Talk:Pulse-code modulation/Archive 1

First message
How is the information encoded digitally? — Preceding unsigned comment added by 195.186.148.xxx (talk) 23:35, 28 November 2001 (UTC)

More recent history about PCM and applications
This article has a lot of info about early invention and development of PCM (Alec Reeves, 1937, Bell Labs work during WW2 and later), but has no information since then. I urge the writers to add information on the development of the D1/T1 system by Bell Labs that entered commercial service in 1962, and the subsequent development of E1/T1 in Europe that became the mainstay of telecom communications for voice worldwide by the 1980s. In addition there is no mention that PCM is the underlying technology for Compact Disc audio, without which CDs would not have been possible.

Thanks. Tony (talk) 15:17, 1 September 2010 (UTC)

Potential Copyvio (it wasn't)
Concerning the possible copyright violations noted by User:Muriel Gottrop on 11:57, 19 Oct 2003 (UTC). I have cowritten much of this article, it has partial histories at Pulse-code modulation and Pulse code modulation.

The history section was directly copied from: http://www.derivaz.fsnet.co.uk/ahr/pcm.htm

The document at http://ccc.inaoep.mx/~cferegrino/cursos/comprcrip/Lossy_Methods.pdf seems to have copied the article literally from Wikipedia without a proper attibution or an acknowledgement of the license under which Wikipedia is distributed. -- Ap 14:21, 19 Oct 2003 (UTC)


 * So is reversed copy violations. I'm glad. :) Muriel Gottrop

PCM also stands for phase change material, disambiguation needed?. Thadk 00:12, 2004 May 14 (UTC)

Number format
Does PCM imply or mandate anything about the number format? Fixed vs floating point? u-law vs linear quantization levels? — Omegatron 19:49, 22 November 2005 (UTC)

This article badly needs an illustration or two
Twang 00:58, 11 March 2006 (UTC)
 * Well, it has one now. If I have time maybew I'll make a DPCM one too.


 * Shouldn't the current image go beside the input section since it starts with "In the diagram to the right" Adam Bryzak 10:53, 8 October 2006 (UTC)


 * Also, the image does not show a discrete time samples but continous time stairsteps. This is a common source of confusion for people. The usual and correct way to illustrate discrete time sampling is with pulses (perhaps I should just make a new image?). IMO there is not enbough emphasis on the *pulse* nature of PCM i the article. AL - 15:45 4 March October 2007 (UTC)
 * The image currently in the article matches the text I have on hand (Stallings). I do, however think you may have a point as this image illustrates the result of decoding PCM, not the process of encoding it. Perhaps a larger image with 3 panels (source, pcm, output) could be created and added to the article? --Ktims 04:19, 7 March 2007 (UTC)

There should definitely be a diagram which shows impulses rather than the "stair step". Every time someone sees this "stair step" picture they tend to think that in some sense the PCM data represents a square wave, which it doesn't, and that tends to screw up their understanding of everything else. So let's get a diagram with impulses (ie vertical lines, spaced equally apart) instead of stair steps. 82.69.171.90 12:59, 30 July 2007 (UTC)

Bell Labs also invented PCM
While I understand that Reeves is widely known as the inventor of PCM, it was also invented independently at Bell Labs as part of X System (aka SIGSALY). An important aspect of the work at Bell Labs is that it was implemented and used in securing Allied communications in WWII. 

Because this information was classified for 30 years it has naturally not become as widely known as Reeves' work. Miller is probably the only remaining member of the team at Bell Labs and some of his knowledge needs to be included in this article and the one on SIGSALY.

For external confirmation of Miller's role at Bell Labs, please refer to "National Service in War and Peace (1925-1975), Volume II of A History of Engineering and Science in the Bell System (Bell Telephone Laboratories, 1978)" p. 297. Table 5.5 mentions R. L. Miller as inventor on four of the patents related to the X System.


