Wikipedia:Peer review/Phase-shift keying/archive1

Phase-shift keying

I rewrote this a while back and finally got around to adding the extra stuff that was needed. We are sadly lacking any FAs on telecomms, and since this is a commonly used modulation scheme, I thought it might be a good place to start. With this lack of precedent, I'm particularly interested in whether there is enough/too much mathematics and whether the lead-in is thorough enough at the same time as not being dumbing-down. As well as other thoughts, of course. -Splash 23:40, 21 August 2005 (UTC)

Hi Splash,

Here are my suggestions for the article:

• Single quotation marks and italics are used throughout this article. Since they break the article's flow, try reviewing each use and consider whether it is really necessary.

• Working on this...

• Review the choice to make bold the phrases "binary phase-shift keying (BPSK)" and "quadrature phase-shift keying (QPSK)"

• They both redirect here, like DPSK.

• It might be worth minimizing the number of acronyms used in order to aid readability consider "The bit error rate (BER) of BPSK in AWGN can be calculated as:"

• This is true. I will look through and see where acronyms may be avoided. Sometimes, the phrasing would become repetitive with them included, but I will squeeze them out where I can. The particular case of AWGN is because I don't like referring to it simply as "white noise" — a statistically inadequate phrase.

• The obvious problem that stands out to the reader with differentially encoded PSK is an error in transmission seemingly won't be undone until there is another error in transmission. It might be worth briefly mentioning how communication systems overcome this difficulty in practice.

• It may be prudent to add something about duobinary encoding in the article you think? --HappyCamper 11:06, 22 August 2005 (UTC)

• Hmmmm. I'll go and read up on this a bit. We should take care not to conflate two distinct, but similar, techniques. -Splash 12:51, 22 August 2005 (UTC)

• The article should start by explaining phase shift keying not listing the three most common classes of digital modulation. Instead place this in the body of the article.
• Rephrase the first sentence so it avoids the use of brackets. Maybe "Phase-shift keying (PSK) changes, or modulates, the phase of a reference signal to convey data.".
• "necessarily" is unecessary.
• "comprises the 'symbol'" is bad grammar.
• Since DPSK redirects to this page it should be bolded.
• The sentence, "In exchange, however, its performance in terms of how many erroneous demodulations are made, is worse." has far too many redundant words.
• Sentences shouldn't begin with "which".
• The paragraph and sentence, "As for many digital modulation schemes, the constellation diagram is a useful representation and is relied upon in this article." goes nowhere.

• I've expanded this to give a compressed overview of constellation diagrams. The wikilink to the main article should do the rest.

• The equation for s1(t) is mathematically wrong. I think you forgot to drop pi from the last part.

• Yes. The basis-functions were added by another editor who I mentioned this peer-review to.

• "where 0 is represented by...and 1 is represented by...This assignment is, of course, arbitrary." might be redundant.

• Think so? It shows why the basis functions are a handy representation.

• Then I have no problem with leaving them in - it was just my opinion.

• "the I- and Q-channels" introduces new unexplained notation stick to talking about the in-phase and quadrature-phase components.
• It's a small point but instead of "The constellation diagram shows that" try "In the constellation diagram shown"
• Consider merging the last two sentences to make "This yields much lower amplitude fluctuations than non-offset QPSK and is often preferred in practice."

• Yes. They were originally so merged.

• Instead of "power of 2 (2,4,8,...)" try "power of two".
• Avoid the use of pronouns to address the reader (e.g. "we").
• Introduce the function Q(x) textually before you define it mathematically.
• Rather than defining Q(x) twice mathematically define it once in the form Q(x) = ... = ...
• The "just" in "it uses just two phases" is redundant.
• In "spacing around a circle which gives maximum phase-separation", "which" introduces a restrictive clause, try using "that".

• Made 'power' a wikilink instead.

• Instead of "which, since there is only one bit per symbol, is also the symbol error rate." try "Since there is only one bit per symbol, this is also the symbol error rate." Consider using a new sentence for all paragraphs that proceed equations maybe with the exception of those that just define variables.

• I'm not sure what your second sentence means exactly.

