Re: Multiple pass moving average filter

Clearly your generally statement is not applicable to all
data sets and applications as it appears to apply. Otherwise
there would not be experts in the dsp field that state that
moving averages are optimal in most cases (it a given that
other filters can be just as good). Since I'm not a dsp
person I'm out of my league to even attemp to debate the
issue or whether multi pass being "a variant" of moving
averages -- has similar properties.

I do know the data in my domain exceptionally well. The
same applies to the application side. So 95% of the time I
never need or use any smoothing in model design. The
performance benefits have to be huge for me to consider
dealing with the type of lags that smoothing filters introduce.
I can verify the performance benefits are huge with probability
based data and "do not expect" any significant performance
gains from using a custom designed filter.

My goal at this point is merely to gain a better understanding
of the equivalent filters that are being created with custom
modification to the ends. While I know various formulas as
outlined in this discussion on treating the ends are commonly
used and considered acceptable by dsp professional -- I'm
looking to understand the mathematical results the weighting
strategies are equivalent to. If at some point in the future I
decide to explore the possibility of an improved solution I would
have more useful information to pursue that goal. Right now I'm
not looking to chase incremental gains that might or might not
be possible using other filters.

Multi-pass moving average filters are a conceptually simple but
computationally inefficient way to implement a low-pass filter. They are
a highly constrained subset of FIR filters which sacrifices performance
for intuitive presentation.

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

.

Relevant Pages

  • DSP ideas wanted AND opinions on DSP software
    ... autonotch and selectable preset filters would be useful, ... My DSP processor adjunct to my homebrew receiver is documented at ... Using the very narrow filter described on the site, ... radio work; I hope to have the remote interface running within a month or so ...
    (rec.radio.amateur.homebrew)
  • Re: Top 10 contributers to field of DSP
    ... on dsp or contributers (some lighter moments ... DSP textbook to discuss digital filters and ... is also the inventor of fuzzy-logic.) ... between black socks and dark blue socks. ...
    (comp.dsp)
  • Re: DSP
    ... Jay in the Mojave wrote: ... The DSP does reduce some noise and such. ... But the Digital IF Filters ... you to only receive one sideband if the station you wanted to understand ...
    (rec.radio.cb)
  • Re: Book alert: Emphasis on digital filter design.
    ... Signals, Systems and Filters. ... > How on earth did he manage to cover DSP in just 950 pages? ... the "standard filters" standard reference, ... and a text dedicated to filters and filter design. ...
    (comp.dsp)
  • Re: How reliable is wireless over wired?
    ... part of the SYSTEM after the filters. ... The attenuating device is located ahead of the filters, ... overload the transmitter is done in the DSP and that is well after the ... So in answer to your post, the limiter works on all frequencies but is ...
    (rec.arts.movies.production.sound)
Loading