Talk:Modified frequency modulation

Symbols per flux transition

The article says "MFM allows more than one symbol per flux transition — up to three — giving greater density of data than FM." I think this is nonsense. FM encodes a one as a 11 and a zero as 10. MFM encoding is as shown. They have the same density, given the same clocking; each takes two "possible transition" periods to encode a bit. The reason MFM achieves greater data density is it can be clocked at twice the rate (because there are never two transitions in a row). Or am I totally off-base here? See http://www.storagereview.com/guide2000/ref/hdd/geom/dataMFM.html

Nybbler 23:29, 5 March 2006 (UTC)

It may be confusing, but I think the statement is correct. Your statement is also correct. Mirror Vax 00:34, 6 March 2006 (UTC)

I'm still not seeing it. A one followed by a zero is 0100, two bits with one flux transition. A zero followed by a zero is 1010 or 0010 two bits and one or two transitions. A one followed by a one is 0101, two bits and two flux transitions. A zero followed by a one is 0001 or 1001, again two bits and one or two flux transitions. There's never three symbols with one flux transition.

Nybbler 15:28, 6 March 2006 (UTC)

Questionable question marks in the examples

It would be nice if the question marks in the examples would be explained somewhere. --89.49.249.90 22:13, 17 August 2007 (UTC)

Verify the illustrations?

Can someone double check the illustrations?, when I study them in detail. They just don't seem to make sense! Esp, a "clock" signal of 11000000 seems weird. It ought to be 10101010 ? Electron9 (talk) 17:22, 5 May 2009 (UTC)

Introductory statements are incomplete and incorrect

Article begins with:

Modified Frequency Modulation, commonly MFM, is a line coding scheme used to encode information on most floppy disk formats, which include the floppy disk formats used in the classic versions of Amiga OS, most CP/M machines as well as PCs running DOS.

This is incomplete and/or incorrect in at least two respects:

1. MFM was also used for hard disks. Some links:
   1. ST-506
   2. Tandy MFM Hard Disk Drive Reference
2. This has almost nothing to do with the OS ... or even the computer:
   1. Other OSes can be used with MFM (including, say early Windows)
   2. Those OSes can be used with other modulation types
   3. Amigas, PCs and "CP/M" machines can use other modulation types. Sometimes and on some of these machines there is a built-in MFM circuitry, at other times it is on extension cards.
   4. Even today's floppies are still MFM ... So today's OSes (say Vista, Windows 7, etc.) still support it (but then again, they have nothing to do with MFM directly - it is the onboard circuitry that does)
      1. If this was true, a typical PC floppy drive would be able to read an Amiga floppy. It is not. The Amiga, however, can reprogram its floppy drive to support both MFM and GCR, enabling it to read both Amiga, Mac and PC floppy disks. The difference between MFM and GCR is also what enables the Amiga to put 880 KB on a floppy where PCs store only 720 KB. This difference should be more clear in the article IMHO. JoaCHIP (talk) 12:40, 18 October 2015 (UTC)

PC's were able to format DS/DD 3½-inch or DS/QD 5¼-inch Floppies, usually marketed as 720kiB (formated 512 bytes per sector, 2 sides, 9 sectors per track, 80 tracks) with 800 KB or even more, with additional drivers (i.e. 800.com or 900.com). I do not know if it was possible to change modulation with these programs, but you could increase number of tracks and sectors per track and gain more place on disk. To read such disk there had to be one of the drivers loaded, which enabled to format it beforehand. From DS/DD 5¼-inch Floppies (Marketed at 360 kiB), it was possible to get maximum 430 kiB, but 410 kib was more reliable and more supported variant, as we found out later). 195.14.165.61 (talk) 07:12, 25 August 2016 (UTC)

MFM was just an early system and was used for the devices built at that time, such as floppy disk drives and early hard disk drive MFM controllers, such as WD1002. While hard disk drives, such as Seagate ST-225 were intended for MFM, some users connected them to newer RLL controllers to gain additional 50% of capacity. One example I was able to find at Hard Disk Hardware Explained, Part II

RLL

Some hard disk manufacturers use something called RLL. RLL is a different way of recording data on the disk drive. Previously, the industry standard was MFM (Modified Frequency Modulation), or "double density," the same technique used on a floppy disk. RLL (Run Length Limited) lets you pack more data onto the hard disk, a sort of "double-double" density. In practice, it gives you about 1.5 times the amount of storage you presently have. For example, a 20-megabyte drive can become a 30-megabyte drive, just by changing the disk controller and reformatting. —Preceding unsigned comment added by 208.65.73.202 (talk) 14:56, 12 May 2009 (UTC)

BIT order on media

I vaguely remember that bit-order on floppies (may ST-506 HDD, too) was LSB first. If anyone knows that for sure, it would be nice to put it here for reference. —Preceding unsigned comment added by 91.187.225.16 (talk) 08:08, 31 October 2010 (UTC)

Here's the signal from my floppy drive that I'm working on right now (top track, signal; bottom track, sync). It is clearly MSB first -- Wesha (talk) 22:34, 22 April 2020 (UTC)

Digital signal?