• You specify the basis function for BPSK but not QPSK. If you are going to discuss basis functions I believe that those of QPSK are equally deserving of mention. Especially since they might help with understanding in-phase and quadrature-phase components.

• These were only recently added to the article by me. I'll look into adding the QPSK ones if they seem useful. --HappyCamper 11:06, 22 August 2005 (UTC)
• We have a choice here (a)Leave the constellation diag as it is, and present the basis functions for QPSK - this is instructive since they take the general PSK form which BPSKs basis function does not; or (b)I can rotate the constellation diagram to be on the axes. Then it is readily apparent that QPSK is two BPSKs, and the basis functions' presentation may be more straightforward. -Splash 12:51, 22 August 2005 (UTC)

• There is no signal versus time diagram in the article. At least one would be appreciated.

• Added for most of the schemes now. Only differential-encoding remains wanting, but I will do it. -Splash 03:50, 27 August 2005 (UTC)

• I think an extra constellation diagram for π/4-QPSK might be helpful but may be difficult to fit in the article.

• Yeah, I'm still thinking about this. I could show the 'other' constellation with faint points or something.
• Done. -Splash 03:50, 27 August 2005 (UTC)

• The final sentence before the example is an important one. There are two things that could be done to improve it. First make it clear that "only a small increase in Eb / N0" is needed to overcome the increased BER. Second you might like to stress that the BER does not account for phase shifts in the communications channel caused by something other than AWGN and that in such channels the differentially encoded PSK may come out on top.

• This is a very good point. I've expanded the sentence into a short paragraph. What do you think? I decided not repeat the point in the final sentenc of the whole section; the one that deals with the graph.

• "k<supthth" is probably a typo.
• Try putting the footnotes in a section "Notes" and the other references in a section "References" in accordance with Wikipedia's other articles.

• I intend all the current entries as references to back up the Applications section. Which ones do you think should be footnotes?

• I really just meant that, by convention, most articles have a separate section for their Notes and their References. I've done this now. Some like the Nirvana article combine them and if you want to do this feel free.

• Ah, I see what you mean. Thanks. -Splash 12:51, 22 August 2005 (UTC)

Good work and I hope this helps.

Cedars 08:54, 22 August 2005 (UTC)

Wow, that's mighty thorough. I'm working my way through the list, so I've struck out the ones I've done. Along the way, I added some questions about some of your comments. -Splash 10:58, 22 August 2005 (UTC)

• looks good, but it's too technical. show a diagram of a single bit (e.g. a sine wave reference from -20 to +20, a signal starting as a sine wave then turning into a cosine from about -5 to +10 and then returning to the reference sine wave from +10 onwards). I've had some success generating this with gnuplot and the erf function. If you don't have an easy way to generate these diagrams, message me.
  • That bit done. -Splash 03:50, 27 August 2005 (UTC)
• I think you want to move discussion of other modulation techniques out of the article line, possibly to a navigation template.
• move the applications section quite early in the article and include pictures
• possibly split some of the maths out, particularly error definitions. Remember Wikipedia isn't paper so the article doesn't have to be complete in its self.
• ${\sqrt {E_{b}}}\phi (t)$ for example might be better represented by a warm fuzzy picture.
• I find the statement "Analysis shows that" a bit strange for "this may be used either to double the data ..." since it's in some sense just a clear true statement / follows directly from the definitions of these concepts. I think just dropthe phrase.
• try to make it so that you could know all of the basics of PSK without passing an equation. They should all stay in, but later and/or in more detailed articles.

