3 ms·
Meh. The author confuses DFT with STFT. They refer to the STFT as "DFT", but then introduce the usual IDFT as the inverse transform, which is not the inverse of
by meindnoch 3y ago
Meh. The author confuses DFT with STFT. They refer to the STFT as "DFT", but then introduce the usual IDFT as the inverse transform, which is not the inverse of the STFT (the inverse STFT can be computed with the overlap-add method).
- quatrefoil 3y agoI don't think your comment is accurate. The formulas appear to be DFT and IDFT, in line with: https://home.engineering.iastate.edu/~julied/classes/ee524/LectureNotes/l5.pdf https://home.engineering.iastate.edu/~julied/classes/ee524/L...
- meindnoch 3y agoDid you even read the article, or just looked at the LaTeX formulas? I'll give you a hint: Cmd+F "window". See any results? Because the DFT doesn't have a window function. The STFT does have a window function. Second hint: look at the image in the article (https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fed62bf6d-9b51-442f-8a39-a07648b7d080_990x875.jpeg https://substackcdn.com/image/fetch/w_1456,c_limit,f_webp,q_...). Do you think this shows the time domain and frequency domain representations (i.e. the DFT) of a signal? No, it does not. It shows the time domain representation, and the STFT of the signal, aka the spectrogram. Do you have any questions?