|WikiProject Radio||(Rated C-class, Mid-importance)|
- 1 History
- 2 Definition of terms, clarity
- 3 Zenith/GE system Pilot tone system
- 4 Basic concepts
- 5 Wide and narrow
- 6 Why 87-108
- 7 Why "Non-Broadcast Uses?"
- 8 Wrong technical information about FM Stereo
- 9 Conferences about FM
- 10 Commercial References
- 11 Um, what?
- 12 Addition: FM Audio Quality comparison
- 13 Removal: Oscar Bonello research promotion
- 14 Restore the Oscar Bonello work
- 15 Why did you delete?
- 16 Target of FM Redirect
- 17 It is the name of a bar
- 18 Audio performance
- 19 Bad Information in Pre-emphasis and de-emphasis?
- 20 This is due to the need to lower the modulation index of the main (sum) signal to accommodate the presence of the 38 kHz DSB-SC (double side-band suppressed-carrier) subcarrier and 19 kHz pilot tone.
This article used to have a lot of US-specific material. To fix that, the original article was moved (along with its history) to FM broadcasting in the USA, then the non-US specific material was restored here. Therefore much of the editing history is now at the new article. --Tony Sidaway|Talk 23:32, 14 Mar 2005 (UTC)
Definition of terms, clarity
I've linked to a number of terms (mostly used in the first paragraph after the introduction) which are technical jargon and undefined in the article. Some of those links probably don't go where they should.
I know a fair amount about radio, certainly well beyond what the average person knows, and I find major portions of this article fairly incomprehensible. It was obviously written by someone who knows the subject and is used to communicating with engineers, not with nonspecialists. Because of this and the undefined terms, I nearly slapped a "clarity" tag on the page. But maybe I'm just having a bad day. In any case, I suspect this article reads like a lot of the mathematics articles here - mostly obscure to those outside the high priesthood. --126.96.36.199 15:00, 6 January 2006 (UTC)
The article really seems as if it starts half-way through - as if there should be a simple "How FM broadcasting works" bit at the beginning, then all the technical jargon. Perhaps someone who understands the intricacies could work on 'FM 101'? Cammy 11:05, 10 January 2006 (UTC)
- That is covered by more fundamental articles on radio, broadcasting and frequency modulation. This article should not duplicate what is already written in those elementary articles, but instead concentrate on the specifics of FM as it is applied to broadcasting. Harumphy 12:14, 10 January 2006 (UTC)
Zenith/GE system Pilot tone system
Can anyone explain why 19 KHz was chosen as the pilot tone frequency for FM stereo and why it is necessary to have a such a wide (4KHz) guardband around the pilot tone. Surely a higher pilot tone frequency and/or a tighter guardband spec could be used to improve the audio bandwith beyond the 15KHz limit ?
Also I have heard there were/are rival systems to the Zenith/GE system. Does anyone know how these worked when/where the were/are used and how well they worked ?
- A higher pilot tone frequency would imply a higher stereo subcarrier frequency, which would make it more vulnerable to noise. A tighter guardband spec would require steeper filter slopes, adding to the cost and complexity of receivers. When it was developed, this system pushed the available technology to the limit. Harumphy 09:14, 23 July 2006 (UTC)
The math for FM is complicated, but pretty much there is so little above 15kHz, and except for babies we can't hear it, anyway. It is usual to build amplifiers that do 20Hz to 20kHz as convenient numbers, and to make it more likely that they are good up to 15kHz. Another factor might have been SCA that was already at 67kHz, so it has to fit below that. It seems that there was also a 41kHz SCA, the presumably got pushed out by FM. Also, note that stereo TV sounds uses the same modulation with a 15.73427k pilot. Gah4 (talk) 01:16, 6 December 2015 (UTC)
Basic stuff about frequency modulation, that is not specific to FM broadcasting, surely belongs in the article about frequency modulation. It should not be duplicated here. This article should describe FM as applied to broadcasting, and not FM in general, IMHO. Therefore basic stuff about wide versus narrow FM, Carson's rule, etc., belongs there. Harumphy 15:49, 23 July 2006 (UTC)
Wide and narrow
The wide versus narrow thing is too simplistic, because in practice there is a wide variety of systems using a whole range of parameters which do not fit into narrow and wide camps. Where is the boundary between narrow and wide? How many kHz? Is an outside broadcast link, using 10 kHz peak deviation in a 50 kHz channel assignment to carry about 12 kHz of audio bandwidth wide or narrow? It may make more sense to talk of the modulation index, but again this belongs in the general FM article, not the FM broadcasting one. Harumphy 15:49, 23 July 2006 (UTC)
Why is 87-108 (roughly) used for civilian FM broadcast? Who chose these frequencies? --188.8.131.52 01:53, 6 August 2006 (UTC)
- The radio spectrum is allocated to its various uses by international agreement, under the auspices of the International Telecommunications Union (a UN agency). These agreements are reached at a serious of conferences held (typically) every few years. Harumphy 08:13, 6 August 2006 (UTC).