Mozzerati 21:50, August 22, 2005 (UTC)

Thanks. I'll work my way through the review up above first, since I'm already part way through dealing with it. But a discussion of some of your points is important now too. In the same order:

• I will make a timing diagram, which lays a bitstream against a signal. I'm going to do this probably only for QPSK since it will also cover BPSK. The BPSK section can just point to "below".
• Yes, a navbox. I've been thinking about making one of those. I'll do it if people want one, although the linear series between the three classes isn't very clear (what order would they go in?). My inclination is (as it was originally) simply not to mention the others, but rely on the wikilink to modulation to do the job. This article isn't about modulation or communications in general, after all, its about one particular shift-keying.
• The various points about mathematics. I'm generally not keen to strip this down to words and pictures. The topic is more than that and, whilst we're not-paper, we are an encyclopedia. The coverage should not shy away from the more difficult parts of things. Spinning the maths out into other articles is something I considered, but decided I didn't like too much. This is because I would have to reproduce a large part of the contents of this article in each subarticle, and the only different material would be a few lines of math. Holding them in here I don't think does much harm — the non-technical reader can either use the rest of the Wikipedia to learn the necessary terms or just look at the graphs.
• To the particlar point about basis functions. They are an important aspect of all digital modulation schemes, and I wonder what warm fuzzy picture you have in mind? A picture of a cos wave?

• Ah, click! You mean a block diagram of the modulator. Yes, I agree. That would be an excellent idea and would show the meaning of the mathematics nicely. I shall make one! Again, I'll make it for the QPSK system since that absorbs BPSK and avoids repetition; perhaps it should just go straight in the Higher-order PSK section?

• I'll have to think about the question about bandwidth. This is obvious to the technical reader, but is it obvious to the non-tehcnical? Is it intuitive to say "twice the data doesn't need any more bandwidth" is the same thing as "keep the data the same but halve the bandwidth"?
• As laid out at present, you can cover the basics of each type of PSK without passing an equation. I'll think about an alternate layout that presents the mathematics in a seperate section or two, but will have to see if that leaves the non-maths parts at a couple of sentences each. That might be no bad thing, though. -Splash 23:18, 22 August 2005 (UTC)

Nice to know that telecom articles are being improved but the article is too complicated for a casual reader. I would like to see an overhaul in the entire structure of the article. Here are my suggestions:

• The lead is very complicated, it mentions terms that a casual reader won't understand. I would like it 'dummed down' and translated into simple English. Instead of saying phase of a reference signal (the carrier wave) you can use the term 'sinusoidal wave' instead. BPSK is the most simple to understand, so the BPSK concept should be introduced in the lead. In its most simple form a regular sinewave is used to represent binary 0; and a 'cosine' wave (180° phase shift) is used to represent binary 1. This should be supported by a diagram alongside. (You can clip the QPSK image, first line-0110 to illustrate this). This example is the most simple illustration, and a first time user should the concept through this example.
• Now in the =introduction= mention the need for modulation techniques, and ASK, FSK. Explain about NRZ signals which are used.
• Now create a new section =types of= . Under this include BPSK, DPSK, DEPSK, QPSK, OQPSK, M-ary PSK, MSK etc.
  • Under each of these headings briefly mentioned the mathematical representation, graphical representation (waveform), basic information, and advt+disadvt.
  • Remove all details such as constellation diagrams, error probability etc. Push them into dedicated articles where the generation receiving, and probability can all be discussed. =Nichalp «Talk»= 09:12, August 23, 2005 (UTC)

Thank you for being very specific with your suggestions. My concern with the route you suggest is that it will result in a series of articles which have to repeat a lot of the same background material (e.g. what is modulation, what is PSK, etc). These articles would also each be rather short, and would have to import a lot of context from elsewhere: the error-analysis for example would be incomplete without the pro/con discussion that would also be necessary in the main article. I suppose seeking an FA is less important than having good coverage, but would these short articles stand any serious chance at being an FA? A rearrangement of the content as suggested by the second reviewer above would keep it all in context, but move the mathematical features etc. to later on in the article: people can just stop reading when they've had enough. -Splash 21:23, 23 August 2005 (UTC)

• You wouldn't have to define what is modulation in each section. The lead is a summary of the article. It should be written last, and it is perfectly ok to have repetition here. So for the moment ignore the lead.
• The intro will have more information on the basics of the topics. NRZ etc.
• Under each section just explain the generation on the topic, such as Sq. Law Dev, Band Pass Filter etc., and why they are used. You can cover the entire topic (of say QPSK) in four paragraphs. Now 4x7=35+3 (give or take a few), which is of acceptable length. Short articles are allowed to be FA worthy, if anyone objects on the grounds "too short"; they have to provide expansion suggestions, else it is deemed inactionably in FAC.
• I don't support the derivation of error analysis on this page. It becomes too technical. For example you can say that the disadvt. of a DEPSK system is that 1) a complex demodulator is needed; and 2) errors occur in pairs because of the XOR operation. More than enough.
• If the article is written for dummies, I see no problem why this shouldn't be featured.

