Monitoring Sound Outputs

Hi,

I'm not sure if this is possible...

I want to write a program that will 'listen' to any sounds generated by the system, and generate a waveform graph or some other visual representation of the sounds being played.

Why? - I want to automate some testing activities that I have here. At the moment its a manual job of loading in sound files, playing them, and checking that they sound right. I was thinking of capturing these sounds as waveforms (or whatever) and just comparing them to a benchmark waveform.

Any ideas if this is feasible ?

Cheers dudes...