Timeline for FFT basic concepts
Current License: CC BY-SA 4.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 15, 2023 at 12:01 | comment | added | Neil_UK | @KaleM 5) No. 6) No. You completely misunderstand. Look up Welch's Method on wikipedia. With the exception that instead of averaging all the periodograms as you would to estimate a stationary spectrum, you plot them against time to get the STFT plot, or you average a few neighbouring ones (rather than all of them) to improve the variance of the STFT plot at the expense of time resolution. | |
| Nov 15, 2023 at 11:52 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 41 characters in body
|
| Nov 15, 2023 at 10:15 | vote | accept | KaleM | ||
| Nov 15, 2023 at 10:15 | comment | added | KaleM | 6) What if I also want to calculate the magnitude of the FFT I just performed ... do I also calculate it on the first half of the buffer when it fills up and then on the second half when it fills up? Or should I calculate the magnitude on the total? (i.e., I should sum the two FFTs calculated on the two buffer halves and then find magnitude of the sum) | |
| Nov 15, 2023 at 10:15 | comment | added | KaleM | 5) So, in practice, as soon as the first half of the buffer of the ADC where I receive the data is filled ... I perform the FFT on that half. Same thing when the second half of the buffer is filled (i.e. I will do the FFT only on the second half of the buffer). Correct? Then should I sum the two results into a single array? | |
| Nov 12, 2023 at 14:17 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 23 characters in body
|
| Nov 12, 2023 at 14:07 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 23 characters in body
|
| Nov 12, 2023 at 12:39 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 931 characters in body
|
| Nov 12, 2023 at 12:33 | comment | added | Neil_UK | @KaleM too long for a comment, so I've added another para or two to my answer. | |
| Nov 12, 2023 at 12:33 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 931 characters in body
|
| Nov 12, 2023 at 11:25 | comment | added | KaleM | Thank you for the explanation! I'm asking you what I asked @Rohat. In point 4) I was referring to a hypothetical case where the signal varies its frequency nonlinearly. Imagine tying a resistive band around a patient's chest that varies in resistance depending on how much it is widened (how much he inhales/exhales). If we assume from a medical point of view that the frequency varies between [0.2Hz and 0.7Hz] ... it doesn't necessarily vary linearly within this range, because he could start running at any moment and then immediately stop and start again. How do you implement a good FFT here? | |
| Nov 9, 2023 at 14:41 | history | edited | Neil_UK | CC BY-SA 4.0 |
added 630 characters in body
|
| Nov 9, 2023 at 14:31 | history | answered | Neil_UK | CC BY-SA 4.0 |