Elementary FIR filters

The descriptions I've seen seen of digital filters seem to jump from 
simple averaging on last n samples to specifying filter as ratio of 
polynomials.

I have Rick's book somewhere but can't find it at moment.

I'm also trying to reconstruct bookmark file.
Can you say "system crash w/o backups" ;{

Richard Owlett <rowlett@atlascomm.net> writes:

> Can you say "system crash w/o backups" ;{

"System crash w/o backups" --- Never! Provided I also say "250Gb USB drive"?

:-(

Ciao,

Peter K.

"Peter K." <p.kootsookos@remove.ieee.org> wrote in message 
news:uu0dp94fr.fsf@remove.ieee.org...
> Richard Owlett <rowlett@atlascomm.net> writes:
>
>> Can you say "system crash w/o backups" ;{
>
> "System crash w/o backups" --- Never! Provided I also say "250Gb USB drive"?

I think external USB (or 1394) drives are great for backup.  They are cheap 
($80-150), fast, easy-to-use, and portable enough to be easily taken "off site". 
You don't even need to use any fancy back-up software if you don't want--copying 
with Windows Explorer works fine.  I keep my a USB drive with my home backup at 
work, bringing home occasionally to refresh the content.

If you have a fast Internet connection and a reasonable amount of data to back 
up (say a few gigs, rather than hundreds), www.streamload.com works pretty well 
as a free on-line backup site. 

Jon Harris wrote:

> I think external USB (or 1394) drives are great for backup.  They are cheap
> ($80-150), fast, easy-to-use, and portable enough to be easily taken "off site".
> You don't even need to use any fancy back-up software if you don't want--copying
> with Windows Explorer works fine.  I keep my a USB drive with my home backup at
> work, bringing home occasionally to refresh the content.

Yup!  When I moved from Oz to the US a while back, I needed space to
transfer all my data... rather than write 20+ CDs / DVDs like I had the
last time.  So I bought a 250Gb USB drive.

I only filled up just over half, so now I also use if for backups.

> If you have a fast Internet connection and a reasonable amount of data to back
> up (say a few gigs, rather than hundreds), www.streamload.com works pretty well
> as a free on-line backup site.

Another possibility is to get a Gmail account and use their 2Gb limit.
:-)

Ciao,

Peter K.

This group has *WANDERLUST* and has completely verified my own advice 
about backups ;}

Are there any examples/visual examples of digital FIR filters between a 
simple averager of last n samples and coefficients magically generated 
by a ratio of polynomials.

Richard Owlett wrote:
> This group has *WANDERLUST* and has completely verified my own advice 
> about backups ;}
> 
> Are there any examples/visual examples of digital FIR filters between a 
> simple averager of last n samples and coefficients magically generated 
> by a ratio of polynomials.

Generally, filters that involve the ratio of two polynomials are IIRs. 
Repeatedly convolving two-point averager with itself yields a a binomial 
(or Gaussian or Pascal's triangle) filter which has better stop-band 
performance than averaging, but not as good as windowed sinc, which is 
nearly as good as iteratively optimized (Parks-McClellan and all that) 
FIRs.

I'm particularly partial to binomial filters because I invented them all 
on my own to clean up the display on my Nicolet scope. They worked well 
enough, even though I now know of better. Haven't you been reading 
Lyons? Check out http://www.dspguide.com/filters.htm.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

It's interesting that one of the trends that are expected in
technology use is to maintain disconnected data.   The idea is that a
person could have a small, but high capacity, hard drive that they
carry with them that contains their data.   That data is then
available (wirelessly) to any device that they may be using, such as
their cell phone, their laptop, PDA, MP3 player, etc., etc.

The usage models are kinda cool if you think about how people might
exploit such a thing.  Security issues are important, obviously, but
it looks like that's a solvable problem to a usable extent.

One ought to be able to configure such a thing that whenever it's in
proximity to another device on which you want to mirror the data it
automatically synchronizes.   I do this now manually between my laptop
and my desktop, but I can easily see it happening automatically such
that your data winds up being mirrored on several "trusted" or
"stable" devices so that backup diversity is almost a no-brainer.

Until then, though, we get to do stuff like this.  ;)


On 4 Dec 2005 06:53:26 -0800, "Peter K." <p.kootsookos@iolfree.ie>
wrote:

>Jon Harris wrote:
>
>> I think external USB (or 1394) drives are great for backup.  They are cheap
>> ($80-150), fast, easy-to-use, and portable enough to be easily taken "off site".
>> You don't even need to use any fancy back-up software if you don't want--copying
>> with Windows Explorer works fine.  I keep my a USB drive with my home backup at
>> work, bringing home occasionally to refresh the content.
>
>Yup!  When I moved from Oz to the US a while back, I needed space to
>transfer all my data... rather than write 20+ CDs / DVDs like I had the
>last time.  So I bought a 250Gb USB drive.
>
>I only filled up just over half, so now I also use if for backups.
>
>> If you have a fast Internet connection and a reasonable amount of data to back
>> up (say a few gigs, rather than hundreds), www.streamload.com works pretty well
>> as a free on-line backup site.
>
>Another possibility is to get a Gmail account and use their 2Gb limit.
>:-)
>
>Ciao,
>
>Peter K.
Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Jerry Avins wrote:
> Richard Owlett wrote:
> 
>> This group has *WANDERLUST* and has completely verified my own advice 
>> about backups ;}
>>
>> Are there any examples/visual examples of digital FIR filters between 
>> a simple averager of last n samples and coefficients magically 
>> generated by a ratio of polynomials.
> 
> 
> Generally, filters that involve the ratio of two polynomials are IIRs. 

Ooops, no wonder things weren't making sense.

> Repeatedly convolving two-point averager with itself yields a a binomial 
> (or Gaussian or Pascal's triangle) filter which has better stop-band 
> performance than averaging, but not as good as windowed sinc, which is 
> nearly as good as iteratively optimized (Parks-McClellan and all that) 
> FIRs.
> 
> I'm particularly partial to binomial filters because I invented them all 
> on my own to clean up the display on my Nicolet scope. They worked well 
> enough, even though I now know of better.

 > Haven't you been reading Lyons?

Student "hems" and "haws". Then admits he can't find his copy.
But I think you had already hit on my key problem -- not realizing ratio 
of polynomials were primarily IIR filters.

I think I've been reading "carelessly".
I think I've been associating passive LTI R-L-C filters with FIR -- is 
there a fundamental error there?

I was associating filters with "feed forward" and "feed back" terms with 
active networks ( ie containing op amps etc ).

 > Check out http://www.dspguide.com/filters.htm.

Took quick look. Think left side of Table 14-1 gives me useful hints.


> 
> Jerry

Richard Owlett wrote:

   ...

> I think I've been associating passive LTI R-L-C filters with FIR -- is 
> there a fundamental error there?

Passing over what is fundamental, it's an error. Passive R-L-C filters 
and their active emulations are generally minimum phase (lattice filters 
excepted), and most closely approximated by IIRs, especially well below 
half the sampling frequency. FIRs can approach minimum phase, but not 
the symmetrical ones usually designed.

> I was associating filters with "feed forward" and "feed back" terms with 
> active networks ( ie containing op amps etc ).

Active, shmactive. It's what they do that counts, not how they're built.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Richard Owlett <rowlett@atlascomm.net> writes:

> This group has *WANDERLUST* and has completely verified my own advice
> about backups ;}

I resemble that remark.

Ciao,

Peter K.

PS: :-)