=Nichalp «Talk»= 05:53, August 24, 2005 (UTC)

• Ok, I will have to think about this. I'll add mentions of BP filters and NRZ, but we already have Band-pass filter (and they're not really part of the modulation consideration) and NRZ, and this article isn't about modulation or digital communications in general, so I don't think we need the full exposition of each component in an end-to-end system. We don't appear to have square law detector, but again it only needs a mention and a brief explanation — it's generic material that should be in a generic article.
• I must confess to be somewhat disappointed that the only way this could become featured is to write it for dummies. Is that what a featured article is about? The article before I rewrote it was suitable for dummies, and was a disgrace. -Splash 00:22, 25 August 2005 (UTC)

No, that's not what FA is all about. You can understand these technical terms, I can too, and so can 100-odd wikipedians. What about the thousands of readers who come to read about article content? Do we want to drive them away by using 'technical terms' that are not explained here? For example look at the cricket page. What we've done is to initally avoid the useage of cricketing terminology so that first time users won't be put off. We've gradually built up the terms by defining them, and using them in the article. It isn't crafted for dummies, but at the same time it hasn't lost its essence or quality.

No, I'm not asking for a fullblown account of the BPF etc. A line on what is does is enough instead of clicking the terms.

I want to see this through FA. I hope you are as optimistic as me. I'd like to help u out. However, I would prefer you structure the content as I've suggested before, before I move in and simplify the terms. =Nichalp «Talk»= 06:40, August 25, 2005 (UTC)

Ok, well, I (and HappyCamper) will finish adding the stuff that's been suggested and then any refactoring/etc can happen. I was short on time yesterday, which was why not much happened to the article. -Splash 10:52, 25 August 2005 (UTC)

• 1) I'd have to echo Nichalps suggestion that the article be made more accessible to those that don't already know the subject. That doesn't mean dumbing the whole thing down, but does require careful planning and the adding of context to explain terms/jargon. It may require moving highly detailed material off to sub-articles in order to allow the article to be properly balanced. A careful attempt is needed to make sure the article gives proper amount of coverage to each topic in relation to its importance to the overall topic. That's basically the idea behind good Wikipedia:Summary style. As for accessibility, the majority of the lead should be understandable to the well educated person that doesn't know the subject at all. That is possible to do and still be perfectly accurate if you properly add context. Yes that can seem a little redundant, but doesn't have to if done right. The same for the introduction to each major subtopic. After that, proceed into as much of the detail as you need to. That way you get all the needed detail in, it's not dumbed down, but people who don't already know the material can at least dig into it if they want to. 2) To a higher level view the whole article seems to spend the most time explaining how each works and very little on where it is applied and how important/widespread its use is. The applications section is it as far as I can tell, but since the explanation of DQPSK, etc are split between the lead and the intro, I have to refer back and forth to those just to try to grok the applications section. The lead basically covers nothing but how the technology works, while it should give a broad overview of the subject. - Taxman ^Talk 19:57, September 6, 2005 (UTC)
  • Thanks, Taxman. Well, I think the refactoring is 'all' we have left to do, along with a general copyedit to minmise acronyms and things. We're missing an image of the receiver structure, but HappyCamper tells me he's going to make that. So I'll go away and think carefully about what you've said. To the applications section in particular: yes, having more of it in the lead is a good idea. PSK is used so widely that listing all its usages would be an article of its own probably; it started with the NASA space program in the '60s I believe. Originally, and imo more helpfully, the applications section at the bottom so you only got to it once you knew what/where the PSKs are. The other reviewers have suggested moving it upward, and an anon already did it hence its current location. Anyway, I've been taking an extended break from the article but will come back to it now. Thanks for your suggestions. -Splash 00:13, 8 September 2005 (UTC)