endolith

The problem is likely in these hardcoded lines: ```py lowermin = int(true_i * 0.9) uppermin = int(true_i * 1.1) ``` They are not eliminating the entire fundamental when used with...

[REW's method](https://www.roomeqwizard.com/help/help_en-GB/html/spectrum.html#rtadistortion), for reference: > If the Use AES17-2015 standard notch option is selected in the [Distortion settings](https://www.roomeqwizard.com/help/help_en-GB/html/spectrum.html#distortionsettings) the fundamental figure will be the power within a one octave span...

This is your actual spectrum, which doesn't have very clearly defined harmonics (green) and has noise at higher frequencies with a peak around -48 dB:  [-48 dB should be...

> What's the drawback to changing those two lines? There isn't one; they shouldn't be hardcoded like that in the first place. It should ideally use a width that eliminates...

My plots are in dB: ```py … freqs = rfftfreq(len(windowed)*5, 1/fs) spec = 20*log10(abs(f)) plt.figure() plt.semilogx(freqs, spec) plt.axvline(frequency, color='red') for h in range(2, 10): plt.axvline(frequency * h, color='green') peak =...

The THD function should be correct now, too, though note that THD uses THDF while THD+N uses THDR. In the future I would like both functions to support both as...

Looks like https://github.com/mahyar-amiri/keras-visualizer is a more recent fork

https://github.com/ashishpatel26/Tools-to-Design-or-Visualize-Architecture-of-Neural-Network/pull/13 also does this

Or just cache the FFT function.

@PoorvaTi This would replace the readme with a python file, which is most likely not what you intended.