How to process audio in real time?

Question

I have some audio input through microphone. I recorded it in Audacity, it looks something like as shown below.

I want to process this audio in real time. I mainly want to do this:

1. See real time audio amplitude vs time graph.

2. Perform some actions based on event like if a wave pattern is seen similar to some predefined wave pattern, then call a function.
   In the upper half part of the image, you can see four disturbances. I want that whenever this disturbance is seen, the script should call a function, otherwise, when the signal is almost straight, do nothing just keep analyzing signal.

Is there any Python module or C library that can allow me to do this?

Answer 1

I'd Suggest to get yourself introduced to FFT first, that will make you able to do lot of analysis and control over the processing on your audio at realtime.

1# A forward FFT will give you amplitude within the time-domain, converted from your frequency domain (the audio data itself) 2# Based on the amplitude patterns within the time / frequency - you can decide to call extra thread or function; for example - when you get no amplitude on certain frequency bin or timeslot - do something, skip otherwise.

There are a plenty of opensource C libraries to perform FFT on audio data, like http://aubio.org/ or https://github.com/zaphire/Monocle-Engine/blob/master/Core/Audio/fft.cpp there are a lot more. But be sure to make you understand FFT and sound analysis before you proceed. I've personally not used any of these C libraries - instead did a lot with iOS audio analysis, but the working principals should be similar