Label Question & Place of Use

MusicBrainz
,
1

This question is one I intended to ask @fmera , but I think this is a better way to do it. When viewing other places, such as the ISRC database and iTunes files, the “label” used is not always the label MB wants. That topic I do not wish to ask or discuss here.

My question is… what shall, or can, we do with that information? The names/titles applied to the containing fields include, specific examples, “Vendor”, “XID”, “Release Label”, etc. As said above, I know that although that is either a match or close match to what MB asks for on the label entry field on the release page, it is not always the case, but could be only. Now, for others, me included, this information is the information wanted.

fmera has stated prior that there are other fields and relationships where labels can be associated with other areas, but in cases like this, the listed title is not a match to MB, so how would one know where to place them? I would make a guess as distribution, but that is a somewhat narrow minded thought and also is not always going to be the case.

This question may need some specifics for an example, so if this is the case, I will provide. I am unsure the knowledge of the usage of such things to all here, so I will leave it as just ask if you need more info to properly provide input. And to repair for clarity, this question is not at all regarding what to enter into MB for a release where it asks for label and catalog number. It is also worth noting that this question applies mostly to digital releases vs physical.

1 Like
2

Are you speaking about digital releases or physical releases?
IMO, for physical releases, we should include all visible labels on the package, booklet, mediums, stickers, etc.

3

Digital primarily. This question may apply to a physical release, so I do not exclude them, but focus is digital.

4

iTunes specifically seems to have (c) on their site (or used to at least, haven’t checked in a while and they keep changing things). If so, that goes as “copyrighted by” on the release level.

I have no idea what the first two even mean. The latest sounds like it should fit what we want, doesn’t it?

5

Ok, I can explain further.

iTunes lists the ℗ on their site I know for sure, along with who it is licensed to sometimes. The Vendor and XID refer to atoms in the metadata. XID = vendor supplied code. An example of the data might be “WarnerMusic:isrc:US23A1500057”. Vendor is, well, the vendor, like “Warner Music” for example. I would be happy to supply real data examples if anyone wants, but that is the format of the data in those atoms.

The third, Release Label, is used, for example, on the ISRC web site. It appears to resemble the Vendor / XID quite often, but not enough to be the same. None of the three examples match what MB wants as the label on a release, thus I disregard that field. I am trying to figure out where and how I can add the data. There are other places I have seen such data, for example in “copyright” which is obvious and “Publisher” which is not as obvious but also not as commonly used for such things.

On physical releases, it is clear that the logos are all important to the label listed. For digital there are no logos and the info used and listed is different. With digital, I basically always leave it blank as there is nothing that lists the label as MB wants to see it, but there is data there and that data is very useful… but MB does nothing with it. There is a lot more data there too, but I was just starting here. One of the others I think MB would be able to use is the NR/Explicit/Clean flag. Makes it real easy to know if explicit just by reading the atoms. That is another topic though.

EDIT: Here is an example I recently added data to.


In MB, someone used Atlantic as the label. Right or not, Atlantic is there. In the metadata, I have:

  1. XID – Warner:isrc:USAT21801395
  2. Copyright – ℗ 2018 Atlantic Recording Corporation for the United States and WEA International Inc. for the world outside of the United States. A Warner Music Group Company

This release and recording are not listed on the ISRC site, so I cannot pull data from there, so I will need to locate another example for that. But you can see here, there is already a slight conflict going on.

For me, those I talk to, digital retailers, etc, that tells me the label is Warner, the ISRC is USAT21801395 and the copyright is with Atlantic. Issue is that MB uses different data for its label, so I would like to find a place for the information that applies to digital releases, such as what I listed here. Please note that the example provided is an iTunes file example.

1 Like
6

Hi there,

I am reviving this old discussion as it seems to be the most comprehensive one to date about (steps towards) guidance on how to identify labels for digital releases, but it didn’t quite reach firm conclusions.

I hope we can move it forward at least partially, as most editors are entering (manually or via scripts) the most readily available information presented to them on online stores without questioning it, and the scale of the issue is only growing as both digital media and import scripts play an ever larger part in the database. In my view, the end goal should be a reliable methodology for determining the label(s) for digital releases that editors can apply (differentiating by store as necessary) - as a digital complement to https://musicbrainz.org/doc/How_to_Identify_Labels.

I have to admit I have more questions than answers myself, especially after reading @thwaller’s insightful posts above, so I will keep this brief and hope that someone with more direct knowledge of commercial metadata and/or of how the music industry and online shops work in this respect can step in.

  • For new releases where both physical and digital releases are launched at the same time, would it be generally safe to assume that the correct label as identified on the former also applies to the digital release (if in lack of an authoritative source for the latter’s label)? Or is it common for digital releases to have a different legitimate label than physical ones (note the emphasis)?

  • Can we develop a common understanding of the most appropriate field (if any) to determine a release label for each of the most common online stores? Once we have an answer, we should also be able to clarify whether the process of entering this information is something that can be expected of users, or if it should be delegated to bots.

  • Based on the answer to the questions above, should we encourage script authors to modify/remove the automatic import of labels for digital releases (if based on fields that we consider unreliable)?

I hope this is enough input to restart the discussion on such a complex topic. Feel free to tag any user who you think might be able to contribute based on their expertise. I look forward to reading your replies!

2 Likes
7

I regularly import Japan digital releases. I use my own tools and my workflow includes interpreting text strings, rather than copy-pasting them like most “off-the-shelf” scripts would. My answers:

For new releases where both physical and digital releases are launched at the same time, would it be generally safe to assume that the correct label as identified on the former also applies to the digital release?

