History:What Defines A Unique Release

From MusicBrainz Wiki
Jump to navigationJump to search
Status: This Page is Glorious History!

The content of this page either is bit-rotted, or has lost its reason to exist due to some new features having been implemented in MusicBrainz, or maybe just described something that never made it in (or made it in a different way), or possibly is meant to store information and memories about our Glorious Past. We still keep this page to honor the brave editors who, during the prehistoric times (prehistoric for you, newcomer!), struggled hard to build a better present and dreamed of an even better future. We also keep it for archival purposes because possibly it still contains crazy thoughts and ideas that may be reused someday. If you're not into looking at either the past or the future, you should just disregard entirely this page content and look for an up to date documentation page elsewhere.

Style

When does a work actually become a "separate record" from MusicBrainz point of view? This is a somewhat complex issue which relates to both the current data model we are using, the tagging service, and to some extent the way users interact with the data. MusicBrainz has been widely hailed for its GUID approach to identifying music, but what does this exactly mean?

Roughly, that MusicBrainz base data units are unique sets of recordings.

Hence, a release which is a reissue (or reprint) of another (same recordings, same TrackTitles, same track times, same ReleaseTitle, same ReleaseArtist) is not considered to be a different release and should not be duplicated. Rather, you may enter an additional ReleaseEvent for it.

Now, in the real world, there is nothing like an exactly identical reissue (even the same edition may present some variants between units, depending on the plant that manufactured it for example), and the differences between reissues may be huge.

Though, we usually don't create a different release entry in the database for any of the following reasons (assuming the two releases contain the same tracks presented in the same order):

  • packaging or cover art variations
  • format difference (cassette vs. CD)
  • slightly different track times
  • minor track title variations
  • a.k.a. track title variations
  • slightly different ReleaseArtist
  • slightly different ReleaseTitle
  • promotional copies of official releases
  • "remasters" that don't present significant sound difference
  • boxsets bundles packaging together releases already available in a standalone form
  • first disc of editions containing an additional bonus disc

All these cases are detailed below with some additional information that you may find helpful to sort out non-straightforward cases.

Case by case detail

What if the cover picture/photo, or more generally the packaging is different?

Don't duplicate.

What about different formats? (Cassette vs Vinyl vs CD)?

They should not be duplicated. As a side-note, multi cassette and multi-vinyl sets should be treated the same way as multi-cd sets - each disc of the set should be filed separately in the database, following DiscNumberStyle.

But the track times are (slightly) different...

It's almost impossible to obtain a "standard" track time from a vinyl, or a cassette. Furthermore, pressings from different plants (for the very same edition) may very well have slightly different tracktimes (even for cd). Also, please remember that track times in MusicBrainz come from a number of different sources, and they may not be trusted up to 10 seconds differences (eg: homebrew cd copies may very well add/remove multiple 2 seconds gaps).

Hence, track times differences up to 10 seconds is usually not a reason to duplicate. On the other hand, though, if the difference is more than that, or if it's known that the two songs are indeed different recordings (eg: remix, single version, etc), then there should be two different releases.

What if the TrackTitles present minor variations (eg: spelling, punctuation, additional extra title information like "take")?

Don't duplicate.

What if one track has multiple (completely) different known titles in different reissues?

This cover several cases:

  • a subsequent edition corrects a (previously erroneous) title
  • a subsequent edition mistakes a (previously correct) title
  • the track is also known under an alternative title, used on some editions

Usually, this is not a reason to duplicate the release.

You may use the correct title for the track (see ArtistIntent and possibly ConsistentOriginalData for some information about how to determine that), and either mention the alternative title in the TrackTitle itself (eg: "Official Title (a.k.a. Alternative Title)"), or use the ReleaseAnnotation for that.

What if the ReleaseArtist is different?

This boils down to how different.

