Eric Dolphy, cleanup of Summer 2023

Rationale

What is it?

Started early July 2023, an effort to cleanup the discography of Eric Dolphy on a large scale has been initiated (and to some extent, related musicians and labels - McIntyre, Booker Little, New Jazz, etcetera).

Why?

The main issues motivating this effort were:

large amount of recordings that were actually duplicates (probably in the range of 70%)
poor to very poor general state of affairs for live performances
largely bogus information being replicated from Discogs, or sleeves, without appropriate research - specifically performance AR
large number of poor album artist and track artist credits, including many bogus artists

Hoping for what?

The purpose of this effort is double:

significantly increase the point in time quality of our discographic information, to the level of Simosko / Alan Saul at least
widely use disambiguation fields on every entity, specifically recordings, to help future users make new release additions from existing recordings, and future maintainers to cleanup low quality new entries, to foster a state of the database where high quality is easier to maintain moving forward

Who is involved?

A small group of editors, either involved formally (like the author of this page), or just informally.

We regularly see edits from riwood, highstrung, dmppanda and others (<- add your name) - and many sharp but supportive voting from the likes of drsaunde, chaban, salo.rock, phanboy_iv, CatCat, MrH, etc.

What is in scope?

Everything that involved Dolphy in any way, shape or form.

This extends to VA compilations, and orphaned titles credited to other artists.

No exception, no prisoner :-).

How long is this going to last?

As long as it takes to get there.

Very likely until end of October 2023.

Interesting things happening

Normalizing informal use of recordings disambiguation (inc. matrix numbers)

Ongoing discussion:

https://wiki.musicbrainz.org/Talk:Recording_Disambiguation

Current agreement:

https://wiki.musicbrainz.org/Recording_Disambiguation

What do we do with crazy garbage like that?

https://musicbrainz.org/edit/102432588

Informal tasks and status

Release groups:

[x] no duplicate
[x] no over-merge
[x] have meaningful disambiguation
[ ] allmusic links
[ ] discogs links

Releases:

[ ] no duplicate
[ ] have meaningful disambiguation
[ ] discogs links
[ ] has annotation modeled after template (see below)
[ ] release artist credit respect sleeve
[ ] track artist credit is the same as release artist, unless there is a strong reason not to

Recordings:

[ ] no duplicate
[ ] have meaningful disambiguation following format of https://wiki.musicbrainz.org/Recording_Disambiguation
[ ] recoding artist reflects the session leader without any of the fancies of specific releases
[ ] are all linked to a work

Works:

[ ] no duplicate
[ ] have composer credit

ARs:

[ ] go through every releases credited to Dolphy one by one and make sure each edition has the appropriate release level credits (producers, remaster info, liner notes, artwork)
[ ] go through every releases credited to Dolphy one by one and make sure all corresponding recordings have the appropriate track level credits (recording engineer, performers, work)
[ ] go through every recordings that Dolphy ever participated in on other people releases and make sure they have appropriate track level credits (same ^)

Release group annotation proposed template (work in progress)

Every part is optional. This is just meant as a guideline to harmonize the way we present similar information.

== Wanted / Issues ==

[write here any significant current problem or missing data that is hard to solve and could use help from someone who would own the release]

== Listening notes ==

[if you have notes to share from your listening of the releases, put them here - about who is playing what instrument, who solos when, who is sitting out tracks]

== Reissues history ==

=== Tracks ===

[info about tracks selection over original recording session and subsequent reissues, alternate takes, etc]

=== Credit ===

[info about how this has been credited over time]

=== Spelling ===

[info about common mispellings or mistitlings]

=== Cover art ===

[info about cover art over issues]


== Remasters history ==

The normal New Jazz trajectory is:
1. originally recorded by Van Gelder in his studio
 * quality is good, because of RVG, but this is obviously mastered in the 60s
2. Original Jazz Classics reissues it first on CD during the late 80s or early 90s with a remaster done in Fantasy Studios, generally by Phil De Lancie
 * generally very decent albeit a bit cold - it is more a small cleanup / update to the originals
3. There generally is a handful of limited edition / special releases for the Japanese market with a distinct Victor remaster
 * typically uses K2/20bit/24bit
4. RVG remaster then come out with a significant upgrade, available in CD and digital
 * these come with significant differences compared to the original mastering and are generally vastly superior to other prior masterings and are typically the gold standard if you go digital
5. occasionally audiophile vinyls come out, typically from Analogue Productions, with their own remastering
 * these are sounding very different, and very good - they are typically the gold standard if you go analog

