In the context of FIR filters, what are the tradeoffs between using a low pass filter designed using a windowed sinc versus an optimally iterative algorithm such as Parks-McClellan? When would you use one over the other?
$\begingroup$
$\endgroup$
3
-
$\begingroup$ Not completely, as it doesn't address tradeoffs. The tradeoffs can be ease of implementation, passband/stopband properties, etc. If they are equivalent then that would also be good to know. $\endgroup$– BigBrownBear00May 21, 2022 at 17:14
-
$\begingroup$ I thought it did in the answers of when you would use one vs the other. You can achieve the same stop band and passband properties with either depending on the window chosen: the Kaiser window comes close to the the least squares which is optimum in the least squares sense. So least squares is best but windowing comes close and useful when you don’t have a lot of resources to compute coefficients (dynamic filters that change on the fly) or when least squares does not converge. But this is all explained in that post I believe? $\endgroup$– Dan BoschenMay 21, 2022 at 17:33
-
$\begingroup$ (And Parks-McClellan when we want optimum in a min max sense vs least squares- Least squares is my typical go-to filter except under conditions I explained above) $\endgroup$– Dan BoschenMay 21, 2022 at 17:36
Add a comment
|