  • if the release artist is entirely different (eg: different person or band), then it may be worth to duplicate the release
  • on the other hand, it's usually not worth the duplication if the difference is:
    • on one version, one of the sidemen is prominently featured on the sleeve, while he is not on the other
    • on one version, the artist changes from "John Doe" to "John Doe's formation" (eg: "sextet" for example)
    • on one version, the artist is presented as a collab ("A & B") while on the other "A" is the PrimaryArtist and "B" is just featured
    • addition or omission of the artist middlename, or of an additional "nickname"
  • Last, if...
    • the artist changed his name, be it for personal or for legal (eg: trademark) reasons: this should be decided both on the artist level (eg: is it worth it to duplicate the artist entry? See SameArtistWithDifferentNames for that), and if it is duplicated, then it's up with the artist subscribers to assess the releases history and make a call.

What if the ReleaseTitle is different?

Here again, it's a matter of how much different: addition/omission of definite articles, slight spelling alterations, addition/omission of a subtitle are usually not considered a reason to duplicate.

What about promotional copies?

You shouldn't create duplicate entries for promotional copies that are identical in content to the official release.

What about remasters?

This is a touchy topic, for the reason "remastered" is not a clearly defined concept, and labels tend to print "remastered" on releases for very various and puny reasons.

Usually, you should not create duplicates for remasters, unless there are significant changes in the audio. For example, most jazz (and possibly classical) reissues advertised as "remastered" are not worth the duplication (most are nothing more than analog to digital transfers).

What if one of the edition has different/additional AdvancedRelationships that don't apply to the other editions?

That generally include cover art or liner notes credits, or possibly production credits (for the re-release).

This again is not a reason to duplicate. You may use the ReleaseAnnotation to mention to which edition such a specific credit applies.

What about BoxSets?

A BoxSet consisting exactly of a set of discs previously released in a standalone form should not be added (see BoxSetNameStyle), but the standalone discs instead. That obviously means that such a BoxSet discs may be merged into the standalone versions.

Now, please pay special attention to the facts that:

  • you should not confuse BoxSets with re-releases containing an additional BonusDisc
  • second, you should not confuse the previously mentioned BoxSets (eg: a nice packaging wrapped around a set of previously released albums or singles), with anthology boxsets (eg: like "The Complete Label Recordings of John Doe") that usually feature additional material, different tracks split across discs (even if such an anthology happens to actually duplicate one release as one of its discs).

What about 2 discs editions that contain a BonusDisc?

Titling of such editions is addressed by BonusDisc and BonusDiscStyle. The "main" disc of such editions should not duplicate the standalone edition (assuming of course their content is identical).

Other non-duplication cases

There are even cases where there really is a substantial difference between two releases (eg: really different track length), but still the artists subscribers decided not to duplicate for a number of reasons. Ultimately, as stressed earlier, it's in the end a matter of common sense.

But are we going to lose all these fine grain differences between editions?

No, it's not lost. You should use ReleaseAnnotations to store any relevant details (alternate spellings, edition specific information, etc).

When NOT to merge, and keep "duplicates"

On the other hand, in the following cases, different releases should be kept separate.

Obviously, there should be multiple different entries in the database if:

  • the two releases have a different number of tracks
  • the two releases present different tracks ordering
  • the two releases, despite having the same content, are split differently in multiple discs (this usually happens when a 2xLP vinyls set - which should be filed in the database as a 2 discs release - is reissued as a one disc cd).
  • one release had its tracks titles translated/transliterated, be it as an official release, or as a MusicBrainz VirtualDuplicateReleases
  • the track time of one of the tracks differ by more than 10 seconds, or it is known to be a different recording (remix, single version, re-recording, different master of the same event, etc...)
  • the entire release is a re-recording of the other
  • "clean" versus "uncensored" song-versions

But why not duplicate everything after all?

Our approach is different than Discogs': we don't keep one separate entry for each and every reprint, but rather one different entry per different track-set (eg: different recordings, or same recordings presented in a different order).

That approach being at the heart of both our current data model and tagging service, trying to duplicate anyhow will have several bad impacts:

  • degradation of the overall data quality by making the maintenance task much harder
  • loss of data consistency
  • severe degradation of the tagging service quality: PUIDs and DiscIDs will get randomly attached to the "wrong" edition, making the (supposedly) different releases just a mess of wrongly attributed fingerprints and TOCs
  • increased discrepancy in documentation (eg: ARs) between the various "reprints"
  • un-usability of the current UI (and probably data model), due to the huge increase in material quantity

Relevant readings