CD manufacturing and matrix guide

MusicBrainz Style
, , , , , ,
1

after a brief dive into the world of CD manufacturing, I thought it’d be good to have this information in one spot.

edit: now a wiki user page!~ User:UltimateRiff/CD Manufacturing and Matrix Guide - MusicBrainz Wiki

edit 2: now a proper wiki page!~ CD Manufacturing and Matrix Guide - MusicBrainz Wiki

The Basics

but what even is an matrix? well, you came to the right place! :wink:

you’ll find the matrix on the inner ring on a CD, DVD, or Blu-ray, usually on the bottom (silver) side of the disc, but on some older discs it’ll be visible from the top. a couple examples from CDs:


in addition to the alphanumeric code around the ring (aka mirror band), on discs made after 1994 there are two numbers marked IFPI; these are the mastering and mould SIDs.

taking the second example above, the mastering SID can be read as IFPI LT05 underneath where it reads “6037 - FIVE IRON”. mastering SIDs appear in the mirror band and always start with “L”. this can tell you where the glass master for the disc was made. (note that mould SIDs can also start with “L”? needs checking)

the mould SID is usually harder to spot. it’s molded into the plastic of the disc in the (usually) clear part around the hub. as such, it doesn’t usually show up on scans (IFPI A603 is barely visible near the top on the second example). an example picture of a typical mould SID; top left, just on the edge of the light. this code can tell you where the disc was pressed.

sometimes, you’ll also find the pressing and mastering credits in plain text in the matrix or pressed into the hub. for example, many EMI Uden glass masters are credited as such, and the second example has RAINBO CA stamped in the clear hub area (can barely be made out near the bottom of the image).

each SID actually refers to a different machine, either a Laser Beam Recorder (LBR) for glass masters, or a stamper or unique mould for pressing.

there are several lists of IFPI codes, including ones from redump, Pink Floyd Archives, Madonna Discography, and others. some may be more up-to-date than others. Discogs label pages also often include a list of SIDs related to a particular pressing plant or label, see EMI Uden above or MediaMotion, for example.

do note, however, often SIDs can be transferred due to different plant owners, so they don’t always relate to a label exactly one-to-one (although places are more stable? need confirmation). see how IFPI 15## is both used by EMI Uden and MediaMotion at different times.

How to Credit

note, there is currently no real standard as to how to enter matrix codes in the annotation, this is just how I do it.

to get nice-looking headers like the ones here and to make annotation searching easy, you can copy this into the annotation:

== Matrix ==
=== Disc one ===
    * {company logo} #####@@-##@@ ('''Mastering SID:''' IFPI xxxx '''Mould SID:''' IFPI xxxx)
=== Disc two ===
    * {another matrix code goes here}

you should start the code at a place that makes sense, usually after the barcode section (when it exists)

for company logos, I put them in curly brackets, occasionally I’ll also link to an image of the logo from Discogs, if I think it’ll help. if there’s already an annotation, I’ll usually add the matrix to the bottom, since more often the annotation that’s already present is more generally interesting and helpful for identification than the matrix code.

On Label and Place Pages

I don’t usually see SIDs listed on MusicBrainz label and pressing plant entries, but there’s no reason not to add them if you wish. an example I recently added is MediaMotion’s annotation. you can also add links to other labels and pressing plants that inherited SIDs from or to the entity in question.

an example to copy-paste:

== Identification ==
The following SID codes may appear on CDs manufactured by {company name} (but as this manufacturer has been known by other names before and/or after, SID codes should not be used to credit {company name} where its name is not shown).
=== Mastering SID codes ===
    * IFPI ****
=== Mould SID codes ===
    * IFPI *****
    * IFPI ****

Helpful Links

  • a MetaBrainz thread with a lot of helpful information
  • (more to come?)

Helpful People

this is more to get eyes on this post, but @IvanDobsky, @ernstlx, and others are quite familiar with pressing details, especially in Europe.

lemme know if there’s anything I missed, and feel free to add to this~

15 Likes
2

Thanks, great!

In general, places are more stable. But there are exceptions.

  • SID codes are assigned to companies. And the company can decide to use the same SID codes on several of their plants.¹
  • Equipment can be transferred to other plants (mainly after a plant was closed)
  • In very rare cases there is unauthorized use by another plant (mostly fake mastering SIDs)

¹) Example: Sony DADC Austria with plants in Anif (until 2019) and Thalgau. Both plants used mastering SIDs in the range of L551 to L559 and pressed 94** (** = 2 alphanumeric characters)
or MediaMotion with mould SID codes AAH** on all of their plants.

4 Likes
3

update~ I’ve finally copied this over to the actual wiki~ User:UltimateRiff/CD Manufacturing and Matrix Guide - MusicBrainz Wiki

even though it’s on one of my user subpages, feel free to edit it there if you’d like~

5 Likes
4

Great!

