Talk:Digital signal/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Too technical

Everybody knows about digital sound now, but this article is just too technical for the ordinary person to understand. It doesn't explain technical jargon at all; it sounds as if it were written by a specialist who assumes that anyone reading the article is also a specialist. AlbertSM (talk) 02:41, 7 October 2011 (UTC)

This is a technical topic. You're probably looking for Digital. I've added a hatnote. ~KvnG 17:04, 19 June 2013 (UTC)

Digital data

I find it problematic that Digital data redirects here. This article only talks of a physical signal, while digital data could be considered in a more abstract way as information, not necessarily requiring a discussion on the physical representation. There's also Data (computing), but it doesn't discuss the digital aspect. Olli Niemitalo (talk) 12:23, 1 February 2012 (UTC)

Digital data has been changed to redirect to Digital. ~KvnG 17:04, 19 June 2013 (UTC)

Two topics

The What Links Here list for Digital_signal has 309 entries, and oddly, Digital Signal Processing is not one of them. Perhaps the name "Digital Signal" is a misnomer. Would "Boolean signal" or "Data signal" be more appropriate? It is hard to tell what those 309 links are trying to reach, because Digital_signal is about two very different topics. If we purify it, by removing the sampling/quantizing parts, how many of those 309 links will lose the content they wanted?
--Bob K (talk) 15:54, 1 August 2015 (UTC)

We definitely want to get to the point where this article is about a single type of signal. I'm going to propose we split this: ~Kvng (talk) 17:05, 1 August 2015 (UTC)

I concur. If Wikipedia has a mechanism for alerting the 309 linkees that they might need to relink to Logic signal, this would be an occasion to use it. Otherwise, we should add something to Digital signal to re-direct the Logic signal seekers.
--Bob K (talk) 00:01, 2 August 2015 (UTC)- Bob

We can put hatnotes on both articles to help address any misdirection. Reviewing the 309 links is not an unreasonable chore. I usually find other useful improvements to make along the way when I do a project like this. ~Kvng (talk) 13:20, 4 August 2015 (UTC)

Digital signal begins with the statement: "This article is about digital signals in electronics. For digital data and systems, see Digital data." So I propose we convert it into a disambiguation page as follows:

As you can see, those links now exist. I created the articles from the material in this one. Disambiguation also solves the problem of the 390 linkees.
--Bob K (talk) 13:51, 20 August 2015 (UTC)

I've reverted these changes, for multiple reasons:

  • digital signals do not necessarily have a clock
  • disambiguation pages are used when two topics are completely different. In this case they are effectively the same topic (they're deeply interrelated), so a disambiguation page is inappropriate.
  • material was apparently deleted during the move; it appeared that the material was still necessary.GliderMaven (talk) 18:09, 21 August 2015 (UTC)

OK, so I've essentially installed the digital electronics sense of digital signal here (a continuous signal that is meant to represent numbers), because I think that's what people really mean when they talk about digital signal, and I've added a hatnote to point to the article that Bob split off which is about the discrete sequences of numbers that are specifically used in digital signal processing, which is a much less frequent use of the term.GliderMaven (talk) 20:19, 21 August 2015 (UTC)

What data supports that opinion? Here is a counter-indicative Google Scholar link: scholar.google.com
--Bob K (talk) 12:13, 23 August 2015 (UTC)
Your search for 'digital signal' mostly returns 'digital signal processing'. But does the term 'digital signal processing' mean '(digital signal) processing' or 'digital (signal processing)'??? I think the latter makes more sense.GliderMaven (talk) 12:44, 23 August 2015 (UTC)
But when you dig deeper into the content of the links, I think you will change your mind.
--Bob K (talk) 14:04, 23 August 2015 (UTC)
I thought you agreed that there were two (related) definitions, one used in digital signal processing, and the other that refers to digital logic.GliderMaven (talk) 14:30, 24 August 2015 (UTC)
What I agreed to is right up above (Kvng 17:05, 1 August 2015), repeated here for your convenience: Two articles,
I didn't say they are related, and I don't think so now. I subsequently realized that two more-similar names and a disambiguation page would be more consistent with other Wikipedia treatments. Note that "similar names" is not synonymous with "related definitions".
--Bob K (talk) 21:11, 25 August 2015 (UTC)
Having looked into this further, the term digital signal does not refer to quantized discrete time signals in any reference I've examined, and I'm finding that a 'logic signal' appears to be a subtype of 'digital signals'.GliderMaven (talk) 02:02, 26 August 2015 (UTC)

In terms of linkage issues, that seemed to work for all of the articles I checked, but I only checked a few, but logically speaking the strategy should rarely fail.GliderMaven (talk) 20:23, 21 August 2015 (UTC)