It's not obvious to me whether MFM applies to the analog or digital world. "Pure" FM is clearly an analogy phenomenon, whereas the article discusses streams of bit symbols, which is clearly digital in nature. So where does that leave MFM? Is it an analog technique that happens to have digital applications, or is it a strictly digital technique? — Preceding unsigned comment added by 70.247.162.60 (talk) 04:18, 18 November 2015 (UTC)

For the precursor of MFM, called FM, delay code or digital FM are better names. It is just a digital signal, not an analog one, which is why delay code or digital frequency modulation are much better terms. For MFM the same applies: it is a purely digital signal. This should also be taken care of if the merge of the MFM and delay code articles is to happen.87.179.128.35 (talk) 02:23, 26 January 2017 (UTC)

Delay (en)coding

The lead says:

MFM is a modification to the original digital FM (digital frequency modulation also known as delay coding)

And now since the merge, delay encoding is discussed. This delay encoding is MFM, right? So is somehow the difference between FM and MFM the difference between delay coding and delay encoding? I get the feeling delay coding is a different way of saying delay encoding, and the lead is simply wrong: FM is Differential Manchester encoding and MFM is delay coding is delay encoding. If I'm right, we should change the lead to

MFM is a modification to the original digital FM (digital frequency modulation also known as differential Manchester encoding)

It also couldn't hurt to be more explicit that delay encoding is actualy MFM; the current merger doesn't make this obvious to the reader. Digital Brains (talk) 13:47, 5 February 2018 (UTC)

How much recording technology is appropriate to this article

@Myrtonos: has attempted to add fragments of recording channel design to this article which I have reverted both on the basis that they were incomplete and inappropriate to this article. The lede accurately states that the "the minimum spacing between flux transitions that is a property of the disk, head and channel design, ..." An article on recording theory might note that the isolated pulse is a function of gap length, flying height and media thickness. The isolated pulse shape along with signal to noise (including off track signal) and data separator stability all go into the code chosen which has a minimum and maximum number code zeros between code ones. This is all summarized in the lede and adding parts of this complex subject into the body is inappropriate. I note that none of this appears in the Run-length limited article which is the parent category to this article on MFM. It doesn't even appear in Magnetic recording. WikiPedia would be improved by a magnetic recording theory section possibly in Magnetic storage#Design which could be linked from this article put the partial details that Myrtonos has added simply don't belong here. TMI for this article! Tom94022 (talk) 17:39, 26 May 2018 (UTC)

The relation between frequency and playback output is why the recorded signal needs to be D.C free.--Myrtonos (talk) 11:36, 27 May 2018 (UTC)

I'm not sure what this has to do with the edits removed, but is it even true? The MFM recording of "B6DB6D..." is not DC free which is why magnetic disk recording channels differentiate. Tom94022 (talk) 19:03, 27 May 2018 (UTC)

Missing sync in A1 marker???

I'm investigating my old floppy disks right now, and I totally do not see a "missing sync bit" in the A1 sector marker this (and some other) article are talking about. Help? (top track, signal from the disk; bottom track, meander manually added for reference that's supposed to represent the real sync signal)-- Wesha (talk) 22:38, 22 April 2020 (UTC)

OK, maybe I'm starting to understand. In terms of signal, "flux" is transition of low to high or high to low, so LH or HL constitutes a flux, and HH or LL do not. The actual recording on the track is HL LL LH HH LL LL HH HL (or the reverse) which according to MFM rules means 1 0 1 0 0 0 0 1 but the catch is the LL LL part, as the signal was supposed to flip there (LL HH). -- Wesha (talk) 17:44, 23 April 2020 (UTC)

Frequency modulation vs. Digital frequency modulation

Tom94022 insists that the precursor to MFM cannot be called Digital frequency modulation. But I don't think what's described in Frequency modulation is right. How do we make this disambiguation clear to readers? Would creating a Frequency modulation (digital) redirect help? ~Kvng (talk) 14:47, 16 January 2021 (UTC)

