For audio classification we are using the cepstral coefficients and their derivatives. Which I calculate using the Matlab libary voicebox.

I am looking for some literature for how to choose the frame length and overlap of frames. Most literature that I know of are considering speech recognition, but for this application are the audio signals not generated by humans and they are chancing slowly over time. I suppose the last part have an impact when choosing the frame length?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.