What seemed to work for all articles? ~Kvng (talk) 15:40, 29 August 2015 (UTC)
All of the linked articles I checked were not incorrect when linked to the current article on digital signals; and digital signal as it is currently defined is pretty clearly the WP:PRIMARY topic, and logic signal is also very clearly a subtopic of that.
I could not be more strongly opposed to turning digital signal into a disambiguation page.GliderMaven (talk) 16:27, 29 August 2015 (UTC)
I have not proposed to turn Digital signal into a disambiguation page. My proposal is to WP:SPLIT material having to do with digital logic to Logic signal. Do you have any comment on this proposal? ~Kvng (talk) 15:02, 2 September 2015 (UTC)
There doesn't seem to be any reason at all to split articles right now. I would support a merged article covering digital signals/logic signals here, or at logic signal. Splitting should really only ever be done if an article becomes unwieldy and it isn't.GliderMaven (talk) 15:29, 4 September 2015 (UTC)
Please read WP:SPLIT, specifically WP:CONSPLIT. ~Kvng (talk) 15:39, 5 September 2015 (UTC)

It looks like GliderMaven has edited the article so that it contains only the content I proposed to split to Logic signal. It didn't take long to find a case where this does not work for links to this article. What's the idea for fixing, for instance the link here from Analog signal? I don't think we have consensus here and would support reverting these edits until we do. ~Kvng (talk) 15:02, 2 September 2015 (UTC)

Not quite right. There's a difference between a digital signal and a logic signal. A logic signal is of necessity a binary digital signal, whereas the concept of a digital signal is more general, and includes ternary and higher signals. This article is about digital signals, hence it's name.
And yes, a lot of the article is about logic signals, because that's the most common type of digital signal, but certainly not all digital signals are logic signals.
We clearly need a real article at digital signal, to cover the generality, and right now, there's not enough material to support a separate logic signal article, but I have no problem in principle with having that as well as a subarticle.GliderMaven (talk) 18:14, 2 September 2015 (UTC)
I've also just noticed there's an error in the definitions above; logic signals are not necessarily 'baseband', manchester encoding is logic signal, but not baseband.GliderMaven (talk) 18:16, 2 September 2015 (UTC)
It concerns me very greatly that you earlier rewrote the lead of the article, and have made multiple proposals that are clearly deeply flawed in similar ways. If you don't know what a digital signal, or a logic signal are; and you do not appear to, if you make an article on a topic into a disambiguation page which has hundreds of links to it, and that almost certainly should remain an article, then in my opinion you're displaying deeply problematic behaviours.GliderMaven (talk) 18:24, 2 September 2015 (UTC)
The topics that you digress to above are worthy of discussion but the immediate issue we need to resolve is how to handle all the incoming links. Your recent changes to this article have narrowed its scope and so many of these links are no longer appropriate. I see the following options:
  1. Revert your recent changes and discuss further
  2. Review all incoming links and re-target as appropriate
I presume you would like to avoid 1. I'm not afraid of a little work and so would be happy to help with 2 but I got stuck pretty quickly trying to carry this out. Do you have any suggestions e.g. where the Digital signal link from Analog signal should be targeted? ~Kvng (talk) 18:30, 2 September 2015 (UTC)
I disagree. It seems to me that 2 of the 3 interested editors do not agree with hijacking digital signal for the exclusive use of "a continuous signal that is meant to represent numbers". It should not exclude signals that actually are numbers, and there are plenty of DSP books using that term. Changing tens or hundreds of links based only on a minority opinion seems unwise. I still vote for the original proposal (2nd) or disambiguation (1st).
--Bob K (talk) 14:00, 3 September 2015 (UTC)
Yes, to be clear, I am not convinced GliderMaven's changes are an improvement. I'm concerned about the incoming links not because of the work required to fix them but because there is now nowhere to point many of them. I do not support converting this article to a disambiguation. I would like to implement my proposal and create Logic signal. If we don't hear from GliderMaven, since we don't have consensus, I think we need to do another revert. ~Kvng (talk) 14:51, 3 September 2015 (UTC)
I don't have a problem with merging with logic signal, provided you still define the more general digital signal correctly in the article, as well as defining logic signal. We can discuss about whether it's at logic signal or digital signal. I have massive problems with merging digital signal with Bob's aphysical definition which is related to a dirac comb; that seems to be a hugely more obscure and niche definition and is clearly not the WP:PRIMARY one.GliderMaven (talk) 15:46, 3 September 2015 (UTC)
The digital signals I'm talking about can be obtained from an analog-to-digital converter, or even a computer program. There is nothing aphysical about them.--Bob K (talk) 02:40, 4 September 2015 (UTC)
That's not what the article you created says: Digital signal (signal processing) "a digital signal is a discrete-time signal for which not only the time but also the amplitude has discrete values". What does discrete time mean in that context??? Discrete means non continuous, so that intervening portions don't exist at all. It's an aphysical abstraction. A real world A-D converter cannot produce a digital signal in that sense.GliderMaven (talk) 03:49, 4 September 2015 (UTC)
The Dirac comb is an abstraction. ADC output is not. And of course physicality is irrelevant anyway. Mathematics does not require physicality. Some signals exist physically. Others are mathematical formulas.--Bob K (talk) 12:33, 4 September 2015 (UTC)
@GliderMaven: I'm not aware of any merge proposal. Can you clarify?
If we don't get a response to how you intend to address problems with the links to Digital signal introduced by your changes, I intend to revert your changes. ~Kvng (talk) 17:15, 3 September 2015 (UTC)
What problems are you referring to? I'm aware of precisely zero linkage problems right now, and certainly none of my making; in spite of a your repeated claims to the contrary. Indeed, the edits I reverted, turned this article into a disambiguation page and immediately created over 300 separate link problems; since articles are not supposed to link directly to them. You guys are the ones breaking links, I have not broken any links at all. Indeed, between you two you deleted the only article we had on digital signals and created two on subtopics. Look, I have a background in digital electronics, digital signal processing and computer science and computer engineering. I do actually know this topic quite well.
I can see we need to get other people involved, it's pretty clear to me that your are not interested in achieving true consensus.GliderMaven (talk) 20:52, 3 September 2015 (UTC)
I'm now repeating this for the third time in this discussion section: Do you have any suggestions e.g. where the Digital signal link from Analog signal should be targeted?
WP:RFC is the usual mechanism to involve other editors in the discussion. ~Kvng (talk) 14:55, 4 September 2015 (UTC)
Link it to this article, where it's already linked. What's the problem with that?GliderMaven (talk) 16:58, 5 September 2015 (UTC)
What's discussed in Analog signal and many of the other incoming links is a signal that is a sequence of digital values i.e. Digital signal (signal processing), not a voltage signal as is currently described in your most recent revisions. ~Kvng (talk) 18:05, 5 September 2015 (UTC)
Discrete time signal
Um. No. The type of digital signal referred to at Digital signal (signal processing) "is a discrete-time signal for which not only the time but also the amplitude has discrete values". That kind of signal looks like the red arrows on this diagram:
A digital signal has two or more distinguishable waveforms, in this example, high voltage and low voltages, each of which can be mapped onto a digit. Characteristically, noise can be removed from digital signals provided it is not too large.
This kind of signal is aphysical. No real world signal can be like that, because the red arrows don't exist in between, they are discrete time values. However, it is an abstraction extensively used in digital signal processing.
The type of digital signal referred to in analog signal is not that kind of signal, it is a real world signal, as currently described in digital signal. A diagram of that kind of signal may look like this: GliderMaven (talk) 23:06, 5 September 2015 (UTC)

