Right now the number of samples (441) is based on the small SDL3 audio buffer size on Windows.
We should accumulate more samples (at least twice as much) to run the FFT on to get a larger power spectrum density array, yielding more fine grain results. Right now 441 samples gives us just 257 values, which means each value represents ~85 hertz (22050 / 257).
This should help getting better/more accurate frequencies and frequencies band graphs.
Right now the number of samples (441) is based on the small SDL3 audio buffer size on Windows.
We should accumulate more samples (at least twice as much) to run the FFT on to get a larger power spectrum density array, yielding more fine grain results. Right now 441 samples gives us just 257 values, which means each value represents ~85 hertz (22050 / 257).
This should help getting better/more accurate frequencies and frequencies band graphs.