Greg Berchin's filter design method

Folks,

a bit over a year ago, in message 
<q99bu0lqffhjhd40vc1n0689nfcidv7s12@4ax.com>, Greg Berchin outlined a 
filter design method he had invented and there was talk of writing an 
article about it. Has that come to pass in the meantime?


Martin

-- 
Quidquid latine scriptum sit, altum viditur.

Rick Lyons and I were exchanging drafts of the article, and then he had
to put things on hold for more pressing matters.  I have not heard from
him in quite some time.  (I am astonished that a year has passed.  Time
flies.)

Greg Berchin

Greg Berchin wrote:

> Rick Lyons and I were exchanging drafts of the article, and then
> he had to put things on hold for more pressing matters.  I have
> not heard from him in quite some time.  (I am astonished that a
> year has passed.  Time flies.)

That's a pity. Would you be comfortable with sending me a draft? 
Otherwise, do you have any quick hints on using what you wrote back 
then? (Your message is at 
http://groups.google.com/group/comp.dsp/msg/2c307e8e77fb5178 .)


Martin

-- 
Quidquid latine scriptum sit, altum viditur.

On Wed, 26 Apr 2006 21:21:49 +0000 (UTC), Martin Eisenberg
<martin.eisenberg@udo.edu> wrote:

>That's a pity. Would you be comfortable with sending me a draft? 

No, but in the next few days I'll try to put together some Matlab code
that implements the algorithm.  Then perhaps I can make it available to
interested parties on a "not for commercial use" basis.

>Otherwise, do you have any quick hints on using what you wrote back 
>then?

Yes; two of them.

1. Artificially adding a few samples of delay to the input data almost
always improves the least-squares fit, especially at high frequencies.
(If you use a pure cosine input and your frequency response tends toward
an odd multiple of 90&#4294967295; as you approach half the sampling frequency, the
output samples occur near the zero crossings and you have an
observability problem -- a "finite-input-zero-output" situation.  Delay
minimizes the problem.)

2. Modifying the algorithm to perform a "weighted least squares" fit is
trivially easy.  I have found that the fit can often be improved by
weighting the data by "1/magnitude" (lower magnitude => higher
weighting).  Furthermore, for input data that are linearly spaced in
frequency, weighting low frequency data higher than high frequency data
is often beneficial.

Greg Berchin

Thanks, Greg!


Martin

-- 
Quidquid latine scriptum sit, altum viditur.

On Mon, 24 Apr 2006 20:49:58 -0500, Greg Berchin
<76145.2455@compuswerve.com> wrote:

>Rick Lyons and I were exchanging drafts of the article, and then he had
>to put things on hold for more pressing matters.  I have not heard from
>him in quite some time.  (I am astonished that a year has passed.  Time
>flies.)
>
>Greg Berchin


Hi Greg,

   Sweet merciful Jesus in Heaven!!

It's been so long!  I've been waitin' to hear from 
you and you've been waiting to hear back from me.
All of a sudden it's nine months later!

Somehow we (perhaps I) have screwed up our 
communications.   I took a quick look at 
my old E-mails and I see that you and I exchanged 
10-15 E-mails last year.  

Darn, sometimes I put things "on the back burner" and 
then I don't realize they've falling off the stove top 
and out of sight completely.

Greg, I'll get in touch with you.

[-Rick-]

I have Matlab code for the algorithm.  If interested in a copy, send me
a private email message -- remove the "w" from my email address. 

Private, non-commercial use only.  

No guarantees, warranties, promises, assertions, hopes, dreams, or
delusions of any kind.  Please inform me of any errors that are found.

Greg Berchin

Greg Berchin <76145.2455@compuswerve.com> wrote in 
news:k5q452ti91k19iksmnjtuq93ujgr20ru82@4ax.com:

> I have Matlab code for the algorithm.  If interested in a copy, send me
> a private email message -- remove the "w" from my email address. 
> 
> Private, non-commercial use only.  
> 
> No guarantees, warranties, promises, assertions, hopes, dreams, or
> delusions of any kind.  Please inform me of any errors that are found.
> 
> Greg Berchin
> 

Hi Greg,

What do you specifically mean by non commercial use? Do you mean don't use 
my method in a filter design product or do you mean don't use the results 
of my method (like filter coefficients) in a commercial product?


-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

On Fri, 28 Apr 2006 20:27:44 GMT, Al Clark <dsp@danvillesignal.com>
wrote:

>What do you specifically mean by non commercial use? Do you mean don't use 
>my method in a filter design product or do you mean don't use the results 
>of my method (like filter coefficients) in a commercial product?

I mean the algorithm was first published in 1985, so it's in the public
domain.  If you write your own software that implements the algorithm,
then it's yours to do with as you please.  But don't use software that
*I* wrote to make *you* money.

Greg

"Greg Berchin" <76145.2455@compuswerve.com> wrote in message
news:390552t5dnsvd3lo4g9et0nsgesd5qh5r3@4ax.com...
> On Fri, 28 Apr 2006 20:27:44 GMT, Al Clark <dsp@danvillesignal.com>
> wrote:
>
> >What do you specifically mean by non commercial use? Do you mean don't
use
> >my method in a filter design product or do you mean don't use the results
> >of my method (like filter coefficients) in a commercial product?
>
> I mean the algorithm was first published in 1985, so it's in the public
> domain.  If you write your own software that implements the algorithm,
> then it's yours to do with as you please.  But don't use software that
> *I* wrote to make *you* money.
>
> Greg

I dunno if this answers Al's question.
I can see that you wouldn't want your code to be used in a filter design
package (or similar) that Al might be selling for money.
I'm not sure if you'll be upset if Al (or anyone) uses your software to help
in creating filters that in turn are used in my application, which sells for
money.

Not that I plan to use your software, but I'm still curious.

Cheers
Bhaskar