Talk:Bandlimiting

Untitled
Is it finite support or compact support? — Preceding unsigned comment added by 64.118.25.228 (talk) 16:48, 14 July 2014 (UTC)

Hello! I hope that adding the following definition is useful:

Definition: The simplest band limited signal (SBLS) has two lines into its spectrum (direct current and a sine or cosine wave) and is given with one of the equations below:


 * A(ss) = Am sin (2*pi* F*t + j) + Bdc
 * A(cs) = Am cos (2*pi*F*t + j) + Bdc

References: ET 4 CO 198.pmd 

Best regards Petre Petrov ppetre@caramail.com


 * Oh, it's Petrov again. First of all, that's not a definition.  Secondly, it's not clear why the "simplest" shouldh have two components.  Why wouldn't the simplest be just a DC signal?  Thirdly, the two (sine and cosine) versions are exactly equivalent, so why write both of them?  So, no, I don't see how your publication is useful, as it's essentially uninterpretable. Dicklyon (talk) 19:40, 25 December 2008 (UTC)

SBLS
Hello! Yes that is me again. I have read several articles and I have found that something can be added.

1. Why that is not a definition? Please specify!


 * It sounded like a property, but if you meant it as the defining property of "simplest", I guess that's OK (but not OK for wikipedia, just for your idiosyncratic paper). Dicklyon (talk) 00:36, 27 December 2008 (UTC)

2. Usually a “band” has two borders (limits)? Or not?


 * Yes. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

3. Yes, the DC is the simplest signal.


 * Yes. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

4. There are huge difference between the DC and AC.


 * Yes or no, depending on your representation; by why to you need both to make a "simplest" signal? Oh, by definition... Dicklyon (talk) 00:36, 27 December 2008 (UTC)

5. The “simplest band limited signal” has a DC and AC component? OK? Or you have another definition?


 * It doesn't matter if I have a definition; what matters is whether there's a definition in a reliable source. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

PP: Sorry! It is very important to have a good definition! You will never advance in the right direction with inaccurate definition! —Preceding unsigned comment added by 78.90.230.235 (talk) 07:57, 27 December 2008 (UTC)

6. Sine are cosine are not “exactly” the same. They are two different mathematical functions. From the engineering point of view there are “90 degrees phase difference” between them and that is important in most of the cases.


 * When they include a free phase parameter, they are exactly the same family of functions. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

7. If you cannot interpret my publication you will not use it. (Sorry!) But if you have questions you could ask them.


 * Yes, why in the papers do you speak about the number of samples needed "per cycle" as if the cycles were independent? That seems to be what leads you to needing 4 samples per cycle; the family of signals you can describe that way is not bandlimited, though, so it hardly seems relevant. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

I hope now the situation is clear.


 * Not at all. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

BR

PPetrov —Preceding unsigned comment added by 78.90.230.235 (talk) 07:15, 26 December 2008 (UTC)


 * Even if your definition was meaningful, the fact is that no-one else uses it, so there's no reason to add it to the article. Oli Filth(talk 13:33, 26 December 2008 (UTC)

?? So, you do not use new meaningful definition because it is new!? Have you ever done something for the first time? Have you ever correct a mistake? Br P Petrov —Preceding unsigned comment added by 78.90.230.235 (talk) 15:09, 26 December 2008 (UTC)


 * Please see WP:Notability, and also WP:Original research. Oli Filth(talk 15:13, 26 December 2008 (UTC)

Many thanks. Obviously we are talking different languages. BR P Petrov —Preceding unsigned comment added by 78.90.230.235 (talk) 15:55, 26 December 2008 (UTC)


 * Yes, and as Oli points out, we speak the language of wikipedia policy. See also WP:V and WP:RS. Dicklyon (talk) 00:36, 27 December 2008 (UTC)

Why the things are not clear?

--

Hello!

Many thanks for the answers in that talk.

It was real pleasure to read them!

1. The situation is not clear because the terminology is not clear.

2. In order to make the situation clear you should always start from the beginning, from the simplest terms and tests. OK?

3. I am discussing the Analog to Digital Conversion of band limited signal according to the model “take and memorize”.

4. I am testing the AD Conversion with SS, CS and SBLS and looking for differences between the analog values and the digital values.

5. I am testing the AD Conversion with SS, CS and SBLS and looking for differences between the analog values and the staircase function of the output of the DAC during the direct reconstruction process.

6. The “classics” are discussing ANALOG MULTIPLEXING into analog transmition line! (Nyquist, etc). Nothing to do with the ADC and DAC! OK?