Why "Non-Broadcast Uses?"
This article is focused on "FM Broadcasting", and, largely the technical details of the science. I propose that the section entitled "non-broadcast uses" be moved to its own article or deleted. In truth, there are many other uses of frequency-modulated RF carriers than broadcasting. Wireless two-way communication, amateur radio, and occasionally data communications (though FM is an inefficient use of spectrum for packet-type radio). It may be of more use to have a section under "[Frequency Modulation]" on "application" rather than a "non-broadcast" section under the "broadcasting" article.--Gdickinson 02:08, 12 August 2006 (UTC)
- I've changed the title to something a little less contradictory. IMHO the section should be kept because the uses listed are within the allocated FM band, using FM broadcast technology, and therefore are relevant to the article. By and large these uses are just extended uses of the standard band, generally one-way communications that would not necessarily be practical otherwise. It is true that a great deal of FM is used in two-way communications, but that doesn't mean that it belongs in this article. I am arguing that microtransmitters, however, do. Haikupoet 05:44, 12 August 2006 (UTC)
Wrong technical information about FM Stereo
The article states:
Stereo FM signals are far more susceptible to noise and multipath distortion than mono FM signals. This is due to several factors, including the following:
the addition of the two sidebands of the difference subcarrier to the baseband signal increases the noise bandwidth of the signal by a factor of three (9.5 dB) as compared with a mono signal. as mentioned above, the pre-emphasis is applied to the audio signals before encoding. This results in the pre-emphasis acting in the wrong direction on the lower sideband of the difference subcarrier, i.e. decreasing the level as the frequency rises, which will have a further deleterious effect on the S/N of the difference signal.
- This is quite obviously completely incorrect.
- The pilot tone is 10 percent, leaving 90 percent available for information. That is a -0.915150 dB reduction in available information (signal) over monophonic operation. Further, on the average, fifty percent of this information is in the subchannel, that's an additional -6.020600 dB reduction in information. The sum of these two shows the reduction of the available S/N over a mono signal is -6.935750 dB.
- Then, the argument of the 23 kHz to 53 kHz information having pre-emphasis in the wrong direction is further incorrect. This information in the subchannel cannot be broken up in such an illogical way. The subchannel contains a pair of sidebands, with the originating carrier suppressed at least 40 dB. This L-R information exists in both sidebands. They exists in a phase-linear, flat-frequency-response portion of the supersonic audio spectrum. Guessing that baseband preemphasis somehow affected this is not supported by mathematics or information theory. At one time in the United States, there were three FM Stereo generators available, that compiled with the FCC Rules and Regulations. All three were designed by me. The January, 1974 Broadcast Engineering magazine, on page 25 starts a three-page article that I wrote as an expert. --LymanSchool 03:41, 28 December 2006 (UTC)
- It probably isn't all that helpful to argue about the S/N disadvantage of stereo signals, because there are so many variables that would need to be defined first. First, are we talking about S/N in the M channel, the S channel, or each or the L & R channels? Is the pre-emphasis 50 or 75 uS? Are we talking about unweighted or weighted noise? If the latter, A-weighted RMS, ITU-R 468 or something else? I suggest we should identify the sources of stereo noise without quantifying them. Harumphy 12:19, 28 December 2006 (UTC)
Totally agree that the "wrong direction pre-emphasis" on the lower sideband of the stereo subcarrier is incorrect. The stereo sub-carrier must be demodulated BEFORE de-emphasis is applied, so by the time the audio reaches the de-emphasis network the "upside down" lower sideband has already been turned back upright, so to speak. Timothy Stockman 02:53, 5 October 2007 (UTC)
I should add that random channel noise is an equal-energy-per-unit-bandwidth phenomenon, and the transmission channel will add the same amount of noise whether the audio is "right side up" (mono and upper sideband of stereo subcarrier) or "upside down" (lower sideband of stereo subcarrier. The stereo signal simply occupies three times more bandwidth than a mono signal; three units of bandwidth will contain more random noise than one unit of bandwidth. The original article makes the assumption that somehow units of bandwidth at higher frequencies have more random noise than those at lower frequencies; this is untrue.
