Classification algorithm

Jan 31, 2011 at 8:57pm

Classification algorithm

Hello everyone,

In a sound bank, would it be possible to compose an algorithm that classify sounds depending on frequencies they are made of ?

Fictive example: i would like all the sounds that contains 10 000 Hz or higher freqs to the “left” and all the sounds that contains 30 Hz or lower freqs to the “right”.

If anyone has an idea, how could i do that ?

Thanks a lot.


Feb 1, 2011 at 10:19am

My first thought is a fft~ analysis of the sound. In this analysis you could extract which frequency is most common. Next thing would be labeling the different frequency groups.

Feb 11, 2011 at 11:19pm

Thank you, indeed that is a possibility.

Feb 12, 2011 at 12:03am

Once you’re there the ZSA Descriptors are definitely worth having a look through.


You must be logged in to reply to this topic.