Here's the sentence from Analog signal that I'm referring to, "It differs from a digital signal, in which a continuous quantity is represented by a discrete function which can only take on one of a finite number of values." An example context of audio signals is given. The nature of digital audio signals is best described by your first picture and Digital signal (signal processing). ~Kvng (talk) 15:50, 6 September 2015 (UTC)

It's not at all clear to me that that sentence is correct, complete or self explanatory, but I think the key phrase here is continuous quantity, which refers to the continuous digital waveform, and the discrete space referred to are the digits that that continuous waveform represents. If it doesn't mean that, then it's probably wrong, since I'm 100% sure that there's digital signals flowing in my computer, and they for 100% certain are real digital signals, and real digital signals are, of necessity, differentiable. Discrete signals are non differentiable (although they are subject to related operations).
I also note that that entire article is completely and totally unreferenced, so drawing conclusions from it to make major changes to the structure of Wikipedia, seems to be rather unwise in the extreme.GliderMaven (talk) 19:13, 6 September 2015 (UTC)
So now you need to claim that Analog signal is wrong in order to try to show that the link to your new take on Digital signal is right. These gymnastics indicate to me that there may be an ownership being defended here. ~Kvng (talk) 03:28, 7 September 2015 (UTC)
I don't even have to claim it is wrong; given that that article is completely unreferenced.GliderMaven (talk) 07:05, 7 September 2015 (UTC)

Another proposal

Since the previous version of this article covered two topics and there is not consensus as to which is the main topic and GliderMaven has edited this article to now cover just one topic and BobK has created Digital signal (signal processing), I propose to rename this article Digital signal (electronics) and convert this page (Digital signal) to a disambiguation page. I know that GliderMaven has voiced opposition to creating a disambiguation page arguing that the topics are too interrelated for that treatment. While making that argument, GliderMaven has edited Digital signal so that it covers just one topic. Maybe a disambiguation is not where we eventually want to be but we need to work at incremental improvements here at WP and having a disambiguation page would be an improvement over where we are right now and would be an improvement over reverting all of GliderMaven's changes. ~Kvng (talk) 16:12, 5 September 2015 (UTC)