Yes, especially if both are “manufactured” by the same record company.

Or is it common for digital releases to have a different legitimate label than physical ones (note the emphasis)?

It can happen, but it’s usually accompanied by “obvious” differentiating factors like a mismatch in the barcode prefix. This is usually because of release country rather than physical/digital format though.

Can we develop a common understanding of the most appropriate field (if any) to determine a release label for each of the most common online stores?..

Well, no - the problem is not that many stores exist and differ. The problem is that many record companies exist and differ. Consider these 2 examples from the same store:

  1. https://www.amazon.co.jp/-/en/BACK-HORN/dp/B01NCUOZTP/ says the label is “Speedstar”. Speedstar is a known release label in the Victor Entertainment record company. The CD and digital release match here. So far, so good.

  2. https://www.amazon.co.jp/-/en/gp/product/B08HQDKFDZ/ says the label is “Universal Music LLC”. This won’t do, because it’s the legal name of Universal Music Japan - that makes it a huge record company, not a specific imprint / release label. However, if we shift our eyes to the copyright field, which can also be seen at Apple,

    Label: UNIVERSAL MUSIC LLC
    Copyright: A UNIVERSAL J / Perfume Records release; ℗ 2020 UNIVERSAL MUSIC LLC

    UNIVERSAL J is a known division of Universal Music Japan, and Perfume Records is a known release label. The CD and digital release match here.

The purpose of these 2 examples is to demonstrate that the question is not,

Can we develop a common understanding of the most appropriate field (if any) to determine a release label for each of the most common online stores?

Rather, the question is,

Can we develop a common understanding of the most appropriate field (if any) to determine a release label for each of the most common record companies?

This is not a technical challenge, but a domain knowledge challenge. It requires the scripter/editor to consider the input critera (record company/barcode) to deviate the process of picking a release label in MusicBrainz.

Frankly, public script authors aren’t going to implement this. It requires a firm grasp on the domain knowledge. The script author might resist, either because of the scope of the task, or because they don’t believe they’re doing anything wrong by facilitating the copy-paste of the invalid text string!

4 Likes
8

This is still an interesting topic to me, so I wanted to share what I am doing since not using the MB data. First, I no longer use MB for identifying digital releases as it is most times incorrect, populating data into my files that not correct to the release and easily overwriting proper data. I am able to use it to help standardize file names, but that is about the extent of it. I sill like and use MB for physical releases as it is the best resource I have seen and used. While I may be blunt, I do not intend to argue my points, there is no need to prove what works to others. Where I note disagreements with those at MB, it is not meant to be insulting to anyone in any way, it is simply to paint a picture of why and how I got to where I am today with this issue.

I have taken the approach that the data saved in my digital files should help me identify the file in the release it is from. This data structure does vary though, it is different from store to store. I found this to be an important aspect of the metadata. Unlike a CD (except cases like BMG where the actual printing and often barcode are different), they all supply the same physical CD. This is not the case for a digital release. Examples… for Amazon releases, I like to keep the ASIN in the metadata and for iTunes the AppleStoreCatalogID.

I particulally like iTnes as they have great metadata. It includes a copyright field for that version of the label and a vendor field containing the ISRC and the ISRC issuing label. iTunes does not use a barcode to identify their releases, so neither do I. When an iTunes release is presented in the Apple store, no barcode is ever disclosed and when you get the files, no barcode exists there either. While it is fair to argue that Apple does have a barcode, it is used internally and plays no role in me identifying the release I have.

Amazon is not as good, and causes some conflicts with the mindset at MB. They are known to have what MB considers an incorrect label listed. This is problematic for me, while I understand the logic used at MB that says this is wrong, if I am going to identify my release, this information is correct. If it says “Company A” and MB says it is “Company B”, this is creating a mismatch which is counterproductive in my identifying my release.

I have also switched to using ISRC as a primary identifier. Contrary to what people at MB have stated, I have zero issues with this. While I understand that on CDs and prior physical media it was a bit more common to have duplicates, I just do not see problems existing in my over 5TB of digital music files… and it makes it really easy to locate duplicate recordings across multiple releases and multiple store sources. This is a change I am happy I did despite the hard disagreements here, it brought my file organization to a new level.

I will also add that I differentiate the store because not all files are created equal. Again, contrary to what those at MB say, I can and do find differences in even FLAC files. Those that want to argue that there is simply no difference in 16 vs 24 bit and 44.1, 96 and 192kHz either simply do not FULLY understand what they are talking about, or just do not look enough into the data in the file to realize and see the differences. All are free to have opinions, and I will also say that those opinions might be just fine for MB standards. I am sure the differences do not matter to many people so it may not be worth the time to be so detailed. This is even more true with MP3 and M4A files, as well as the other various lossy compression formats.

To summarize… I am using the ISRC as a primary identifier, similar in saying this release name by this record label. It allows me to locate the same recording included in multiple releases easily. My next level of distinction is the store/source of the release. IF there is no store, say I RIP a CD of mine, then I populate the CD information as the source in the metadata… listing out the barcode, cat number, etc… the stuff MB does very well at recording for physical releases. This also allows me to have recordings with different names and ACs. That to me is simply a name given to it on a release, the actual recording ID is the ISRC. So that is what I moved to after stopping use of MB data, and all I can say is that it works, it is what I use for each and every file I have.