Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: Iterative Match Filterings
Replies: 33   Last Post: Apr 17, 2012 6:13 PM

 Messages: [ Previous | Next ]
 Eric Jacobsen Posts: 107 Registered: 12/7/04
Re: Iterative Match Filterings
Posted: Apr 6, 2012 5:58 PM

On Fri, 6 Apr 2012 14:11:55 -0700 (PDT), Bret Cahill
<BretCahill@peoplepc.com> wrote:

>> >> >>>>>> Do you mean matched filtering or is "match" filtering something
>> >> >>>>>> different?

>>
>> >> >>>>> North filter.
>>
>> >> >>>> Far more commonly known as a "matched filter",
>>
>> >> >>> How long did it take you to figure that out?
>>
>> >> >>>> which requires no
>> >> >>>> special kind of convolution, only the ordinary, mundane kind. =A0The=

>> >re's
>> >> >>>> no deconvolution involved.
>>
>> >> >>> How is the original waveform recovered with a North Filter?
>>
>> >> >> Bret, can you stick with the common semantic? =A0before today, i have
>> >> >> never heard of a "North filter".

>>
>> >> > Is anyone still whining about "match filter" or "matched filter", both
>> >> > of which google up tens of thousands of on point hits in less time
>> >> > than it takes to play trifling word games.

>>
>> >> >> i am assuming from Eric's response and
>> >> >> yours, that it is synonymous with "matched filter".

>>
>> >> >> a matched filter is not about recovering an original waveform.
>>
>> >> >http://en.wikipedia.org/wiki/Matched_filter
>>
>> >> > Scroll down to the match filter recovery of the binary signal
>>
>>
>> >In that example the "detection" of the signal requires recovering the
>> >original waveform.

>>
>> You're calling it a waveform, other people may call it the modulated
>> data,

>
>And some may call it Shape XYZ.
>

The point is that people can't read your mind regarding what YOU mean
when you throw phrases around like "recover the original waveform" in
a context where it isn't typically used. Matched filtering is not
typically used to recover a waveform, although it can be and that way
of looking at it isn't necessarily wrong, but it's up to you to make
clear what you mean if you want people to understand you.

>. . .
>

>> >> also, Wikipedia should not be taken as authoritative
>>
>> >You think an authoritative article is necessary for an _example_?
>>
>> >Maybe you think the author put that nonsense in because he wanted to
>> >help out anyone trying to mislead on how match filtering can be used
>> >for wave form recovery?

>>
>> Again, you miss the point by quite a long ways.

>
>The communication problem is on _your_ end.

You're right, I still have no idea what you're talking about, despite
doing DSP and matched filtering applications professionally for a few
so far you haven't explained anything new about matched filtering that
makes any sense, and you seem to have some significant misconceptions.

I await your clarifications using explanations or terminology that can
be understood by an experienced practitioner.

>The link below shows another example of match filtering, not merely
>detecting, but recovering the original waveform before the noise was
>
>http://www.complextoreal.com/chapters/mft.pdf
>
>It may help explain why you are having trouble with the term
>"match[ed] filter."

We notice that you've not addressed any of the points brought up that
could help clarify what you mean, and now you point to a document
which properly explains the usual concepts yet doesn't clarify
anything about what you've been describing regarding deconvolution and
frequency-domain processing, or the use of terms as you've used them.

Charan's tutorials are pretty well known and respected for
introductory treatments to topics, which he's good at. I can't figure
out what you're talking about, though, and you seem unable to clarify
it and unwilling to address the clarifying points that have been
raised by others.

Eric Jacobsen
Anchor Hill Communications
www.anchorhill.com