FYI, I already implemented exactly that proposal. But it was reverted by GliderMaven so quickly that you may have missed it. It was easy to do, and no WikiLinks were affected.--Bob K (talk) 17:23, 6 September 2015 (UTC)
Yes I missed it :( ~Kvng (talk) 03:14, 7 September 2015 (UTC)
To sound like a broken record, I don't think a disambiguation page here is a good idea. It would be a lot of work, and pointless work, to move over 300+ links, since you shouldn't link to disambiguation page. Or in terms of WP:PRIMARY TOPIC, I think that the current definition is the one everyone really means (to the extent that they've thought it through, although very, very often the term is used synonymously with logic signal, which this article also covers). I think that digital signal is a WP:BROADCONCEPT. I would much prefer to see the broadconcept article here moved to logic signal, leaving a redirect behind or leave it where it is.
But there's also material under this topic, which has not yet been described here, incredibly closely related to digital signal (signal processing) since a discrete time, discrete value 'signal' is the logical effect of clocking a buffer or sampling to an A-D converter at discrete times, and in this sense the digital signal (signal processing) article is really a subarticle of this one.
When topics are broadconcepts you never really want to make disambiguation pages, it tends to obscure the subject matter, and the reader cannot easily get an overview and there ends up with excessive duplication and errors fixed in one place persist elsewhere and so forth.GliderMaven (talk) 16:54, 5 September 2015 (UTC)
I should have mentioned that I am prepared to do the edits to fix the incoming links. I'm prepared to all the work, in fact. I understand your view on where disambiguation should not be done. I actually don't know if I agree or disagree with this. There is room for a judgement call here and sometimes it is not clear where things need to go until you start going somewhere. I am not looking for THE ANSWER, I am trying to propose incremental changes that are clearly improvements. I think that the changes you've made are an improvement in once sense, we have a good coverage of the digital signal as it relates to electronics. Thanks you. I'd like to keep these contributions. The changes are a regression in that material that relates to the signal processing aspect has been removed. So on balance, arguably not an improvement and so at risk of being reverted. ~Kvng (talk) 18:18, 5 September 2015 (UTC)
I added some literature and illustrations showing definitions used in different contexts. I tried to use some of your formulations but partly reverted to "my" old version. I hope you can continue the good work utilizing the sources.Mange01 (talk) 22:37, 6 September 2015 (UTC)
The pictures themselves are mostly quite good, but there's huge problems with the changes you've just made. The idea that a digital signal, as most people consider it, is discrete in time is simply incorrect; that's an abstraction you can derive from a digital signal. Also, digital signals aren't necessarily quantised amplitude, for example you get uniform sinewave amplitude waveforms with FSK, but the frequency varies. No quantised amplitude. Then there's also big style problems; articles in wikipedia aren't about a term, we're not talking about the term "digital signals" we're talking about digital signals, the thing; so we can't have 3 different definitions, we have to have one definition, and this doesn't look like it's it.GliderMaven (talk) 22:48, 6 September 2015 (UTC)
You have a point. Wikipedia should reflect different contexts and meanings of the term. The easiest way is to discuss them in one and the same article, but perhaps it is better to create a Digital signal (disambiguation) page and several articles with different meanings of the term, for example Digital signal (electronics), Digital signal (communications), and Digital signal (signal processing). What you say is true in accordance with the tele transmission/data communication sources I gave (which also is my own background), but not in electronics and computer networking. Please put the sources you removed back, or motivate each of the deletions here and provide your own sources. Have you seen sources in digital electronics supporting that your definition also is valid also in that subject? Mange01 (talk) 23:11, 6 September 2015 (UTC)
Pulse-with modulation is digital according to definition (2) if the pulse withs are discrete
Another suggestion for illustration to the right. I wait with touching the article to give you time to work on it. Mange01 (talk) 23:25, 6 September 2015 (UTC)
I think we have to agree on what we want from this article. At the moment the links to it seem to treat it like: Digital Signals- University of St Andrews (not a reliable source) or Analog vs. Digital. But we don't necessarily have to stick with that.GliderMaven (talk) 00:25, 7 September 2015 (UTC)
Thankyou for improving the article. My image captions are not that well-turned, so don't hesitate to improve them.
Before we discuss if we should revert the disambiguation page to its old version, first we need to agree on if "digital signal" has the same definition in electronics as in communications (as you claim) or different meanings in the two contexts (which I think). Please put back or comment on the academic sources provided in my last revision, or provide an academic source in electronics that supports your view. Mange01 (talk) 16:28, 7 September 2015 (UTC)
Well, I actually used to work in telecommunications- specifically fibre optic equipment R&D, and, so far as I am concerned, all modern telecommunications can be considered to be fully digital end-to-end. Certainly all fibre is digital, all cell phones, WiFi, Ethernet, USB etc. The days of Strowger exchanges are far behind us! But we certainly need to find some more references. There's still some analog broadcast and two-way radio that is still just about hanging on, but that's about it I think.GliderMaven (talk) 18:34, 7 September 2015 (UTC)

