I want to create an application that analyzes human voice pitch, but spectrograms are very noisy. However, in Adobe Audition, there is a feature called the spectral pitch display, and it successfully filtered the spectrogram so that only relevant pitch info is displayed. What is this type of graph called? And where should I start searching if I want to implement it in code?

Example Spectral Pitch Display:

enter image description here

Example Spectrogram of the same audio segment:

enter image description here


1 Answer 1


You can usually trade off time versus frequency resolution in a spectrogram by changing the length, offset, and overlap of the STFT/FFTs.

If you compute more than one set of STFT/FFTs for a single spectrogram (say, one short FFT set for better time resolution, and one longer FFT set for better frequency resolution) you can combine the two (or more!) sets into a single graph by using one set to gate the display of the other (via some non-linear function).

However, (perceived) pitch and spectral frequency are not the same thing.

Some methods I've tried for my music spectrogram app are to use a pitch estimator (one of autocorrelation/ASDF/AMDF, cepstrum, harmonic product spectrum, et.al.) instead of just a fixed length FFT frequency estimator, and graph the tracked pitch and its associated overtones with or on top of the STFT/FFT spectrogram.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.