I am writing an application using FMOD to detect voice input from a microphone. I am not doing any processing on the voice audio, but simply specifying the binary state of ‘talking’ or ‘not talking’.

I have based the design on the Pitch Detection example to determine if the input signal from the microphone is within human vocal range. However, to make this more robust, I also want to combine it with some volume threshold detection so that any background sound does not trigger the ‘talking’ state.

Can this be achieved with DSPs? I have experimented with FMOD_DSP_LOWPASS_RESONANCE and FMOD_DSP_NORMALIZE_THRESHHOLD, but they doesn’t seem to have any influence on the detected audio signal 😮 Maybe I am using them wrongly. I am adding them to the system object (I only have one channel), setting the parameters accordingly, then activating them.

Any suggestions for achieving this ‘simple’ volume threshold detector very welcome.

  • You must to post comments
Showing 0 results
Your Answer

Please first to submit.