Hi All,

I want to add to one an application a feature that enables the comparison of two *similar* audio files and produce a *similarity* metric.

The two files could be different recordings of the same performance or even different "takes" of the same riff.

Of course the files will be binary different and might need to be time-aligned and normalized before any comparison takes place.

I am planning to break down the audio files in discrete segments and then analyze frequency spectrum, energy levels, "level envelopes", etc...

I have already searched this and others resources but found no concrete pointers, so my question is:

Is there any code, algorithm or paper that can get me started with this?

Thanks in advance

www.voind.com

Google on acoustic fingerprint

Thanks. That lead me to some interesting code that I need to explore