Prev: Porch
Next: IFFT of custom frequency array - confusion with the time domainsampling frequency
From: Andor on 5 May 2008 15:32
Greg Berchin wrote:
> Andor,
>
> Can you give an assessment of the typical performance difference
> between least squares FDLS and minimax FDLS (which, I guess, should be
> renamed to FDMM in this context)?  Or is it just too dependent upon
> the particulars of any given problem?

Given the nature of the numerics, I would use FDLS to design filters
from noisy frequency domain measurements (for example to fit a
periodogram or the like) and FDMM to design filters from analytic
(smooth) frequency domain prototype. LS allows for wiggles that won't
occur in MM (assuming that the frequency grid is dense enough).

Having said that, I have never actually used FDMM to design any
filters (I used FDLS in my negative group delay blog), so YMMV.

Regards,
Andor
From: Rune Allnor on 5 May 2008 15:36
On 5 Mai, 20:04, DigitalSignal <digitalsignal...(a)yahoo.com> wrote:
> An IIR filter invented in early '70s? Are you kidding?

Well, Armstrong did his moon walk in the late '60s...

I found the reference Clay Turner directed me to a few
years ago:

[1] K. Steiglitz, Computer Aided Design of Recursive Digital Filters,
IEEE
Trans. Audio Electroacoust., vol. AU-18 June 1970, pp. 123-129.

If you have a general optimization framwork that paper
contains everything you need to implement your own
design program.

Rune
From: Fred Marshall on 5 May 2008 20:42

"DigitalSignal" <digitalsignal999(a)yahoo.com> wrote in message
news:bc68d4ff-b0ef-4d34-8954-b5e30d308faa(a)k1g2000prb.googlegroups.com...
> An IIR filter invented in early '70s? Are you kidding?
>
> James
> www.go-ci.com

A huge part of the important work was done in the 60s and 70s if not
before....

One can go back in history to find discrete sequence processing algorithms
back 100s of years. Thus, "DSP".

Fred


First  |  Prev  | 
Pages: 1 2
Prev: Porch
Next: IFFT of custom frequency array - confusion with the time domainsampling frequency