Should I remove this incorrect information from the article? Timothy Stockman 03:08, 5 October 2007 (UTC)
- I believe the idea is that adding noise to the FM signal has an unusual relationship to the sidebands, and so to the decoded signal. But there is another way to look at the FM stereo signal, which shows that there is no such change. You can consider it as alternating between L and R at 38kHz. Added noise will add equally to the two channels, and deemphasis will then be applied. Gah4 (talk) 01:24, 6 December 2015 (UTC)
Conferences about FM
I added a small section about ITU conferences; don't you think we should expand that section? I can do it, if we think we could do it. [Hamlet 26 Mar 2007]
Would it be out of line to mention some commercial things under Small-scale use of the FM broadcast band? Specifically, Mr. Microphone under FM radio microphones and the Ramsey series of FM transmitter kits either under Microbroadcasting or pirate radio? I think they are both notable, but the Ramsey link in particular might be against WP:EL or WP:SPAM.--Mdwyer 20:27, 26 April 2007 (UTC)
"A new device named IM cancelled high frequency clipper is able to give a heavy audio clipping at high audio frequencies with low ear annoying."
English, please! --184.108.40.206 11:14, 29 July 2007 (UTC)
Addition: FM Audio Quality comparison
IMHO it would be worthwhile to add an extra paragraph comparing FM quality (example: 50-16000 Hz, 60 db S/N) to AM quality (50-10000 Hz) and CD quality (20-20000 Hz) audio. I actually came here hoping to find that information, and was surprised to see it's not listed, since the opening paragraph labels FM as "hi-fidelity". Is it? I don't see any numbers to back it up. Thanks. - Theaveng 16:48, 27 September 2007 (UTC)
- By itself, there's nothing to say that FM is higher or lower fidelity than AM, SSB, or what-have-you. It all depends on the S/N ratio and bandwidth of the channel (so, for FM, the modulation index). While it's true that broadcast FM has lots more baseband bandwidth and fidelity than broadcast medium wave AM, that's more a function of the channel bandwidth consumed than the fact that the modulation is FM. Perhaps no one has been willing to write the explanation yet. Feel bold? - Atlant 23:32, 27 September 2007 (UTC)
- This article is not about "FM" in general, but specifically about the FM Broadcasting technology, and thus can be stated to have a certain measurable sound quality for the ~200 kilohertz wide channels used in Europe and the U.S., versus the ~10 kilohertz wide channels for E.U. or U.S. AM radio. - Theaveng 12:22, 21 October 2007 (UTC)
Removal: Oscar Bonello research promotion
Removed clearly unencyclopaedic references to research by Oscar Bonello that don't enhance the overall article. He may well have important work, but unless it is referenced by wider works on radio, it's promotion here is inappropriate. Martyvis (talk) 10:11, 11 March 2008 (UTC)
Restore the Oscar Bonello work
The particular point of view of Mr Martyvis is dangerous for the Science. If every time a researcher shows the results of a published investigation ( accepted for the reviewers of the Journal) we say "Discard it because Newton is promoting the Gravity Laws" probably the Science Books will be only blank pages... As Mr Theaveng said, above this lines; " This article is not about "FM" in general, but specifically about the FM Broadcasting technology" Then, there are two main questions to solve 1) How we can use pre-enphasis with music of high sound density above 5 KHz (like a CD recording) instead of the sound material used by Armnstrong in 1933 ? We indicate where to find this answer 2) Yet more complex (and frequently asked by my pupils at University) Why a signal that has ever the same power (in FM the amplitude remains the same) is able to get longer transmission distances when the audio is processed ? This is not a minor point because the FM commercial sucess is vinculated to the coverage area. For the first time the JAES march 2007 paper solves this aparent paradox using Mathematics. I feel that both contributions are pertinent (complements the knolewdge about FM Broadcast) and are scientifically founded and accepted for the scientific community Otherwise you can read my curriculum to see that at my 68 years old I do not need "promotion" like a young film star... Oscar Bonello (email@example.com ) —Preceding unsigned comment added by OscarJuan (talk • contribs) 23:34, 15 March 2008 (UTC)
Clearly you have a NPOVon this then? I am not a radio technical expert my any means. I went to the FM broadcasting article to find an answer about audio bandwidth. The 2 reference sto Oscar Bonello's work stuck out to me like a sore thumb. I assume there are 100s of researchers and 1000s of papers that have written around the usefulness of various techniques. But unless they have substantially changed the way FM broadcasting actually I operates, I fail to see why they should be included. What would this article (or any other technical article on WP) read like if every research put in a snippet of how they think the current state of play would be approved? Thist article should be about what FM broadcasting is - not what it might be if people just listen to me (as an author)! This content is not appropriate and and if you think about it, you should remove this. Martyvis (talk) 07:22, 17 March 2008 (UTC)
The article says Multiband audio processing improves the coverage area of an FM stereo station, however, this knowledge was never rigorously demonstrated prior to the work of Oscar Bonello. with a reference to a 2007 paper. I must admit I know nothing about this subject, but it looks a bit strange that such a basic fact about such an old technology was never considered in a scientific paper before. --The very model of a minor general (talk) 13:52, 3 May 2008 (UTC)
- By 'rigorously demonstrated', he appears to mean mathematically proven. Big deal. There has been ample empirical evidence for around 30 years, and the sudden arrival of a mathematical proof for what has been obvious, intuitive common knowledge amongst broadcast engineers for decades is neither important nor even interesting, except perhaps to other mathematicians, the paper's author and his mum. There's no consensus for these comments' inclusion and only one editor who wants them in, so I'm going to remove them. Please do not reinstate them until and unless there's a consensus here to do so. --Harumphy (talk) 16:07, 3 May 2008 (UTC)
Prof Bonello opinion
Is unfortunately for the Wikipedia future to know that:
- " I must admit I know nothing about this subject, but ... "
- " I am not a radio technical expert my any means..."
