I'm trying to understand how to specify the "desired signal" in adaptive LMS filters such as the following: This one, or this one page 7, or this one.

  • It seems like the example there uses it for noise suppression?
  • But why is the desired signal input + noise?
  • Isn't it what we want to get rid of, rather than desire?

I'm trying to use LMS for "arbitrary magnitude response" filtering (as suggested here).

  • $\begingroup$ I think you need to describe your application in more detail. From your previous questions it sounds like you want a filter with a (totally?) different magnitude response at each new incoming sample. This doesn't make much sense to me. Where do these quickly changing magnitude specs come from? $\endgroup$ – Matt L. Jul 2 '16 at 18:48
  • $\begingroup$ @MattL. Well that's close. I'm wanting to do an equalizer to which I compute magnitude responses from FFT analysis. So I can get data, but I don't know how to shape a filter to that data. And yes, it should vary even on a per sample basis (since you need to compute the envelopes adjusting from one FFT analysis frame to another). $\endgroup$ – mavavilj Jul 2 '16 at 18:50

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.