I don't think a redirect creating an invented term would help. I suggest we fix Frequency modulation. If the hatnote at Frequency modulation, "For frequency modulation (FM) as used in digital coding, see Differential Manchester encoding." is incorrect or insufficient lets fix it. The problem is that FM has two meanings, the conventional meaning is described in Frequency modulation but the other less conventional meaning in computer storage and elsewhere is in Differential Manchester encoding, a synonym for that meaning of FM. Tom94022 (talk) 19:35, 16 January 2021 (UTC)

While to people from this field of expertise Frequency Modulation will clearly refer to Manchester encoding, many others who will also read this article will be surprised when used as such. For them, it should be clear that the term can mean several things and that this article uses the term in a specific sense. If indeed no one in the field refers to Manchester encoding as digital FM, then I understand the desire not to invent terminology. So the article should clearly establish what FM means before exposing the reader to the plain unadorned term Frequency Modulation, to remove any pre-existing misconceptions.

Furthermore, I don't think this article should link to Frequency Modulation for more information on the digital variant. Because that means people first click that link from this article, and then read in the hatnote that they actually need to go somewhere else. That's even worse than a link to a disambiguation page. So the hatnote seems irrelevant in that sense, as we shouldn't link to it for more information on Manchester encoding anyway.Digital Brains (talk) 10:33, 17 January 2021 (UTC)

The hatnote at Frequency modulation is fine. I agree we should not link to Frequency modulation from here. Nor should we use a piped link Frequency modulation because that's too WP:ASTONISHING. Probably best to start by linking directly to Differential Manchester encoding and then explain that in this field it is commonly known as Frequency modulation. I can take a crack at this. ~Kvng (talk) 23:36, 20 January 2021 (UTC)

Digital frequency modulation is not just a single method, it is a group, more commonly called line codings and keying methods. Differential Manchester encoding is only one of them, and Manchester encoding and MFM are others among many. The Manchester methods are probably better called shift-keying methods in angle modulation, as they also involve phase. The articles do not distinguish these sufficiently. The hat note of the FM article (For frequency modulation (FM) as used in digital coding, see Differential Manchester encoding.) is misleading the reader into a false singular identification, what is provided is just one of many examples. I don't see any good justification for this hat note. The article already has a table of all kinds of methods that would be more appropriate for the reader to navigate to find a topic that is a good match for their interest. kbrose (talk) 02:12, 21 January 2021 (UTC)

PS: A pitfall for an encyclopedia is to fall into the trap of using the various industrial and engineering terms of the vast array of modulation methods in too literal manner. When digital circuit engineers talk about frequency modulation it sounds like they own the concept, while analog transmission technology has a more fundamental idea. Therefore, the use of phrases such as "original frequency modulation" in a digital context is terribly misleading, because it isn't the analog method, it is just an earlier keying or line coding method. kbrose (talk) 02:58, 21 January 2021 (UTC)

Hmmm, I was under the impression that Manchester, MFM etcetera are all instances of the concept line coding, and are all in essence baseband signals. But because Differential Manchester happens to look like binary frequency shift keying, being equal to it for a very particular set of frequencies, they called it frequency modulation even though frequency modulation is definitely not a baseband modulation. So the name was coined based on a perceived similarity rather than that it is an instance of the class. And this means that I wonder whether Kbrose's claim that Differential Manchester is part of a group of frequency modulation methods is correct. Digital Brains (talk) 14:45, 22 January 2021 (UTC)

I don't disagree. I believe another term for line coding methods is digital baseband modulation or transmission. kbrose (talk) 23:54, 22 January 2021 (UTC)

Kbrose, this is helpful but it does not give me clear ideas about how to fix the deficiencies in the current treatment of the subject. Do you have any specific improvements to suggest? ~Kvng (talk) 14:57, 25 January 2021 (UTC)

I think the problem we have is the term frequency modulation has two different meanings, a rather specific meaning in magnetic storage and a more general meaning in communications and/or information theory. They really were two distinct arts until the late 1970s when the magnetic folks realized they had a communications channel and started to treat the recording playback process as one. Modified frequency modulation seems to be limited to magnetic storage and since that is the subject of this article perhaps we should limit the discussion in this article to the magnetic storage meaning. I tried to edit the article in that direction. Perhaps a hatnote pointing this out in the Frequency modulation section with a link to the more general Frequency modulation section. Tom94022 (talk) 01:18, 26 January 2021 (UTC)