What is this all about? The title Digital signal (signal processing) won't work because people won't like the repeated word or the made-up nature of the title, and it is confusingly similar to the standard term Digital signal processing. Is there a problem in the current content at Digital signal (GliderMaven's version)? What? Can someone point to a brief discussion showing the problem? I see #Two Topics above, but "Boolean signal", "Data signal", "Logic signal" are not standard terms, so that can't be it? Johnuniq (talk) 21:52, 7 September 2015 (UTC)

The title Digital signal (signal processing) is consistent with Sampling (signal processing), Pulse (signal processing), Filter (signal processing), Decimation (signal processing), Bandwidth (signal processing), Folding (signal processing), Quantization (signal processing), Innovation (signal processing), Noise (signal processing), Energy (signal processing), Clipping (signal processing), and Coherence (signal processing).
--Bob K (talk) 15:35, 8 September 2015 (UTC)
So far as I know a 'logic signal' is a reasonably well used term for binary "logic" signals (also see logic analyser which basically analyses logic signals).
The Cliff notes go something like: Kvng and Bob_K turned this page version into a disambiguation page after this brief discussion: Two topics. I reverted it. They were upset. They've also been basically claiming that none of the 300+ inbound links are really correct, or at least wouldn't be after they'd made it into a disambiguation page, but I had no need to worry, they would fix all of them. If everyone else agrees with them fine, but it seemed a bit odd to me.GliderMaven (talk) 22:35, 7 September 2015 (UTC)
Yes, people talk about "logic signals" when sticking probes on pins, but it's just jargon associated with digital logic and is only a peripheral part of a "digital signal" (I guess a logic signal is a physical representation of a digital logic calculation). The current article here looks fine. Sorry I'm too lazy to check, but if anything was moved from here to Digital signal (signal processing), I would think it should be restored here and the new page made an implausible redirect to here. Johnuniq (talk) 23:23, 7 September 2015 (UTC)


GliderMaven subsequently "resolved" the original issue ("two topics") via rev 14:58 21-Aug explained as follows: "OK, I've decided that the most common use of 'digital signal' refers to a quantised (usually binary coded) continuous signal that may also have a clock; and hence this is that article". In other words, he/she hijacked digital signal for the exclusive use of continuous-time signals and functions. I don't contest the irrelevant claim that it's the "most common use", but that is not synonymous with "unambiguous". And in fact, all 6 citations in Digital_signal_(signal_processing) define "digital signal" as having the discrete-time property.
I do not understand and did not participate in the quarrel over incoming links to the reverted disambiguation page. Those seeking "a continuous-time waveform transitioning between discrete logic levels" will find an appropriate link, and those seeking "a quantized, Discrete-time signal" will find a different link. That's what disambiguation pages are for.
--Bob K (talk) 16:46, 8 September 2015 (UTC)

Definitions

OK I've temporarily reintroduced the definitions, but they seem pretty bad. For example it states that: In digital electronics and computer networking, a digital signal is a pulse train (a pulse amplitude modulated),. OK, first I couldn't find that claim in the attached reference. Second, isn't it the other way around, isn't a pulse train a specific type of digital signal? And pulse amplitude modulated says it is a series of 'pulses', and there's a nice diagram at that article showing me what I believe a series of pulses actually is. To me a pulse is something that starts low, goes high, and then goes low again, where each pulse is of similar length, and while some do, pretty much, digital signals don't do that. And the rest of the sentence says: "i.e. a sequence of fixed-width square-wave electrical pulses or light pulses, each occupying one of a discrete number of levels of amplitude." which backs me up, but looks nothing like a logic signal normally does. I mean, to take a corner case, if a digital signal just stays low, or stays high, it's not a digital signal???GliderMaven (talk) 15:18, 15 September 2015 (UTC)

What is this all about

The article had been stable for some time covering two common meanings of "Digital signal". See, for instance, this 14 May revision. Some editors felt that an article should cover just one topic. Others felt that the two topics were too closely related to merit a WP:SPLIT. Bold edits and reverts and discussion ensued. GliderMaven then took a firm position that there was only one topic, the electronic term, to be covered and edited the article to remove coverage of sequence of discrete values and has reverted attempts to restore coverage of this second topic to which many of the incoming links to Digital signal refer. ~Kvng (talk) 04:03, 8 September 2015 (UTC)

