Re: Optimum length for a mean filter

Hi Jerry, mn and Dale,

Maybe I wasn't clear explaining my problem, but you have made the way to
it. Hehehe. Yes, all your guesses about what was my problem are right.

Dale: yes, there is the threshold issue but it is not important for me
now. You are also right in saying that you lose time resolution if you
increase the shaping time of the filter. What is very interesting from
what you said is that nonstationary of the noise can limit the shaping
time. Could you expand the idea? Is there are any paper explaining this?

Jerry: that's right, the SNR is maximum just on the pic. You can block the
noise but that actually requires to know when it starts and that's also the
purpose of the filter... hehe. The SNR increasing with a longer length of
the filter might come as you say from the idea that the step is not
actually a step, for instance an exponential. But in this case it is a
step, I actually implement a pole-zero compensator that converts the
exponentail into a step.

Nm: right! I am implementing also a matched filter which would provide the
best SNR. But this is still independent from the idea that by increasing
the length of the filter the SNR should decrease.


On Nov 15, 7:14 pm, Jerry Avins <j...@xxxxxxxx> wrote:
.> Jerry Avins wrote:
...
.> Whether in the end you
.> choose stay with an integrator or use a better filter, blocking the
.> filter's input until a step is detected will keep out much of the
noise.
.>
.> Jerry

While this statement is certainly true it is not of much use to the
step detector itself. You have obviously provided another example of
the classic detector optimized to minimize false alarm rate by being
turned off. This is probably no longer patentable as it was widely
known before the end of the last century. But with the patent system
you never know...

Dale B. Dalrymple

.

Relevant Pages

  • Re: Comparing matched and Wiener filters
    ... "filter across all time and sample" viewpoint of matched-filtering, ... the output for the sampling points (i.e. correlation, ... You are interested in comparing MF vs Wiener, ... MF does not maximize SNR in the ...
    (comp.dsp)
  • Re: Quantization Noise in FIR filters and SNR
    ... large order filter. ... The maxum error I will experience after quantization will be: ... The round off power from the Coeff for the whole filter will be: ... How can I calculate the SNR after the filter?? ...
    (comp.dsp)
  • Re: Quantization Noise in FIR filters and SNR
    ... I have a problem understanding the effect of noise in fix-point ... The maxum error I will experience after quantization will be: ... The round off power from the Coeff for the whole filter will be: ... How can I calculate the SNR after the filter?? ...
    (comp.dsp)
  • Re: am/narrow-fm/wide-fm/lsb/usb what else?
    ... >> With what receiver and which filter? ... >> is wide enough and the detector is set correctly. ... available eceivers with genuine ISB and DSB modes. ...
    (rec.radio.scanner)
  • Re: Sky Quality Meter
    ... That is a fast lens! ... detector we are using. ... using visual bandpass CCD photometry with standard stars of known Vj ... and j (i.e. a minus-IR filter) which is most similar to the ...
    (sci.astro.amateur)