[release group specific informations]

== Verbatim ==

=== Simosko ===

[any verbatim from Simosko that is relevant as source to backup editing decisions]

=== Saul ===

[any verbatim from Saul that is relevant as source to backup editing decisions]


=== Original liner notes, XXX ===

[any verbatim from original liner notes that is relevant as source to backup editing decisions]


=== [Reissue additional liner notes, XXXX] ===

[any verbatim from reissues liner notes that is relevant as source to backup editing decisions]

Useful resources

Books:

Online:

Alan Saul:
- http://adale.org/Discographies/EDIntro.html
- http://adale.org/Discographies/Booker.html
London Jazz Collector https://londonjazzcollector.wordpress.com/
Onttonen: https://mingus.onttonen.info/leader.html

Editing philosophy

Listening matters!

Credits on sleeves are always problematic:

at best they are loose - "Eric Dolphy alto sax, bass clarinet, flute" never means he is playing all three instruments on every track, or that he is even playing on every track
quite often they are just plain bogus - bootlegs, low quality reissues, or even legit releases are either scarce on details, or float completely wrong credits, mis-titling, wrong dates and location...

How do you figure these situations out?

You listen.

MusicBrainz is and has always been about music (it is in the name, right?).

We tend to forget that and think that editors are mere bookkeepers moving bits of texts between systems and databases with varying degrees of efficiency, getting street cred for quantity over quality.

This part of the database cannot get better with that ^. So, listen to the tracks you are editing. Every edition of it if you have to. It is 2023. Almost everything from the 60s is on Qobuz, Spotify, YouTube, or "other" online properties. You do not need sleeves to document credits, or identify a mis-titled track.

Discographic research != copypasta

A large amount of information has been simply copy-pasted from Discogs, or RIM, or Wikipedia.

This barely makes any sense

if Discogs (or other) is absolutely superior, why exactly are we staying here and why don't we all move over there? (hint: they are not)
if it is not, why should we turn MusicBrainz into a manually maintained replica of their information?

There is nothing inherently wrong with Discogs, and there is certainly some valuable information there.

There is also a large amount of bulls***, poorly documented stuff, ridiculous organization decisions, unsourced (bogus) information, mistaken production country for release country, or recording dates for release dates, and a ludicrous amount of partial duplication.

Bottom-line: look for scans on Discogs, but challenge almost everything else. Like with every other online resource that does not clearly say from where they are sourcing information.

How do you do better than Discogs?

read Simosko. Read Alan Saul. Read old Billboard, CashBox. Read dedicated blogs. Read other related biographies (Like a Human Voice, or Mingus, or The Coltrane Reference).
train yourself in identifying releases, and learn about labels histories
- spindle ring on one side? deep groove? purple label? AB etching? this is an Abbey pressing
  - it is likely original, from very late 50s or very early 60s
- you have a New Jazz release?
  - original is definitely from late 50s to 1965, recorded by Van Gelder, probably produced by Esmond Edwards - it has very likely then been reissued by OJC under Fantasy in the late 80s or early 90s, usually remastered by Phil De Lancie at Fantasy studios - it has also very likely been issued in Japan by Victor in limited editions with their own remaster - then as part of the RVG remasters collection with a significantly different remaster - and there is usually an Analogue Productions audiophile vinyl hanging around
cross-reference
- you cannot have a nightly gig in New York one day and another one the next evening in Stockholm (the rules of physics, border control, live music, and time travel prevent it). If somebody does, the dates are wrong.
- labels usually issue their releases in sequence - if some catalog number is out of whack, it happens, but then it is worth investigating
do no trust things you cannot independently verify unless they have a solid reputation for accuracy. With 60s Jazz, our personal experience has been:
- RIM is generally just wrong
- AllMusic is a big mess
- Wikipedia *sometimes* has accurate information, and at least provides sources, but they generally half ass it, or copy paste from JazzDisco (see below)
- Discogs is definitely not to be trusted at face value
- jazzdisco.org is regularly wrong or incomplete
- Reinhardt is good, but clearly outdated
- Simosko is very solid, but is missing some of the latest discoveries
- Alan Saul is almost always right
- The Coltrane Reference is top of the class
do not trust any label that is not a household name and that does not provide master information - triple check whatever these labels are saying

Be practical, be realistic

... with recordings...