What is your proposal? Would it be okay to start out from this version, where I provided sources supporting that people in electronics define digital signal differently than people in communications and in signal processing do? Those sources were deleted by GliderMaven. I wrote "please ...provide an academic source in electronics that supports your view" (that people in digital electronics also consider for example a PSK signal as digital), with no response. Simple because such sources do not exist, or are really rare. Mange01 (talk) 16:56, 13 September 2015 (UTC)
I put forward at least two proposals in the unproductive discussion above. I don't choose to work with GliderMaven on this any further at this time. The next step may be WP:RFC but I haven't yet come up with a crisp question to be resolved. ~Kvng (talk) 20:21, 13 September 2015 (UTC)
I initially thought you were questioning whether telecommunications transmitted digital signals, but then I realised you were actually asking whether digital electronics uses digital signals. Which is actually with all due respect, a bizarre question, since digital electronics was invented purely to handle digital signals rapidly to make things like computers work and to decrypt cyphers. (Sorry.) Most of the textbooks tend to use vague terminology like 'signal', 'input', 'output' and if you're very lucky 'logic signal'. If you want academic references to what digital signals are that's fairly easy, although getting a non educational reference is far harder, pretty much nobody except educators bother defining such mundanities because everyone is supposed to know what they are. For example: [1] carefully defines digital signals as taking finite domain, and then separately defines discrete time domain signals (such as you get from sampling) as being related. (i.e. You can have discrete time domain analog signals or discrete time domain digital signals or continuous versions of each.) It seems to me that this article needs to be, and is, about both types of digital signals.GliderMaven (talk) 21:27, 13 September 2015 (UTC)
No, that is not what I meant. Please read this version. My main point is that in literatur on communication the result of a digital modulation scheme (for example a phase-shift keying signal) is considered "digital communication" and "digital signal", while in literature on electronics and computer networking, such a signal is considered analog. Also, in theoretical signal processing, a digital signal is a time-discrete signal, which only exists at the sample instants, while in electronics it is a step-wize continous signal, also known as a pulse-amplitude modualted signal, which also exists in between the instants. The sources I provided that you deleted showed this difference. Please put these sources back. The easiest way to find academic sources is to search at books.google.com, sometimes at scholar.google.com . Thankyou for your new citation, Kulkarni, "Information Signals". It is a book in signal processing rather than communications, and consequently do not discuss if digital modulation results in digital or analog signals.
So, how can we reach an agreement that does not contradict the sources?Mange01 (talk) 07:06, 14 September 2015 (UTC)
If you think about it carefully, all digital signals are analog (or more accurately still, continuous waveforms), it's only how they're created and terminated that determines whether they're 'really' digital or not. And you can treat digital signals as analog, just fine, amplify, distort them, filter them etc. Pretty much if they're able to pass with 100% veracity (provided the noise isn't too great), through any system or between subsystems then they're digital, and that's true whatever they look like in the middle.
Note that the article currently covers the time-discrete signal abstraction; in reality there's always staircase signals (or similar) of course, but sampling/gating any signal means that only the sample timepoints count for downstream processing, so you can abstract the signal as being at only the sampling timepoint; but that's only really an abstraction to make the mathematics easy, as I'm sure everyone here is aware.GliderMaven (talk) 14:29, 15 September 2015 (UTC)
None of that rationalization addresses the issue, which is that digital signal is ambiguous. As I have said before, "all 6 citations in Digital_signal_(signal_processing) define "digital signal" as having the discrete-time property". Wikipedia acknowledges the existence of overloaded terminology, and its mechanism for dealing with it is disambiguation.
--Bob K (talk) 00:54, 16 September 2015 (UTC)
No it isn't. Disambiguations are intended to separate completely different meanings. When there's related meanings, as here, and there's a topic that covers all of them, then we have an article on that topic that covers, compares and contrasts them.GliderMaven (talk) 17:40, 16 September 2015 (UTC)
We don't have to turn this page into a disambig. I'm sure Bob K is willing to discuss alternatives. Can you please address his (and my) concern that there is ambiguity around "Digital signal". This was formerly addressed in content you have removed from the article. How do we get this back? ~Kvng (talk) 16:35, 17 September 2015 (UTC)
We're not here to define the term 'digital signal' we're here to talk about digital signals, how they are created, how they are transmitted, how they are decoded, and what abstractions there are made of them. If you consider that to be 'ambiguity' no. Just because multiple things are referred to with the same label, doesn't make it ambiguous. Is the term 'cat' consider ambiguous? I mean you could say that it refers to house cats, lions, tigers, cheetahs, felids, and they're all different, but no it's not ambiguous, it's an umbrella term. Are lions, tigers and house cats all precisely the same? Of course not, but they're all cats. And are digital signals in telecommunications and digital electronics and digital signal processing, all exactly the same, no, but they're all digital signals in the same sense- they all directly relate to signals that represent numbers. That's not ambiguity in the sense of a disambiguation page.GliderMaven (talk) 22:07, 17 September 2015 (UTC)
But I'm not opposed to there being a disambiguation page as well. For example the term 'digital signal' can refer to someone raising their fingers at someone as an insult; now that's an ambiguity, the current things you're talking about really aren't.GliderMaven (talk) 22:07, 17 September 2015 (UTC)
Point taken. OK, fine, I'm out-voted regarding disambig. So the original article has it about right, which brings us to ~Kvngs question: "How do we get this back?"... or more pointedly... "When?".
--Bob K (talk) 23:09, 17 September 2015 (UTC)
At the very least, there seems to be textual problems with that article version that the current version does not have. The current version also seems to be higher quality on the whole, it has better diagrams and it follows Wikipedia's policies better. I find that reverts should only be performed when it's unequivocal that the previous version is better, otherwise it's best to go forward. Notably, the discussion section in the old version is pretty awful, and definition 2 in the lead seems to be quite incorrect and contradicts definition 1.GliderMaven (talk) 01:47, 18 September 2015 (UTC)
Still, I don't think that the current version reaches C class yet, but that previous version definitely isn't above start. We need an article that is self-consistent, the current one is more consistent, but not quite there yet.GliderMaven (talk) 01:47, 18 September 2015 (UTC)
I agree and appreciate the improvements you have made to the article. What we're looking for when we review changes is an overall improvement. Coverage of the electronics aspect of digital signals has definitely improved. Coverage of discrete-time quantized signals, on the other hand, has been removed, a regression that potentially affects over 300 articles that refer to Digital signal. A mixed bag. ~Kvng (talk) 14:57, 18 September 2015 (UTC)