7. If you have only ONE cycle of the singe wave (SS) and you sample it with signal sampling factor (SSF) N=4 in the WORST case you will have -3dB bandwidth of the ADC. Or you will have a maximum of 3dB difference between the analog signal and the maximal digital sample. OK?

8. “Classics” do not have any evaluation of any error during the signal sampling and analog to digital conversion. OK?

9. Repeating inappropriate and inaccurate terms ant theories 10 000 000 times or more do not make them true, accurate or applicable to the real world. OK?

10. Errors should be defined clearly and evaluated exactly (if possible) OK?

11. You will never reconstruct an analog (real) signal exactly (even in theory) UNTIL  you will not prove that the natural number e = 2.71…. and pi = 3.14… are FINITE OK?

Legal matters: I am respecting the Wikipedia policies but all these legal matters are a bit complicated for me. That why perhaps I will leave the talks. I do not want any troubles and I personally am not very interested to correct the articles here. Sorry!

BR

P. Petrov —Preceding unsigned comment added by 78.90.230.235 (talk) 07:51, 27 December 2008 (UTC)

PS: May be you will not like that but I do not want to repeat again the contents of the papers below. I am open for questions. And I will try to explain the contents if you wish.

References:

9 CP PE 8.pmd www.ieindia.org/pdf/89/89CP109.pdf

Revaluation and replacement of basic terms in the sampling theory www.pueron.org/pueron/nauchnakritika/Th_Re.pdf

BR, P. Petrov —Preceding unsigned comment added by 78.90.230.235 (talk) 09:27, 27 December 2008 (UTC)


 * Please stop pasting links to your papers over and over again. As Dicklyon has already stated, the contents of these papers is invalid.  You've invented a whole set of terminology in an attempt to address a "problem", but that problem is merely due to your choice of a sub-optimal filter, or by trying to sample non-bandlimited waveforms.  This is already well-understood, and none of it affects (for instance) the sampling theorem, and none of it has any reason to appear in any Wikipedia article.


 * Incidentally, there are no "legal" issues! Wikipedia has several policies which dictate what should be included in an article and what shouldn't, but none of it is legally enforcable!  Oli Filth(talk 12:16, 27 December 2008 (UTC)

PP: Where did you find "non-bandlimited waveforms" stated by me?

"sub-optimal filter"??? - Where?

" no one of it has any reason to appear in any Wikipedia article." OK. —Preceding unsigned comment added by PetrePetrov (talk • contribs) 13:12, 27 December 2008 (UTC)


 * In one paper, you discuss sampling triangle waves and so on; these have infinite bandwidth. As for sub-optimal filters, the zero-order hold is a sub-optimal filter; it does not have finite bandwidth ($$h(t)=\mathrm{rect}(t/T) \Leftrightarrow H(fT) = \mathrm{sinc}(f)$$) and hence is not what the sampling theorem is talking about. Oli Filth(talk 13:18, 27 December 2008 (UTC)

PP: OK but not exactly. That is for a geometrical reconstruction with solving equations. Not with DAC and filters. Thank you for paying attention to my papers. —Preceding unsigned comment added by PetrePetrov (talk • contribs) 13:23, 27 December 2008 (UTC)

Definition of timelimited?
I think it'd be helpful and clear a lot of confusion about the timelimited vs. bandlimited proof, if there would be a proper definition of when a signal is timelimited.

Also, I don't understand, why would X seconds of a bandlimited signal require infinite time to transmit? Why not some-function-of-X seconds? PAStheLoD (talk) 10:45, 10 March 2009 (UTC)


 * The section already defines what is meant by timelimited; however, I've now wikilinked support.


 * As to your question, that's the whole point of the proof! Oli Filth(talk 20:00, 10 March 2009 (UTC)

Misleading comparison
I think it is a little misleading to say that bandlimited signals require infinite time to transmit. It is true that they can't have compact support, but often in signal processing, one considers what would happen if the signal (as we observed it) repeated itself infinitely often (the periodic continuation). It is then reasonable to talk about a finite, actual signal being bandlimited in this sense.--2003:69:CD03:4301:2E81:58FF:FEFF:8F4B (talk) 22:23, 13 August 2013 (UTC)

Too Abstract Template
I am adding a too-abstract template to the article because it is highly technical and jargonistic, and contains no examples (unless you count one tangential mention of the mathematical basis for the quantum mechanical uncertainty principal). Examples of popular applications of bandlimiting could also help remedy the issues implied by the "Technical" template that was added in January 2013 (not by me). Perhaps an example in the field of audio signals would help? —Adallace (talk) 01:09, 17 June 2015 (UTC)