Perhaps the relationship between matrix code (LBR written code) and disc ID could be highlighted. → Discs with the same matrix code will always have a distinct disc ID submitted.
And I would like to have a call included, whenever adding a disc ID or setting track lengths from it, to add the matrix code of the CD as an edit note. (If this is a matrix guide, suggestions on how to use disc matrix data on MB may also be included)

Other thoughts: on most occasions it should be “matrix code” instead of “matrix” - the matrix is the whole thing (can still be a “matrix guide” :slight_smile: )
The mould SID is best spotted with a magnifying glass and a side-way light source.

My annotations are somehow different in style (example), but I would gladly adapt to a standardized form.

5 Likes
5

Just curious why do you wish to call it a ‘code’ - it’s normally just a string? Often times we also see logos there which can be quite important, some manufactures do actually use some kind of a barcode or barcode-like thingys which both get largely ignored (at least by me).

1 Like
6

These are barcodes but no known standard and can’t be read easily. But the existence of such a barcode can help to determine when the glass master was made. It’s ignored on Discogs and therefore rarely added here, but it would make the information more complete ← noted as {barcode} :slight_smile:

It is a code that stands for the specific data written to a glass master (sometimes with additions for duplication stages)

7

I’ve been using bulleted lists since there might be slight variations in one release (like for manufacturing dates in the matrix or a different machine with the same glass master in the same plant). I might add additional sub-points for SIDs to my standard style tho, as your form looks a bit cleaner~ (and that would allow for others to add SIDs from their own copies)

we could also start using heading 3 like you use, but I think it might make more sense if it’s under a heading 2 (something like == Identifying this release ==, == Identifiers == or something?). I don’t know if I like that, but using a standard heading would make it easy to search == Matrix == or === Matrix === in annotations (maybe the former search will find the latter tho?)

that said, we could follow the example of Sub Optimal Credits and just use heading 3 with no heading 2… idk…

I’ll try and add those notes about Disc IDs and whatnot later this week~

1 Like
8

Yes, but I do not focus on an exact manufacturing date and mould SIDs are not very useful for this purpose. The first two characters are company specific (and usually the same) and the last two are not really sequential (and often re-used after some time).
And I do not consider represses as different MB releases (of course no artwork changes), even if the manufacturing company has changed its name.
But if we want to allow other users to add their own mould SIDs, we can define bulleted lists for them.

In general, I’m more interested in the glass master code and its related disc ID. This is why I always add the disc ID that belongs to the matrix code.

I have not though much about such things.

At first I used == Identifiers == and === Matrix === for matrix information and other titles for other relevant information, but then I thought, - of course “identifiers” because it’s all about identification of the release - I don’t have to tell. So I only need heading 3.

I have developed this style over a long time (with many changes) - at first I added for mould SID only (e.g.) “IFPI 94** (** = 2 characters)” but more recently I started to add my distinct code, call it example and explain the relevant parts if that’s necessary - in most cases, it isn’t as there will be no similar matrix code in all future.

1 Like
9

I have a proposal for a wiki section, probably best added at the end:

Matrix and Disc ID

The matrix code of a disc stands for a specific data content (o̶n̶l̶y̶ ̶t̶h̶e̶ ̶c̶o̶d̶e̶ ̶w̶r̶i̶t̶t̶e̶n̶ ̶b̶y̶ ̶L̶B̶R̶,̶ ̶"̶s̶t̶a̶m̶p̶e̶d̶"̶ ̶c̶o̶d̶e̶s̶ ̶a̶r̶e̶ ̶a̶d̶d̶e̶d̶ ̶d̶u̶r̶i̶n̶g̶ ̶d̶u̶p̶l̶i̶c̶a̶t̶i̶o̶n̶)¹. This code is unique (very few exceptions) and will not be found on discs with somehow different content.

Part of the matrix data is the TOC (table of contents), a series of numbers (mainly start sectors of tracks and lead-out) from wich the disc ID is calculated. Thus a specific disc with its unique matrix code will always generate the same disc ID, whatever was printed on the top side.

Note that this is not true in the opposite direction. The TOC can be copied to a new glass master and then the same disc ID will also be calculated for a disc with a different code. This is common with new glass master versions made at the same plant, rarely happens with different plants.
Often discs with the same disc ID will show a similar code with only a forward counted number at the end.

Knowing what matrix code the CD submitting the disc ID had is an important information. It is good practice to enter the matrix code as an edit note for each edit Add disc ID and Set track lengths.

