notsosmart: it's a popular song so there are lots of recordings, this isn't unusual
many of them could probably be merged, but lots of editors aren't aware of how recordings work in MB, or want to spend the time to determine if a song matches an existing recording
Lotheric
Should I use an artist entity for the relationship ([video recording] was directed by 16pads) when the a videoclip was credited as being directed by a production company (https://www.sixteenpads.com/about) ?
notsosmart
atj: could I just assign all of them to one of the 5-7 recordings that actually exist and have ISRC? That would be judging just by the time, but I think it's hell lot better than 99 recordings
I have a strange feeling that it's just one person :D
Lotheric
I *could* use Patrick Antoniewicz but can't be sure if he's the only director associated with the company or just the main one
yeah me too
trying to find more info on the web
:)
notsosmart
someone might correct me on this, but I think the approach is not to add companies like that to MB, as it goes beyond music-related information -- it is not a label, so not a relevant company
you can write this in Annotation
atj
notsosmart: there are recordings with duplicate ISRCs and lengths that differ by over a minute
you've also got about 20 recordings that are part of DJ mixes
notsosmart
right, but dupliacate ISRC are quite easy to deal with
I'll merge them first
atj
I've just created a big merge for the main recordings
*recording
based on +/- 2 seconds of 4:43
notsosmart
nice
about the DJ mix issue... maybe there can be one recording without ISRC, sort of "unknown recording", and merge those mixes there?
atj
no
learn the rules :P
how recordings work and how you want them to work are different
DJ mixed recordings fall under "Recordings of different durations can be merged, as long as there is no evidence to suggest that differences in mixing or editing have caused the change in lengths"
e.g. the difference in length is due to the mixing
notsosmart
I see
elomatreb[m]
I don't think that's the common consensus
It is very common to have separate releases for DJ-mixes
atj
elomatreb[m]: you misunderstand me
I was quoting the rule that indicates why DJ-mixed recordings are separate
notsosmart
no no, atj is right
elomatreb[m]
ah sorry, I misunderstood
notsosmart
in that case those recordings should have "DJ mix" relationship with one of the recordings that the DJ has used as source
atj
elomatreb[m]: sorry, my wording was a confusing
Lotheric
and then you have explicit vs clean recordings
elomatreb[m]
I thought you were saying to merge DJ-mix recordings with their non-mixed counterparts because it's just a mixing difference
notsosmart
alright, but what I mean is that there is a set of original recordings, and all those other recordings are in some relation to those base recordings, or to each other etc.
atj
yes, but in order to create the relationships you need a canonical recording
which is rarely self-evident
elomatreb[m]
Ideally that would be the case, but determining which recording is which is very difficult outside of trivial cases
notsosmart
exactly
elomatreb[m]
doing it properly would require having all the releases at hand and actually listening
atj
and this is why we end up with 50 recordings
notsosmart
no I don't think that's good approach
elomatreb[m]
It's better than any alternative
atj
it's the only proper approach
notsosmart
let me explain why I think it is not
elomatreb[m]
for recordings with known provenance, you get full and reliable information, without claiming a false level of accuracy for recordings where it's not known
Remember that there are also Works, which can properly link together the separate recordings even if you don't know the full details
notsosmart
first of all, I think it is not any approach, or any alternative -- it is physically impossible to listen to all of these mixes and to learn which DJ used which source recording. So what you are proposing is not actually possible.
so currently there is no alternative, those unknown recordings will be there, but this should be addressed
some change in schema, like somehow marking the recording as unknown, or filtering them out
elomatreb[m]
there are very different levels of "known", I don't think this is something you can quantify in the database
atj
notsosmart: I suggest you learn about works, they achieve what you want I think
notsosmart
I know about works, and indeed there is a much better situation there
elomatreb[m]
e.g. a recording which nobody knows what it sounds like (e.g. if it's DJ-mixed or not), a recording where it's known to be live but not from which concert (date/location), etc
notsosmart
but it should also be with recordings
atj
well, I suggest you post on the forum and try to build some sort of consensus
but keep in mind that you're not the first person to think about this
elomatreb[m]
The situation with tons of duplicated/unknown recordings is also dramatically worse for "older" artists or for really popular ones where there are tons of compilation releases compared contemporary artists where it's easier to confirm two digital recordings are the samew
notsosmart
elomareb[m]: `there are very different levels of "known", I don't think this is something you can quantify in the database` -- in every database there is always "other"
elomatreb[m]
"Other" is not a useful qualifier though
notsosmart
why?
elomatreb[m]
why would it be?
notsosmart
useful to indicate that it is not the one enumerated or accounted for
elomatreb[m]
if the data about a recording is known, it should be added. If not, it can't be, so I don't see how that would be useful
Compare to the "Data Quality" field which we have on releases mostly for historical reasons, which basically nobody uses
notsosmart
elomatreb: `if the data about a recording is known, it should be added. If not, it can't be, so I don't see how that would be useful` - I see your point
but "DJ mix of an unknown" recording is a common practice in music industry, and should be addressed. For example, in case of Patrice Rushen, there is a lot of DJs who included some recording of her track on their mixes, but Patrice Rushen has no control over it whatsoever
and when looking at her recordings, I see just a bunch of dudes mixing her stuff, instead of her recordings
covers I can easily filter out, problem solved there
but for dj mixes? I can make rec-rec relations if I know the source recording
but if I don't, it's not possible
so why not just an attribute? for a DJ mix?
just like for cover, live, etc
then we would need a filters and/or groups on an artist recording page, like for releases there are not only groupings (albums, EP, singels, other, etc.) but also an option to show/hide official/unofficial/various artists groups
fhe has quit
skelly37 joined the channel
OK so a different issue but same song... a user created a release "101 Disco Anthems" (a compilation) and added a recording of "Forget Me Nots", and shortly added ISRC USEE10183258, ripped from CD
but the recording has length 3:56 instead of 4:45, so it must be either an edit of that recording, or a mistake on the CD
should I make "edit of" relationship without actually hearing the track, or create another recording, because it is unknown?
notsosmart: I'd say don't add "guessed" relationships like that
Linking both recordings to the appropriate work should be enough
notsosmart
think so too
so is the ISRC unknown for this recording?
elomatreb[m]
If the ISRC is ripped from the CD, it should go on to that recording
It's possible to have recordings with multiple ISRCs as well as ISRCs that belong to multiple recordings, since the concepts used to assign ISRCs don't necessarily match the MB recording concepts
notsosmart
ok
so both recordings with same ISRC can be linked with relation "edit of"?
s/both/two/
elomatreb[m]
They could, but I would advise against it unless you listened to both
They should both be "recording of [work]"
If in doubt, duplication is preferable to overzealous merging (since untangling an incorrect merge is really annoying)