Hi everyone,

 I'm new to audio processing and am trying to extract the MFCC of WAV files. I'm using C# and would appreciate any guidance. I understand most of the theory behind it, but there are some aspects which are still confusing me when it comes to actually coding. I hope someone could clarify things for me :D.

I read a WAV file as an input and it's made up of these (for e.g.):
414 202 423 0 0 231

Let's say this represented one second.
What would constitute a frame of 10ms (with overlapping)?
If I was to perform a Hamming Window on each "frame", would that mean I take 414 (index 1) and insert that into the formula? Then do that with every number in that particular frame?

Thanks for any help.

--- (Edited on 7/29/2009 1:01 pm [GMT-0500] by Visitor) ---

Sorry, to clarify the above, by formula, I am referring to the Hamming Window formula :

http://en.wikipedia.org/wiki/Window_function#Hamming_window

Thank

--- (Edited on 7/29/2009 1:03 pm [GMT-0500] by Visitor) ---

There are many Java MFCC implementations. This one is simple

http://www.cp.jku.at/people/schedl/Research/Development/CoMIRVA/webpage/CoMIRVA.html

Sphinx4 also has one. I suggest you to look on any of them, porting from C# to Java shouldn't be complicated.

--- (Edited on 7/29/2009 4:47 pm [GMT-0500] by nsh) ---

Thank you so much for the link! Hopefully I can understand it a bit more with the code and continue with my project! :)

--- (Edited on 7/29/2009 7:57 pm [GMT-0500] by Visitor) ---