 * It would be of interest to list the patent number, grant date etc of Miller's US patent. I am familiar with the sigsaly system and I admire it very much. It is a matter of debate to say that sigsaly is pcm (I think it is a digital compressed vocoder, and not pcm). If you submit the patent number I will take a look at the patent, and will try to find out what it means. Unfortunately, I do not have a copy of the work you referenced. By the way, according to the US Inventors Hall of Fame,, normally a relaible source, Claude Shannon is the inventor of PCM as granted in 'Communication System Employing Pulse Code Modulation,' US Patent Number 2,801,281. It was granted in 1956 (20 years after Reeves PCM patent was granted). Sorn67 14:47, 9 July 2006 (UTC)

It was not until about the middle of 1943 that Bell Labs people became aware of the use of binary coding as already proposed by Alec H. Reeves, of The International Telephone and Telegraph Co. Reeves had filed for a French patent in 1938, but his U.S. patent had not been issued until late 1942. In describing his invention, Reeves later wrote, "Having had it patented, for understandable reasons I then let the invention slip from my mind until the end of the war. It was in the United States during World War II that the next step in PCM's progress was made, by the Bell Telephone Laboratories." Reeves was referring in particular to work carried out by H. S. Black and W. M. Goodall during and after 1943 (a part of the work being for the U.S. government) and described after the war by the first technical publications on PCM.
 * I did some searching and found on Miller's home page. From it, I quote the following relevant text:

This emphasis on binary PCM, which was mainly pointed toward wideband circuits, did make the personnel working on the X System take another serious look at the coding arrangement to be sure that something had not been overlooked. H. L. Barney carried out a series of extensive tests with various possible combinations but came to the conclusion that for the situation it faced, the multilevel arrangement used was better.

I conclude that the PCM by Reeves was strictly binary, and sigsaly was a multi-channel vocoder with 6 levels of quantization. Sorn67 15:37, 9 July 2006 (UTC)

That's consistent with my understanding. Bell Labs' patent was for "N-ary" not just binary. I'm trying to get more info about the patents. --Dobbse 21:17, 10 July 2006 (UTC)

I'm working from the same source as you mention above. There's a disclaimer in that PDF about possible errors due to limits of OCR. This table is subject to that same disclaimer and I'll see what I can do to verify the patent numbers and dates for accuracy. --Dobbse 21:54, 10 July 2006 (UTC)

Nyquist limit
I just edited a typo in the introduction, but I'm not sure that what's being said is accurate. It's said that "the frequency range of the analog signal is higher than half the sampling rate". This sounds like a reference to the Nyquist limit, but the picture shows a signal with a much lower frequency than the sampling frequency.

Stuarta 10:30, 13 July 2006 (UTC)

The wording in the limitations section about the sampling theorem and Nyquist limit indicates that the loss of information above the nyquist limit is aliasing error, which is incorrect. The distortion due to folding when that information is not eliminated by an "anti-aliasing filter" would be aliasing error. —Preceding unsigned comment added by 74.0.49.42 (talk) 20:02, 17 February 2008 (UTC)

Paul M. Rainey, 1926
Hi, I was wondering why this page has no mention of Paul M. Rainey. I quote from an introductory book on speech recognition tecnology called "Spoken language processing" by Xuedong Huang, Alex Acero and Hsiao-Wuen Hon (Prentice Hall, 2001), p.271: "Pulse Code Modulation (PCM) had been invented by Paul M. Rainey in 1926 and indepently by Alan H. Reeves in 1937, ..."

Anonymous user, 17:22, 29 October 2006 (UTC)

There seems to be some limited debate surrounding this. Rainey's patent that this text seems to be referring to is not what I would call an invention of PCM. His patent describes a system by which paper can be scanned and transmitted by encoding it using an elaborate electro-mechanical arrangement. Though the end-result is essentially a signal modulated in PCM, I wouldn't consider it an invention of PCM as it doesn't realize the general applicability of the technology. It is a very specific case and the patent applies more to the device itself than the encoding used. Reeves' patent describes PCM in detail as a way to encode arbitrary waveforms with virtually no noise. It details the specifics of PCM, its limitations, and how these can be overcome. Since Rainey doesn't describe a general case, I don't think I'd personally give him credit for inventing PCM, since his writings and work didn't contribute to today's use of the technology. It may, however, be worth mentioning in the article. For reference, Reeves' US patent is viewable at the USPTO website as patent number 2,272,070. --Ktims 01:26, 19 December 2006 (UTC)

Cathode Ray Tube Encoders
Is there a page about these on Wiki? I think the mention of gray coding and these primitive encoders is worthy of the page - but we need to reference these a bit more than calling them out by name. At least a brief description of their operation should be added.

Additionally, I don't think gray codes are limited to cathode ray encoders, either. During transitions and with uncertain inputs, digital devices would have the same problems, possibly oscillating between vastly different values. It's my understanding that most digital ADCs internally use a gray code in their detection element, with the simple decoder in the digital domain on the output.

Anyone feel like writing this? I'm uninspired lately... —The preceding unsigned comment was added by Ktims (talk • contribs) 21:27, 29 December 2006 (UTC).