Doing stuff

Sometimes it is better to do stuff that to talk about doing stuff. I have created Digital signal (disambiguation). I'm not trying to piss anyone off; BobK has advocated for this and GliderMaven recently said he was unopposed. We have Digital signal (signal processing) and Digital signal (electronics) a.k.a. Digital signal. Now the next thing to do is go through incoming links and decide where to point them. ~Kvng (talk) 14:48, 18 September 2015 (UTC)

Creating a separate disambiguation page is completely fine with me. I'd also like there to be more about the discrete time digital signal abstraction in the article; it is currently subarticled, but I think there needs to be more in that section, I've also just added stuff to the lead as well.GliderMaven (talk) 19:40, 18 September 2015 (UTC)
I am happy with your latest changes Kvng and GliderMaven. THankyou for good work. Mange01 (talk) 21:22, 18 September 2015 (UTC)
Hello~. I have made some edits to the lead that made sense to me, but I am confused, what is the difference between an electronics digital signal and a signal processing digital signal? Perhaps that is what the article is about, but the way it was written was making me worry. From my understanding, a digital signal need not be continuous. Also, it seems to me that the idea of a digital signal is much more broad than this article. For example, the hand signals that a person directing an airplane into the dock at an airport takes on only a discrete set of values and qualifies as a digital signal. Where would something like this be mentioned? Is it appropriate for this article or for others? BourkeM Converse! 05:10, 22 September 2015 (UTC)

A sequence of numbers is not accurately characterized as an "abstraction"

Why the obsession with abstractions?

"In digital signal processing, digital signals are often described by an abstraction where the signal's value is sampled at regular intervals."
"In digital signal processing, a digital signal is an abstract signal that is considered discrete in time and amplitude, i.e. a sequence of codes, [9] representing a digitized analog signal signal."

DSPs operate on actual numbers, not Dirac combs.
--Bob K (talk) 01:48, 25 September 2015 (UTC)

Actually, if you're being ultra precise, DSPs operate on digital signal waveforms that are continuous, and do not operate on numbers per se. And it actually makes a difference in some circumstances; a DSP crashes if the waveforms that are fed into it change at the wrong moment (see metastability in electronics.) GliderMaven (talk) 02:10, 25 September 2015 (UTC)

Better source for the 3rd def?

