(see Bug 2023).

When I replaced X100Ps with TDM FXOs, busydetect stopped working until 
until the RX gain was increased 3dB. Which of the two was correct or 
maybe neither of them?

Given that gain distribution in analog telephony is quite critical from 
a stability/echo performance point of view, it would be great if the 
cards came out of the box correctly  calibrated for spec sheet levels.

There are enough real world variables to deal with, without wondering if 
the TDM FXO TX gain is correct and the FXS RX gain is low or vice versa.

I do realise that test sets, CO and asterisk milliwatt generators are 
needed to achieve a close to perfect result, but looping an FXS into an 
FXO should show unity before starting out.

Regards,

Richard