Gray Code nicely explains a variety of modern uses of the code, and if you read Flash ADC alongside it you can figure why one kind of modern coder can work just as well with Gray as with straight binary. Analog-to-digital conversion also explains a few kinds that don't need Gray. I do not see a Wiki explanation of the ancient sweep coder tube, but I remember the basic theory and will write it up this weekend if nobody beats me to it. Jim.henderson 00:39, 30 December 2006 (UTC)

Attribution to Rainey? Not resolved on talk.
There is unresolved discussion on the topic of whether or not Rainey invented PCM. I disagree with the change made to the article and move that it be reverted. --Ktims 04:00, 21 February 2007 (UTC)

History of PCM and SIGSALY
Please see Talk:SIGSALY for my concerns related to the claim that SIGSALY was an early example of the use of PCM for the transmission of speech. --Gmaxwell 19:36, 19 March 2007 (UTC)


 * The history section already mentions early developments that were TDM but not PCM, I think appropriately. Certainly if the frequency components were encoded by PCM, then PCM is what it was.  Seems to me, the SIGSALY article should make clearer the distinction between PCM and samples that were quantized but not coded, and the distinction should be mentioned here as well.  Jim.henderson 17:42, 23 March 2007 (UTC)

demodulation
Is the DEMODULATION section in this article correct? I've read about something about PCM before and as i recall, the sequence requires some kind of equation-solving process, instead of "instantaneously shifting" as mentioned, to achieve such purpose. But I'm not very sure about this since I haven't really study any authentic documents. --220.129.132.253 13:08, 15 July 2007 (UTC)

Digitization as part of the PCM process
In the text it does not make any difference between Differential PCM and Delta PCM, but there is a difference. The Differential PCM is stated correctly, but the Delta PCM is actually in its simplest form just sending 1 bit per sample, indicating if it is larger or smaller that the previous sample.

I suppose that you are misreaded term "Delta PCM". It is actually delta modulation (ΔM, DM). --Čikić Dragan (talk) 18:54, 7 February 2008 (UTC)

DPCM
At this site I found this explanation:

"Differential pulse code modulation (DPCM) is a procedure of converting an analog into a digital signal in which an analog signal is sampled and then the difference between the actual sample value and its predicted value (predicted value is based on previous sample or samples) is quantized and then encoded forming a digital value."

It's closer to that I learned at faculty. --Čikić Dragan (talk) 16:11, 7 February 2008 (UTC)

Not much info about PCM
Most of this article is about digital quantization and not PCM. In fact, a quick glance at this article makes it look more like its describing pulse-amplitude modulation and not PCM.

A simple example (e.g., using +1 and -1 pulses to transmit 1 and 0 respectively) might be helpful. --TedPavlic | talk 22:29, 17 February 2008 (UTC)

Encoding for transmission?
Do we really need this much detail on e.g. ones-density in this article? These issues are generally part of the transmission system and don't really have anything at all to do with PCM itself. I think most if not all of this section could be removed, and would improve the article.

I'm also thinking some clarity is needed about what actually constitutes the 'PCM' part, and which parts of the article are discussing other technology. PCM itself is hard to separate from its practical implications though, which is why I think this is happening.--ktims (talk) 14:05, 15 June 2008 (UTC)

Change in Referenced Book
My son brought to my attention that someone changed my name as author of one of the referenced books to another name. I have edited the reference to correct this change It is unlikely that this change was accidental and I am concerned the individual is using a Wikipedia citation to misrepresent themselves in a reference. The reference in question is the second reference: Bill Waggener, Pulse Code Modulation Systems Design, Artech House, 1999

Bill Waggener Bill Waggener (talk) 17:17, 18 April 2009 (UTC)


 * Sure enough, an anonymous vandal back June 2008. Thanks for fixing.  Sorry none of us noticed.  On the other hand, you probably shouldn't have added those in the first place; there's no reason to believe, from the edit history, that they were used as sources for the article, and they're not cited, and one shouldn't generally link their own work this way.  Dicklyon (talk) 19:34, 18 April 2009 (UTC)

Old history
Some old page history that used to be at the title "pulse-code modulation" is now at Talk:Pulse-code modulation/Old history. I had to delete those old edits to make way for a history merge. Graham 87 14:12, 21 December 2009 (UTC)

