I was playing with FMOD in windows mobile 2003 and all seems to works pretty well.

Now I’m writing an application in eVC++ 4.0 , basically is a non-graphical application, all that it does is record 2 minutes of audio from the mic and process it, no playback required and at the moment the recording is working ok.

Now I have to insert my processing, basically I have to send 1 sec of raw audio (16000 samples at 16 Khz/Mono or 22050 at 22Khz/mono) to a function that I wrote while recording, without stoping the recording process. All that my function does is some fixed point FFT and save the results in a text file.

I was checking the forum for some ideas of how to do it, but I cannot find the rigth approach.

The first one that I notice is a loop recording (similar to the record app in Win32) and then by locking/unlocking the buffer and getting the posstion copy that portion of buffer and send it to my function. I have the impresion that my processing must be done in a different thread, otherwise the recording process will stop during my process, is that correct?
And if is like that, there is any chance to put for instance a callback function that when I have 1 second will be called without interrumping the recording?

Any suggestions of how to do that or samples?

The other approach that I read is using a DSP callback, but for that as far that I noticed, the DSP must be associated to a play function and I’m not playing audio, just recording, I dont need to playback.

Please any suggestion, idea or comment is welcome.


  • You must to post comments
Showing 0 results
Your Answer

Please first to submit.