It's currently based on an Indian textbook. These sometimes use terminlogy not found in other parts of the world, e.g. "tension" for voltage, or "soft saturation" instead of quasi-saturation (of a BJT). So can someone do a little survey and determine if that meaning/terminology is widespread in the field of DSP or just one book (or perhaps just one region). Thanks. 5.12.38.28 (talk) 19:29, 1 November 2015 (UTC)

I have to say that the 1st def isn't much better (sourced). Again an Indian textbook and one computer security book (hardly the epitome of authority on signals) that only gives a vague definition. 5.12.38.28 (talk) 19:46, 1 November 2015 (UTC)

Incoming links

Since the hatnotes and article now acknowledges that "Digital signal" has different meanings in different contexts and the focus of this article has shifted, it is time to review all Digital signal links and make sure they're appropriately targeted. As a first step, I will replace all Digital signal references with either Digital signal (signal processing) or Digital signal (electronics) as appropriate. ~Kvng (talk) 00:33, 2 August 2016 (UTC)

I've come across another type of digital signal in this work: Digital signal (broadcasting). This isn't well described by Digital signal (signal processing) or Digital signal (electronics). I've created a redirect with possibilities for this. ~Kvng (talk) 18:45, 2 August 2016 (UTC)

I have changed Digital signal (broadcasting) to redirect to Digital broadcasting which I discovered in the course of this work and which seems perfect. I'm going to leave Digital signal (broadcasting) as it is now because WP:NOTBROKEN. ~Kvng (talk) 19:03, 3 August 2016 (UTC)

Alternate targets for Digital signal topics: ~Kvng (talk) 22:29, 2 August 2016 (UTC)

 Done I have examined all links to Digital signal. Digital signal (disambiguation) has been improved to include the options I listed above. Subjects now addressing the electronic or electrical engineering digital signals now redirect here via Digital signal (electronics). With these redirects in place and in use, we're free to consider improvements to the titles and coverage in this topic area. ~Kvng (talk) 15:17, 3 August 2016 (UTC)

digital signal processing and digital signal processor

My understanding is that digital signal processing involves processing signals that represent continuous analog physical quantities, but using digital logic.
Thinking that a digital signal processor somehow processes "digital signals" seems to be a misunderstanding of the same type as thinking that a "red mountain bike" has something to do with "red mountains".
What can we do to straighten out such confusion? --DavidCary (talk) 16:37, 3 August 2016 (UTC)
This looks like a different issue than what I've been working on above. Probably best to start a new discussion at Talk:Digital signal processor or Talk:Digital signal processing. ~Kvng (talk) 17:51, 3 August 2016 (UTC)
Hyphenation is an easy solution. I.e "digital-signal processor" means a processor for digital-signals. "Digital signal-processor" means a signal-processor that used digital methods.
--Bob K (talk) 17:56, 3 August 2016 (UTC)
Perhaps it is grammatically correct but I have never seen digital signal processor (DSP) written out with a hyphen so there may be WP:COMMONNAME conflict. ~Kvng (talk) 19:07, 3 August 2016 (UTC)

Apparently Alan Oppenheim had the same issue, and switched from Digital Signal Processing to Discrete-Time Signal Processing, without changing the DSP acronym. Here is an excerpt from http://dspguru.com/book/export/html/33:

Here are some classic DSP books which have been widely used--but are now out of print. (Darn!).

  1. Theory and Application of Digital Signal Processing by Rabiner and Gold. A comprehensive, industrial-strength DSP reference book.
  2. Digital Signal Processing by Alan V. Oppenheim and Ronald W. Schafer. Another industrial-strength reference. (Replaced by the authors' Discrete-Time Signal Processing)
  3. Digital Signal Processing by William D. Stanley. A very readable book; has a strong treatment of IIR filters.

--Bob K (talk) 16:05, 5 August 2016 (UTC)

Improvement?

GliderMaven reverted this edit as not an improvement. The previous paragraph was, in my assessment, pretty garbled and nonspecific. I came to this new version by reviewing information in Asymmetric digital subscriber line and Orthogonal frequency-division multiplexing. ~Kvng (talk) 17:33, 26 September 2018 (UTC)

I haven't scrutinized the ADSL article, but a quick look did not find the OFDM reference. So I think it is useful to mention it here. I vote for the rev 861326221 version. --Bob K (talk) 08:18, 27 September 2018 (UTC)
ADSL uses discrete multitone modulation (DMT) which is an adaptive form of OFDM. My improvement to the substance here is that instead of saying it's "not binary", we say what it actually is. I had also made readability improvements. ~Kvng (talk) 14:48, 27 September 2018 (UTC)

Xiph.Org's Monty and the lollipop diagram

Please do watch the second video from https://www.xiph.org/video/. Come back and write a better wikipedia article. Alternatively one could translate the german article: de:Digitalsignal. User:ScotXWt@lk 16:43, 25 December 2016 (UTC)

This is a helpful source. I have added a link in the External links section. ~Kvng (talk) 15:51, 19 January 2020 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Digital_signal/Archive_1&oldid=1073106977"