Stft output frequency bin count doesn't match Stft constructor parameters?

[Bambofy]

Hi!

I'm currently using the Stft feature, when i provide these parameters

Stft stft = new Stft( windowSize: 1024, hopSize: 256, window: NWaves.Windows.WindowTypes.Hann, fftSize: 256);

The output spectrums result in a length 513 not 256?

See the image below that displays the number of bins in a single spectrum of the Stft spectrogram

Additionally: Q1) What is the resulting Y Axis graph representing, is it power? Q2) The Y values output from the Stft are normalized, how are they normalized? e.g:

[ar1st0crat]

Hi!

Regarding the STFT parameters - take a look here at n_fft and win_length parameters.

In short, the FFT size can not be smaller than the size of analysis window (so NWaves sets FFT size to 1024 automatically in your case). I guess what you're trying to do is analyze signal in the 1024-samples window and then obtain FFT with lower frequency resolution (256 samples). So essentially, you need to compute 1024-point FFT and then sum values inside each group of 4 adjacent samples. Something like this:

for (int i = 1, k = 0; i < s.Length; i+=4, k++)
    low[k] = s[i] + s[i+1] + s[i+2] + s[i+3]

Q1: yes, Y represents power (spectrogram is the sequence of power spectra). Q2: each power spectrum is normalized by FFT size (by default). Unfortunately, I didn't add the boolean parameter for normalization in Spectrogram() method (Fft.PowerSpectrum() has this parameter, though). This parameter will be added in next version of the lib.

Regards, Tim

[Bambofy]

Hi!

Regarding the STFT parameters - take a look here at n_fft and win_length parameters.

In short, the FFT size can not be smaller than the size of analysis window (so NWaves sets FFT size to 1024 automatically in your case). I guess what you're trying to do is analyze signal in the 1024-samples window and then obtain FFT with lower frequency resolution (256 samples). So essentially, you need to compute 1024-point FFT and then sum values inside each group of 4 adjacent samples. Something like this:

for (int i = 1, k = 0; i < s.Length; i+=4, k++)
    low[k] = s[i] + s[i+1] + s[i+2] + s[i+3]

Q1: yes, Y represents power (spectrogram is the sequence of power spectra). Q2: each power spectrum is normalized by FFT size (by default). Unfortunately, I didn't add the boolean parameter for normalization in Spectrogram() method (Fft.PowerSpectrum() has this parameter, though). This parameter will be added in next version of the lib.

Regards, Tim

Thank you for the detailed reply!