Talk:Discrete frequency domain

Latest comment: 13 years ago by Lambiam in topic Discrete frequency

Dubious

edit

It's possible that one could make a topic of "discrete frequency", such as what one gets from a Fourier seriers or a DFT, but the definitions and factoids in this article are just wrong, apparently just made up, and really about about discrete-time signals. I don't see any way to salvage any part of it. Dicklyon (talk) 22:47, 10 February 2011 (UTC)Reply

The AfD has closed without getting rid of this article and I'm not sure how to proceed. I can't believe that the closing admin has any understanding of signal processing, so is there some way to find an admin who understands the subject and get this rid of this? The alternative seems to be to blank everything in the article, because, although the word "discrete" occurs in "discrete fourier transform" and "discrete time" they have no connection with a "discrete frequency", which has just as much meaning in the realm of continuous time. "discrete spectrum" is a term with meaning, but that also has nothing to do with the article as written. If no-one has any better suggestions I'll probably end up blanking the article on the grounds that it is rubbish. Dingo1729 (talk) 17:51, 18 February 2011 (UTC)Reply

I reduced the article to the small amount of information which can be said about the subject. Dingo1729 (talk) 19:10, 18 February 2011 (UTC)Reply

Discrete frequency

edit
(copied from user talk pages)

[@Lambiam] Hi, I saw you changed my (almost content free) version of the article to a new title. I just wondered where you were planning on moving with this. Discrete frequency domain doesn't seem to have much worth talking about either, and this wouldn't really give much insight to someone looking it up. The only time I can think of when discrete frequency domains arise is for for a periodic signal (which redirects to periodic function). So I wondered whether a way forward would be to add a section about fourier analysis there and turn discrete frequency domain into a redirect to there. Regards, Dingo1729 (talk) 21:28, 18 February 2011 (UTC) Incidentally, I'd probably dispute the statement that a discrete fourier transform goes from a discrete time domain to a continuous frequency domain. It goes from a discrete domain to a periodic domain. Which one we call time and which one we call frequency is just convention from engineers. But this is just quibbling. Regards again, Dingo1729 (talk) 21:38, 18 February 2011 (UTC)Reply

The problem with an article titled "Discrete frequency" is that it is like having an article Automatic soap (the stuff you get from automatic soap dispensers), so I moved it to a title that does make sense. Although I have no particular idea where one might go with this stub, this appears to me a better solution than maintaining an article on a non-existing topic. You are right, the function produced as output by a DFT also has a discrete domain; I've adjusted the article accordingly.  --Lambiam 22:03, 18 February 2011 (UTC)Reply
By the way, a signal does not need to be periodic to have a discrete frequency domain. Any weighted sum of sinusoidals will do; so-called discrete frequency noise sometimes manifests as a mix of almost pure whistle sounds that do not have common harmonics.  --Lambiam 22:10, 18 February 2011 (UTC)Reply
I agree with your analogy with automatic soap. That was the type of example which was being proffered in the Afd. On the other hand, "discrete frequency noise" is clearly parsed as noise made from discrete frequencies so "discrete frequency" isn't always without meaning. My view is simply that it isn't notable.
I'm not sure that I would speak of discrete frequency noise as having a discrete frequency domain. Domain isn't synonymous with support. In fact it seems to be a rather slippery word in this context, not quite meaning the same as "the domain of a function". I'm not sure how to define "frequency domain" although we all use the term and no-one ever queries it. Do you have any suggestion? Dingo1729 (talk) 06:47, 19 February 2011 (UTC)Reply
The notion of "discrete frequency noise" is not well-defined and commonly used concept, but I interpret "discrete" here as referring to the support of the frequency spectrum (at least, in an idealized form); "noise having a discrete-support frequency spectrum", however, is rather a mouthful and needs to be contracted. The plural "discrete frequencies" does make some sense, meaning "frequencies that are discrete from each other", but in the singular we get something like the sound of one hand clapping. In forming a compound noun with a plural for the first component, it is not uncommon to use the singular form while retaining the plural sense; boots that allow you to take seven leagues in one stride are seven-league boots. But an article topic Seven league doesn't make sense. Similarly, a spectrum consisting of discrete frequencies might become a "discrete-frequency spectrum".
I think "domain" in "frequency domain" means essentially the same as "domain of discourse"; it is the set of values over which the frequency variable may range.  --Lambiam 19:37, 19 February 2011 (UTC)Reply