- ... and the sudden arrival of a mathematical proof for what has been obvious, intuitive common knowledge amongst broadcast engineers for decades is neither important ... "
Usually when a person recognizes itself as ignorant of some question he do not writes a single line at any paper cyclopedia (only the best people in each field writes for the Encyclopedia Britannica by example). Usually the people that do not know about an issue the best he can do is to learn. After 30 year teaching at University I used to do the job of converting people to the full knowledge... But what can we do with you my dear friends? Only to say:
1) About "Clearly you have a NPOVon this then?" Please read the Wiki rules. A person can write about his own work if this is published by a recognized scientific journal
2) About: "it looks a bit strange that such a basic fact about such an old technology was never considered in a scientific paper before. " Please give me examples of previous publications. I let you know that before publication of a Paper, the most brilliant engineers of the AES Journal (New York, USA) revised all international literature in English, French, German, Italian, Spanish, Japanese and Russian. If you know a published proof previous to mine, please write it in this same page.
3) About "the sudden arrival of a mathematical proof for what has been obvious" Please note that in Science nothing is obvious. Example: It appears to be obvious during centuries that the Earth is flat, not round. Remember the Galileo trial ? Do you know about the last Fermat theorem? Do you know about the Tales theorem? About specifically the FM statements. Can you explain me what is the reason that the audio processing improves the coverage area if the power of an FM modulated signal is the same not matter the modulation percentage ? Please explain me it !
- Robert Orban has already explained it better than I can, many times over many years. I read his papers avidly over 20 years ago when I worked as a broadcast engineer. Are you not familiar with his work? BTW, if you restore your work without consensus here first you risk being blocked for edit-warring.--Harumphy (talk) 08:20, 23 May 2008 (UTC)
4) My paper not only gives a proof of the coverage area improvement. It gives for first time the way to calculate it. And give the way to improve audio processing
Since a Cyclopedia can not be based on the ignorance I will restore my article (to avoid the ignorance) if you do not give me the reply to my questions
- side note, not related to this discussion: Galileo trial was about heliocentrism vs geocentrism, while the "flat earth" seems to be mostly a recent, modern myth. --Gabriel Bouvigne (talk) 08:11, 23 May 2008 (UTC)
OscarJuan reply: Yes, Gabriel, your are right with Galileo, of course. I cited him in support of the idea: "in Science nothing is obvious" (the earth running fast around the sun was not an obvious idea) Thank you for not said the same of Fermat or Tales theorem (nothing to do with the flat Earth...). Thank you for follow me around the Wikipedia... Maybe we can be good friends at the future.
