[mb-users] Interpreting FeaturingArtistStyle

[Olivier]

2007/2/6, Chris Bransden <chris at whenironsattack.com>:
> how do you find that out, though? take the chemical brothers example I
> gave before - http://www.discogs.com/release/24389 - all these vocal
> credits are big name people. you can be sure that the chemicals knew
> exactly who they wanted on these songs, and invited them in specially.
> hell, they may have even co-wrote the songs. however, they are still
> only given a liner note credit (no feat.). are you saying we should
> promote these to featuring credits, based on their real input? as this
> seems to be the way you define 'feat' in jazz.

Frankly, I have no idea about that specific case (though I like to
hear Chemical Brothers from time to time) :)
I don't have a specific idea that "solves-all" in all areas.
I tend to think our current styleguides (especially "feat"), and
approach (trusting sleeves) *are* sane and good - but (maybe) in a few
specific case corner (Chemical Bros??), and for a good part of jazz
(given what I previously said).

>
> plus it's also near impossible to say how much artist intent is really
> on a sleeve.

Ah! That old song! :D

You play safe - I don't agree :)
In my "area" of editing, I most time have a very good idea of how much
input had the artists both in the production conditions of the
releases, and control over the final product including sleeve.

> you can easily say that Artist A being featured was a
> 'marketing decision', but you don't know that. some artists have
> complete control over sleeves, some have none, and most (in my
> experience) are somewhere in between.

I'm not saying I've a solution for all, or that we know all in all
circumstances either...
I barely would like to address some concrete problems arisen in a
small corner of the db ;)

> if we assume nothing, we can never be 'wrong' :)

<tease>
and can never be right either. :)
</tease>

> this is fair enough. however i think if you're going to go down this
> route, then you can't use 'feat.' AT ALL in jazz, as the meaning is
> evidently not the same as in the rest of the database. instead you'd
> have to either remove them entirely, i think. else how do you decide?

I'm not at ease with the term "jazz" taken in a "genre" way, or as a
well definite entity...

Anyhow, there are IMO legit feats in jazz, as well as legit groups
being named "John Doe Quartet".

I'm sorry and I wish I would have a better answer to provide you:
There's no way you can tell that "Jane Doe Quartet" is not a group,
and "John Doe Quartet" is a group, unless you have knowledge about the
stuff (familiarity with the artist work, biography, and condition
productions of the release).
You *won't be able* to tell from sleeve.
Identically, you won't be able to tell from sleeve if X is a sideman,
or a featuring artist.
On the other hand, by checking all covers of all the different
editions, by reading liner notes and discographic data about session
details and stuff, plus by hearing the stuff, *you'll* be able to make
a pretty documented and reasonably objective assumption.

As for getting rid entirely of "feats", well... why not? :D

About "getting down that route": I think we are already down the
route. I said it previously - I think people editing in these area are
not adding such feats.

> and what meaning would the word "featuring" have if artists who are on
> the cover are not "featured"?

Attempt at it:
Being in the artist intention and/or production intention to have a
specific other artist feature in a given recording session, and have
that fact advertised.

Which at least saves me from the (IMO improperly deduced in the case
of jass): "if he's on the front sleeve, then he's featuring". :)

> > > plus it's a contradiction in terms. an artist can't 'feature' on a
> > > release if they are not on the cover or the track listing, and not
> > > 'feature' on a release if they are.
> >
> > Sorry, that's a contradiction in terms, but that's the way it *exists*.
> >
> > Check:
> > http://ayler.org/albert/assets/images/Debutgh1.jpg
> >
> > which by your reasoning (if I understood well) is a release by
> > Albert Ayler Quartet, featuring Cherry, Peacock, Murray
> > (and this is utterly wrong! Cherry, Peacock, Murray, Ayler *are* the
> > Ayler Quartet - whatever the meaning of that group name)
>
> right - and check http://www.discogs.com/release/619434 - Lou Reed is
> the founding member, and (at the time) frontman for the VU, yet he is
> still given seperate (featuring, essentially) credit.

