Hi,
Is it possible for Asterisk to record (using Monitor or MixMonitor) in a higher sample rate than whatever the codec supports?
For example, the alaw supports an 8000 Hz sample rate, is there a possibility that the audio can be recorded with a sample rate of 48000 Hz when the RTP transmission is happening in the alaw format.
It can do so, but it is likely to be a waste of CPU time and disk space. It won’t extend the frequency response. Why would you want to do it.
In your example, just use a file extension of slin48 (if I’ve remembered correctly). The mixing will, I suspect, be done at 8kHz.
The reason for doing so is that we want to integrate a third-party speech analytics service which only supports the 48000 Hz or higher sample rate recording files.
The reason it only supports them might well be that the loss of sibilants with 3.1kHz audio makes the speech recognition difficult. On the other hand 16kHz sampling is perfectly adequate for speech. 48kHz is total over-kill. You should check with them whether they really support telephone quality speech.
I wonder if they are doing something clever to try to identify different speakers. Obviously that won’t work as the audio has no more information that the original 8kHz sampling.
Sure, I’ll check with the vendor about the 16KHz sampling.
Thanks, the explanation helps a lot.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.