About the Harumphy contribution. Yes I know the valuable work of Orban. I know him personally. But there are different approaches in this issue. He proposes a loudness explanation of the way the distance is improved; that is only partially correct. This explanation was enough for most of the broadcasting people, but never was fully accepted by the Communication Engineers community (please read the FM power paradox I stated above). In your own words, Harumphy: "the sudden arrival of a mathematical proof for what has been obvious, intuitive common knowledge amongst broadcast engineers for decades is neither important “ Some broadcasting engineers usually think as you do... But this is not the way I think If you read my article, you will find a lot of interesting Psychoacoustic explanations that probably will improve this technology in the forthcoming years (as Mr Orban recognized me about the wide use of the Kahn-Bonello phase symmetry networks that I proposed in a JAES paper 25 years ago...) Since the FM BROADCASTING article has only one reference (it is very poor), I suggest other reviewers to improve the quality of information with references to AES papers and books. Probably I do not add nothing to FM BROADCASTING because I feel a sensation of "all is perfect here, please do not touch it" Regards OscarJuan —Preceding unsigned comment added by OscarJuan (talk • contribs) 01:35, 24 May 2008 (UTC)
- As I see it, the problem is rather different. I don't think anyone claims the article is perfect. The article has, to date, barely touched upon the whole subject of FM TX processing at all - even in broad overview. It really needs to have that first before getting into details about a highly specialised aspect of it, especially if the latter involves self-publicity. --Harumphy (talk) 08:10, 24 May 2008 (UTC)
Why did you delete?
In the past I just added the algebric forumala with which the Matrix DECODER acts. I ask myself why you deleted it. It's not an industrial secret and it's even very known. Simply it was not written here in Wikipedia. I worked in Broadcasting for a decade and I was member of S.M.P.T.E. What I written, it' sjust the matrix Law that normally governs ( or governed in the past ) de decoding of a Stereo Signal. What's wrong with it ? Who's the "teacher removed it" and why ?
I repeat HERE that Law:
BEFORE the multiplex era, it was transmitetd the SUM ( L+R ) and the difference ( L-R ) as written in the article. The operation needed in order to give back L and R it was ( and it's even quite intuible, come-on! ):
Of course to obtain L and R now
Now, why this information was deleted ? What's wrong ?
- I assume you mean this edit of mine: . If so, my aim was to simplify the explanation by cutting it down to its bare essentials. There's nothing 'wrong' with the matrix formulas, they just seemed, for an encyclopedic overview of the subject, an unnecessary repetition of information already given.--Harumphy (talk) 09:23, 21 August 2010 (UTC)
- yes it was exactly what you linked.I assumed that to give a clear formula, without diminishing the value of the explanation, was just better, in order to give an immediate idea. If you have a look in "Dolby" chapter, there are also forumlae. Just fro mthe latin "Formula" means: the shortest form. And I find it extremely efficient, where words are creating some confusion, sometimes, due the differences in languanges. Being Wikipedia an Enciclopedia that gives very deep details on arguments treaten, I thought it was useful to write clearly down even this formula once forever. Am I so wrong ?Iw2mln (talk) 20:17, 22 August 2010 (UTC)
- The terms 'sum' and 'difference' are already defined in the article as meaning L+R and L-R, so I'm not sure where the confusion would arise. I agree that sometimes maths is the ideal language, but I don't think this is such a case: the concept we're describing is simple enough for prose. Clearly we have a different view and probably are not going to convince each other, so what does anyone else think?--Harumphy (talk) 22:40, 22 August 2010 (UTC)
It is the name of a bar
The following statement appears in Comparison of analog and digital recording:
- Analog FM broadcasts rarely have a dynamic range exceeding 50 dB, though under excellent reception conditions the basic FM transmission system can achieve just over 80dB.
I don't find any support for that in this article or any discussion of audio performance. It would be a welcome addition here. -—Kvng 16:11, 7 December 2012 (UTC)
- Sounds believable, though. About 50dB was usual for vinyl records. If you get much more, it gets lost in the background noise around the listener, especially in a car. Even if the source has more, it is likely compressed before broadcasting. Rock music starts out with low dynamic range, so maybe isn't changed. See Loudness war. Gah4 (talk) 01:02, 6 December 2015 (UTC)
Bad Information in Pre-emphasis and de-emphasis?
Near the very end of this section it is stated that the BTSC TV sound system utilizes a variable pre-emphasis. This is not correct. I have the BTSC specification and it clearly states in section B.c.2.i, "Fixed pre-emphasis..." (referring to the difference channel.). The sum and difference channels must be encoded the same way, thus both are fixed preemphasis. Yates (talk) 15:09, 30 April 2015 (UTC)
This is due to the need to lower the modulation index of the main (sum) signal to accommodate the presence of the 38 kHz DSB-SC (double side-band suppressed-carrier) subcarrier and 19 kHz pilot tone.
Note that just adding the DSB-SC doesn't increase the modulation index. The sum is never more than the maximum of the left and right channel inputs. It is reduced to allow for the 10% of the pilot, which does add. I suspect that the bigger problem is not enough signal to recover the pilot properly. Gah4 (talk) 00:57, 6 December 2015 (UTC)