"Albert Ayler Quartet" is *not* a group... :/
I'm sorry to *hammer* that... but this is (partly) the main source of
our problems here.

>
> it DOES make sense - he was at that point, more famous as a solo
> artist than his old group was, so they put his name on the cover to
> sell more records. however, this also follows that someone looking at
> this release on their iPod/MBz would expect to see his name, as that's
> what they have on their record sleeve, so we have to reflect this.

Well, we still flirt with tagging questions... sure MB is used for
that... but not only.
IIRC you agreed that blatantly wrong information should be
corrected... and we correct blatant typos and fix style also, don't
we?
While sure we need to take tagging into account, I don't think we
should base our style decision (entirely) on people "search
expectation" (which is an entirely subjective topic btw...)...

> no, like i said before, there shouldn't be one MBz release per feat.
> configuration. there's already good rules as to what gets merged and
> what doesn't and I don't think feat. configurations have ever been
> used as reason to keep things split (this to change when NGS arrives
> but then so does everything).

So? You still don't answer me: which one of these would you keep? How
would you solve conflicts between people who think their edition
should be dominant?
Please apply now these "good rules" on this very example ;)

>
> if 2 seperate MBz releases (ie, maybe 1 has bonus tracks) have
> different feat. configurations for their common material...then so
> what? it might be slightly misleading if you compare the 2 ("why's he
> featured on this but not that??"), but surely the comparison between
> the physical copy (sleeve) and the MBz copy is more important than MBz
> release to MBz release?

Mmm. I like consistency actually. *I'm* dissatisfied if I have the
same tracks on two releases with different credits/feats... be it from
MB or from physical sleeves - in the later case I usually dig data
until I sort it out for myself, and makes a mental note that edition X
has partial/erroneous data...

As I said, I'm working mostly on quite old stuff actually, in a region
of the db where people value (and favor) discographic exactitude above
similarity with physical media. Again, I'm not trying to make this the
religion for all... rather articulating some facts from there :)

>
> > I don't believe (at all) in exactly reproducing cover sleeves, and I
> > don't have the impression "other" music databases do either.
>
> discogs pretty much matches sleeves exactly now, thanks to ANV (which
> is a grrrreeat system and we should steal it :P)

Lol.
Fine for them ;)
I for one don't think we should follow this road *for everything*.
I'm pretty pleased to find consistency/accuracy in MuscBrainz when it
lacks in physical sleeves.

> but you say that jazz releases should have featuring artists on
> occasion, but you don't define what those occasions are? if you aren't

I hope I did (an attempt) just above

> using the cover, then surely the only thing left is (perceived)
> musical involvement.

Well. Perceived, yes. To read a sleeve or a book, you need to "perceive", right?
But subjective, no. I took at heart myself to be as objective as
humanly possible...
And I think a biography relating production details is (at least) as
factual and objective as a cover sleeve - if yourself have in mind to
reflect facts, and not your own musical tastes - which I hope I do...

> old pop duo from the 80s (?) who never sang any of their own songs -
> rather they got someone better to do it, and then danced around with
> their shirts off in the videos :)

Cool stuff!
I'll check that out :D

> well, presumably that's a problem your method faces as well! as you

Sure, there are problems every way ;)

> anyways, to answer: if it's on the cover as "featuring X" then it's
> easy. if it's on the tracklisting as "Song (featuring X)" then that's
> easy to.

Agreed in most cases. But not *all cases* (see my Ayler example above).

> that accounts for most examples. the tricky ones are "X with
> Y" when it's not clear whether it's a collab or a featuring, but
> generally you can easily tell if you know the artist(s).

Agreed again, in most cases. As you said, knowledge is important here...

>
> i still don't see where this fits in with the feat stuff. featuring
> someone doesn't create a new artist (well, aside from the person
> themselves if you add a featuring AR, but there's no harm in that,
> surely?)

T'was a more generic remark, as an objection to the "you should always
use what the sleeve give as a release artist".

>
> hot topic!!! :P
>

Interesting also :)


Hey! Other people sitting around!
Come in and share if you feel like! I'm not biting! And Chris neither
(well, I guess)!


- Olivier