¹) Regarding struck-through text: At some point I have to mention that I’m only talking about the code written by LBR, but I didn’t find the point, nor the words…

PS: I guess some sentences need reworking

2 Likes
10

isn’t the CD TOC technically part of the data area, not the matrix code in the mirror band? I seem to remember for CD-Rs, if they’re burned in multiple sessions you’ll have multiple TOCs, one for each session (don’t know if this is a hint as to how pressed CDs are formatted tho), and they’re just glued to the end of the previous session

just to confirm, the LBR written code is the code in the mirror band? I might mark up one of those scans as a diagram… about the time I make these edits…

11

I’ve written

Part of the matrix data is the TOC

The LBR writes the data and the “mirror band” (except those parts added during duplication - mostly called “stamped”, although more likely burnt in)

I think we only have to consider (red book) audio CDs. Burnt CD-Rs have no useful matrix code and DVDs or BDs no disc ID.

1 Like
12

so would it be fair to say that the TOC isn’t part of the matrix code, but kinda determined by the glass master, which the matrix code identifies? (that sounds a lot more confusing written out than I expected)…

so basically, LBR-written code (mirror band) = glass master = TOC, so therefore the LBR-written code in the mirror band determines (or rather identifies) the TOC in that way… am I understanding that correctly?

I’m also not personally counting anything stamped as part of the mirror band, since it’s on top of the plastic covering the data layer and mirror band, even if it’s over the mirror band

I think it is important to note that CD-Rs, DVDs and Blu-rays do have matrix codes tho, but it’s less important for most CD-R and CD-RW, and currently we have no Disc ID system for DVDs and Blu-rays. I think the matrix codes on these video media are still important to have, since they’re manufactured in a similar way, just with smaller data so you can fit more on a disc

2 Likes
13

Yes, correct, the matrix code represents the TOC (and the whole matrix data) but the TOC isn’t part of the code.

Only the mould SID is formed in the plastic. Other code is usually added to the mother or stamper (son) copy and directly in the metal coated layer, thus to be seen on the same metal foil as the glass master code (under the plastic).

Certainly, and I also add BD, DVD, SACD matrix codes to my annotations. The title should better be “Matrix and Disc ID of Audio CDs”.

1 Like
14

I’ve seen cases otherwise, like the Five Iron Frenzy CD I use as an example has “RAINBO CA” formed in the plastic, not in the mirror band.

edit: I’m also adding a section on re-writable media (drafting currently):

Matrix Codes on Re-writable Media

while nearly all digital disc formats have matrix codes, not all of them have quite the same importance. for example, for CD-R and CD-RW (writable and re-writable respectively) discs, the manufacturing isn’t typically as important as for pressed discs, as the person or company doing the CD burning could have used discs from different companies, factories, etc. in a single run of a release. these can be split into seperate releases, but I haven’t seen any cases where this has actually been done.

2 Likes
15

Ah, yes, these were kind of pressing plant identifications, moulded into the plastic.
I thought mainly on additions like in Uden (1-2-3-NL) or Langenhagen (MADE IN GERMANY BY PDO). I try to find a microscope image I have in mind. It shows both mould SID and a part of the company credit.

EDIT: found the image - focus on mould SID, the company credit is behind
NP-K-mould-01R2

(the microscope looks from the top down, thus reflection between the layers are above)

EDIT2: A 1200 dpi scan of an Alsdorf disc with Warner logo and a “2”, formed in the clear hub area (like Five Iron Frenzy’s “RAINBO CA”):

1 Like
16

They could be separate releases… :thinking:

It should be. The self burnt prolonged release (I know, or I have something like this) could also have a different disc ID if the the originally pressed version had an index 1 pregap (32 sectors kind). And it was certainly released later.
… and actually I can’t think of many other cases, because even if it was made professionally, it would be a re-relase and separate.

17

Wouldn’t this mean that burn-on-demand CDs would have a separate release for virtually every copy?

18

while it’s technically possible, I imagine most burn-on-demand services (mainly thinking of Amazon) would use discs from only a few sources, not a different source for each disc. that said, I’m not certain this would be the case. I did have that same thought while typing that up tho~

19

I don’t think that’s necessary. Raw CD-Rs are not relevant for for the release. They are only tools like a mould for pressed CDs. Even if discs from different suppliers would be used, they can considered to be the same release.

I thought about burnt vs pressed disc releases. They have to be separate. A release can’t have format CD and CD-R.

I have not known that Amazon has such a service, but many smaller pressing plants offer CD-R as an alternative for smaller quantities. (e.g. P+O Pallas). The top side of such discs look no different than pressed CDs .

1 Like
20

Amazon has (or maybe had?) a subsidiary called Amazon Media on Demand (previously CreateSpace) which does it this way, and at least two CD-R manufacturing plants, which are in Kentucky and South Carolina

2 Likes