In theory, it would be good if:

we could have different recordings for mono vs. stereo
we could have different recordings for different remasters
we could have different recordings for live recordings that have the clapping / chatter / warm-up cut out

In practice:

every time you create one of these, you do make the cost of maintaining / editing ARs X times higher
normal users confronted with different recordings for these cases never figure out the right one - when they pick an existing recording at all (which usually they do not...) - even with the most well intentioned disambiguation bit, and you end-up with distinct recordings being referenced by the wrong releases, completely defeating your original goal

Experience has demonstrated that it is simply not practical, a pipe dream, and not bringing any value for anyone to distinguish these recordings, at least as long as MusicBrainz recordings overall are about ~70% duplication otherwise...

https://musicbrainz.org/edit/102417842

https://musicbrainz.org/edit/102246973

... with releases...

In the same train of thoughts, some create separate releases based solely on the factory used to manufacture the physical CD.

This may be a valid argument for vinyls, where there are noticeable sound differences dependent on pressing, and a significant crowd of audiophile buying QRP (including me).

But for CDs, the benefit of such a piece of information is highly dubious - and its impact on sound quality belongs to the same category as "using a cat7 vs a cat5 ethernet cable to enhance webpage colors on your laptop".

Similarly to above, this will create additional maintenance work for release level ARs, and confusion and clutter for users trying to find the release they want to tag, with almost no upside whatsoever.

https://musicbrainz.org/release-group/bc6c2573-bc7d-346a-8880-0e34b8fb7e50

... with works...

The same is true as well for works. Splitting works because lyrics have been added later is not helping anyone, or adding value, but is the guarantee to create a large scale mess:

https://musicbrainz.org/edit/102532636

.... with artists...

In the overwhelming majority of cases, "Artist Quintet/Sextet/Quartet/Septet" does not refer to a defined band, stable group, or artistic project in any way shape or form. This is usually just information about the number of sidemen involved in the recording, given the buyer a sense of what type of music they are buying.

Such information is then routinely discarded or propped-up by release labels on subsequent releases, depending on what they think will entice customers. The same release could be put out under a completely different artist name who was just a sideman, but happened to become more famous decades later.

How to handle this is delicate and requires some case by case assessment.

being faithful to sleeves is paramount in our opinion - if you deviate, people are going to create duplicates - that is why we have "credited as" on album and track artists for
it still does not mean you should create these gazillions fantasy "bands"
it also still does not mean you should collapse every formation under their leader - there are legit bands in jazz going by a formation name - their identity is defined, there is an artistic project associated with it, that is distinct from its leader

Just use good judgement, leverage your deep jazz knowledge, and a little bit of common sense.

Nobody can seriously believe this below makes any sense:

https://musicbrainz.org/search?query=duke+ellington&type=artist&limit=25&method=indexed

.... with release groups...

Historical jazz in the real world overall is now a mess. A lot of stuff has fallen into public domain, attracting a large number of shady people hoping to make a buck without accountability by just recording an old vinyl and slapping that on a CD, or online.

This has created a situation where a large number of release groups are completely muddying information about an artist, while they are all essentially the same original vinyl being ripped over and over with made up titles.

Keeping these separate serves little purpose.

The murkier the copyright is, the worse the situation is, evidently. Live are of course the worst - scene releases, unofficial "trees"...

On the other hand, there is significant historical and emotional value in keeping certain things separate. Live at the Five Spot has been issued during Dolphy lifetime in separate volumes that did acquire individual identities, although all of it has been recorded live in one night. In that case, keeping them separate makes more sense.

Bottom-line?

We all want database and style purity, and incredibly refined qualitative information everywhere.

Now, go out there and go touch some grass.

This is the real world. NGS does smell funny in a lot of places (take it from someone who has been here for 17 years). And overall the quality of our db is quite bad in the jazz department. And some of our official style guidelines are quite often just divorced from practical reality.

We are not advocating going rogue and splintering style or community - but that we do not drop the prey for the shadow. Style is in the service of quality. Quality should not be sacrificed in the name of obedience.

If some guideline is getting in the way of quality AND practicality, deviating from it should be an option worth considering.

Also, be creative. There are free form metadata you can leverage when structured data models are out of touch. Use disambiguation generously. Use annotations. Some stuff is not worth trying to bend a broken data model into an untenable position while you can fit that info just there.

Finally, ask for help - ask for opinions and votes if in doubt. Let's look at issues together and find practical solutions.

Jazz/Dolphy

Contents