Constant Q and Variable Q Algorithms for Frequency Analysis

Feedback and Discussion Forum
1

Hello,

I have noticed that the resolution for lower frequencies is much lower than that of higher frequencies due to the use of the ordinary fast Fourier transform. However, this does not correspond to human audio perception, so I would like to request adding the constant Q and variable Q transforms as options in the spectrogram settings, and I would also like to go one step further in requesting these be the default option, since most people use the spectrogram function to analyze sound; the default would be chosen based on the scale used; i.e., the linear scale would continue using the FFT, the logarithmic scale would use the constant Q transform, and the Mel, Bark, and ERB scales would use the appropriate variable Q transform for each of them. This is similar to how the period scale is designed to work with the EAC algorithm, except my proposal is to automatically change the algorithm when the scale is changed, which in my opinion would increase accessibility.

Please give this some consideration.

Thank you,
Anhilare

2

If you increase the window size there is better resolution at low frequencies

https://manual.audacityteam.org/man/plot_spectrum.html

There are free spectrum plugins where you can tilt the slope of the plot to ~match human hearing

TDR Prism | Tokyo Dawn Records … Voxengo SPAN - Free Spectrum Analyzer Plugin VST, AU, AAX

3

Hello,

Thanks for your reponse, but I’m not talking about the plot spectrum; rather, I’m talking about the spectrogram view that’s accessed by right clicking the track and selecting “spectrogram” and already has different scales built in. I couldn’t find a plug-in that does as I want, and I honestly feel that this feature should be in base Audacity anyway (and it already exists in a partial form with the implementation of the EAC algorithm).

Sorry for being unclear in the title. I should have specified that I was referring to the spectrogram and not the plot spectrum.