Article focus
This article doesn't really seem to say much about PCM - the Modulation and Demodulation sections are all about sampling, ADC's and DACs. Wouldn't it be better to direct the reader to those pages, and concentrate here on PCM instead? GyroMagician (talk) 14:27, 20 March 2010 (UTC)

history
would M.B. Farmer, mentioned here (and many elsewheres) for his 1853 idea about time division multiplexing, actually be Moses G. Farmer, for his US Patent No. 9634 "Improvement in Electric Telegraphs" (March 29, 1853) ? Asfaltics (talk) 20:34, 6 February 2011 (UTC)


 * Evidently that's the one. Dicklyon (talk) 21:10, 6 February 2011 (UTC)

Extra information
This contribution appeared in the Limitations section. I believe it takes readers off track and so removed it. Perhaps there's a better place for it. --Kvng (talk) 14:06, 25 August 2011 (UTC)

Extra information: PCM data from a master with a clock frequency that can not be influenced requires an exact clock at the decoding side to ensure that all the data is used in a continuous stream without buffer underrun or buffer overflow. Any frequency difference will be audible at the output since the number of samples per time interval can not be correct. The data speed in a compact disk can be steered by means of a servo that controls the rotation speed of the disk; here the output clock is the master clock. For all "external master" systems like DAB the output stream must be decoded with a regenerated and exact synchronous clock. When the wanted output sample rate differs from the incoming data stream clock then a sample rate converter must be inserted in the chain to convert the samples to the new clock domain.

PCM vs. RAW
I'm wondering what the difference is between PCM and RAW audio formats. It seems like to me that PCM is more digital? Maybe? Also, which is better, what are their pros and cons, etc. Mchcopl (talk) 23:52, 18 February 2012 (UTC)
 * PCM is not an audio format, it is a method used to digitally represent sampled analog signals. Raw audio format is for storing uncompressed audio in a form without headers describing the file layout. The audio stored may have been encoded using PCM, or any other audio encoding method. Lmatt (talk) 00:32, 20 February 2012 (UTC)

When one sits down and reads what is written &hellip;
&hellip; some errors and omissions leap out at one. This article talks of "Goodall's later tube" without anywhere prior to that point (or even after that point) having told the reader who "Goodall" is or what his or her earlier tubes were. Jonathan de Boyne Pollard (talk) 16:24, 29 March 2013 (UTC)

Requested move that affects a redirect to this page
An editor has requested that the disambiguation page for "PCM" be moved to the base name, which currently redirects to this article. Please see the discussion at Talk:PCM. -- JHunterJ (talk) 11:42, 27 July 2013 (UTC)

Todo
This was left as a HTML comment in the article. I don't know when or by whom. I am moving it here today. ~KvnG 00:50, 2 March 2014 (UTC)
 * To be written - topics to be covered:


 * development of the PCM concept
 * Bell Labs and 8 kHz DS0 standard
 * early ADPCM codecs
 * mu law and A law
 * development of hi-fi digital audio
 * NICAM
 * Sigma-delta modulation

Missing context or corrupted paragraph
The paragraph that reads PCM in the late 1940s and early 1950s...a glitch-free Gray code...a fan beam instead of a scanning beam seems to be out of place or unexplained -- did we suddenly switch from audio to video? Also, the name "Goodall" is employed without being explained or introduced. — Preceding unsigned comment added by 84.227.250.195 (talk) 16:17, 23 March 2014 (UTC)

32-bit
apparently would like us to include information on 32-bit PCM. I'm reverting the restoration of Linear pulse-code modulation (which I had merged into this articl) but have added the material that the article was apparently revived for. I know there is 32-bit floating-point audio. I don't know that 32-bit PCM is a real thing and the reference supplied is not particularly reassuring. ~KvnG 18:44, 28 March 2014 (UTC)


 * The blog/discussion board reference is not good enough to state this as a fact. I see no 32-bit PCM in professional audio. Mac OS X uses to 32-bit floating-point linear PCM, but floating-point is not the same as 32 actual bits. Binksternet (talk) 19:34, 28 March 2014 (UTC)
 * look this up on internet PCM_S32BE. Also if u were to install wave surfer and use it to convert a audio sample there is a option to convert it to Lin32.  Also, VLC media player reads Linear 32-bit audio.Doorknob747 00:42, 29 March 2014 (UTC)
 * Sources Doorknob747 01:02, 29 March 2014 (UTC)
 * These sources are not good enough. None of them are mainstream authors. Binksternet (talk) 03:36, 29 March 2014 (UTC)
 * [1] is definitely a reliable source but is not particularly specific. [2] forum posts are not considered reliable. [3] and [4] are talking about floating point audio, not PCM. [5] isn't bringing up anything useful for me. ~KvnG 12:45, 31 March 2014 (UTC)