FFT Bins, averaging/normalizing

Best,

I’m trying to divide the data from a CHOP spectrum into several bins (or channels).
Willing so to get a magnitude for each part of the audiospectrum (sub, bass, low-mid, mid etc.)
I want to use these values to drive certain parameters so the values have to be normalized or something.
Any help or suggestions that will bring me closer to this goal will be much appriciated.

Thanks in advance,

Bart

Hi bart,

For instance you can connect a few Audio Filter CHOPs to filter out each bands of audio your want (sub, bass, mid, high), and use the output of each of those Audio Filter CHOPs to drive your visualization.

If you open Help->Operator Snippets, you’ll find some examples how to work with these under CHOP → Audio Filter

Also in those snippets see the Audio Spectrum CHOP examples, there’s an example which shows approximate average for 10 bands in the audiospectrum, you can easily adjust that to the number of bands you need.

To normalize soundlevels, as in use a Limiter & Compressor, see the Audio Dynamics CHOP examples (also in the Operator Snippets collection).

You can use the Trim CHOP to select parts of your magnitude values and you can use tha analyze CHOP to get the power in the selected band.
As for normalization, yes, as nettoyaur said, some compression before the FFT and maybe some limiting and rescaling afterwards should more or less do it.