Timeline for Is there a noise cancellation technique for over-lapping spectra?
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 9, 2019 at 12:45 | comment | added | Fat32 | Ok edited now... | |
| Jul 9, 2019 at 12:45 | history | edited | Fat32 | CC BY-SA 4.0 |
added 213 characters in body
|
| Jul 9, 2019 at 2:07 | comment | added | Creator | I cannot undo the voting, sorry as time has passed. I can do that only if your edit the question. In any case, Fourier transfor (FT) loose time information, as it a convolution with a specific function. So people do short time FT but loosing time informaiton is inherent. | |
| Jul 9, 2019 at 1:17 | comment | added | Fat32 | No I am talking about clicks and pops ?? If it's a click and pop (LOCAL) noise, then indicate this clearly in your question. Overlapping spectra is not sufficient to describe click & pop noise. White Noise will also have overlapping spectra. Wiener filter has no good use in click and pop noise. You should use nonlinear time domain filters. So go and edit your question for clearly indicating your noise type. And BTW the Wiener filter will reduce in band noise too. There's nothing wrong in the answer, correct your mistake and undo downvoting, and EDIT your question. | |
| Jul 9, 2019 at 1:11 | comment | added | Creator | No I am talking about clicks and pops. Again expected value means non-local or global measure, hence you loose local information. This means you would remove signal portion of in-band noise as well if it removes in-band pectra. What I am saying is that the filter cannot remove the requied portion of in-band noise without affecting the signal portion.Can you show mw some example of reducing in-band noise without affecting the spectra of the signal? | |
| Jul 9, 2019 at 0:04 | comment | added | Fat32 | Are you talking about clicks and pops that are local in time and wide in frequency? Then a Wiener filter is not the correct tool to remove them of course. | |
| Jul 9, 2019 at 0:00 | comment | added | Fat32 | In band means, having the same frequency range, and as the formula of the Filter frequency response shows, it tries to reduce (not remove) noise in every frequency band including in & out bands. I don't know, where you get the idea of time information vs convolution, but it has nothing to do with how a Wiener filter works. Every LTI filter is equivalent to a convolution sum and this has nothing to do with their capability of noise reduction. Wiener filter works on a statistical basis. A classical filter to remove outof band noise is quite a simple idea, Wiener filter is something different... | |
| Jul 8, 2019 at 23:26 | vote | accept | Creator | ||
| Jul 8, 2019 at 23:26 | |||||
| Jul 8, 2019 at 23:26 | comment | added | Creator | Any convlution is a summation mean time information is lost.. How do you expect to remove in band signal with convolution when you loose all time information? | |
| Jul 8, 2019 at 22:59 | history | answered | Fat32 | CC BY-SA 4.0 |