"recording" ≠ "performance" (Page 2 of 4)

If you want a clear distinction, I think the only objective measure is ISWC and ISRC. If the thing has it's own ISWC then it is a work. This would mean AFAIK that NO remix would be a work. However...

ym wrote:

I think problem with popular music is that people don't always write down scores or lyrics (like sampling, remixing etc..), but this doesn't mean they are not creating music and if the work object denotes a "musical work", IMO, each of those should be treated as a work, either as derived or original.

...suggests you would not agree with this. Do you think every remix should be a unique work?

The term “musical work” covers both new musical compositions and the repertoire of existing musical compositions (from songs to symphonies to jingles), regardless of the medium (material or otherwise) in which they are comprised.
With regard to the origin of work in MB, I agree that it has been largely influenced by classical music but I think it applies equally to a lot of popular music as well, at least in relation to cover versions.

The problem is that if you say all remixes are unique works because they change the music then you also have to say all cover versions are unique as each artist/band is going to play different music to the original, whether it be only slightly or dramatically different. For example, The Byrds cover of Bob Dylan's 'Mr Tambourine Man' sounds very different to the original, but retains the original melody and lyrics.

So if you want to draw a firm line, where would you draw it? You can't simply say that if something 'sounds different' it is a unique work because this is a subjective opinion.

I think it's more instructive to go to http://iswcnet.cisac.org and see how they actually classify things.

Lixobix wrote:

If you want a clear distinction, I think the only objective measure is ISWC and ISRC. If the thing has it's own ISWC then it is a work. This would mean AFAIK that NO remix would be a work. However...
I think problem with popular music is that people don't always write down scores or lyrics (like sampling, remixing etc..), but this doesn't mean they are not creating music and if the work object denotes a "musical work", IMO, each of those should be treated as a work, either as derived or original.
...suggests you would not agree with this. Do you think every remix should be a unique work?

I am not sure how ISWC and ISRC is dealing with derived works, but looking at:

http://www.bmi.com/news/entry/233038 wrote:

The term “musical work” covers both new musical compositions and the repertoire of existing musical compositions (from songs to symphonies to jingles), regardless of the medium (material or otherwise) in which they are comprised.
With regard to the origin of work in MB, I agree that it has been largely influenced by classical music but I think it applies equally to a lot of popular music as well, at least in relation to cover versions.

I think they also think remixes and covers are works too (even jingles!).

The descriptive metadata for an ISWC includes:
the title of the work
all composers, authors and arrangers of the work identified by their IPI numbers and role codes
the work classification code (from the CIS standards list)
in the case of 'versions', for example arrangements, identification of the work from which the version was made.

As I stated earlier, having another database's identifier doesn't mean we have or should have one-to-one relationship with those databases, but opposite is true too, if MB says that the work object is in a one-to-one relationship with the ISWC, that would mean a line too.

Lixobix wrote:

The problem is that if you say all remixes are unique works because they change the music then you also have to say all cover versions are unique as each artist/band is going to play different music to the original, whether it be only slightly or dramatically different. For example, The Byrds cover of Bob Dylan's 'Mr Tambourine Man' sounds very different to the original, but retains the original melody and lyrics.

covers don't change any music or lyrics generally. The artist plays the same music with different style, so performance and also recorded audio changes, but not the music itself. Of course I am assuming covers don't change the original musical material.

Lixobix wrote:

For example, The Byrds cover of Bob Dylan's 'Mr Tambourine Man' sounds very different to the original, but retains the original melody and lyrics.

I think you should check out http://iswcnet.cisac.org/iswcnet/MWI/re … geNumber=2
That pages might give an idea how they deal such examples.

edit: Another interesting example is: http://iswcnet.cisac.org/iswcnet/MWI/re … geNumber=1
searched for Beethove - Symphony no 9
Please take a look how different artists are credited as arranger and author, but only beethoven is credited as composer and still they have different ISWCs.

Lixobix wrote:

So if you want to draw a firm line, where would you draw it? You can't simply say that if something 'sounds different' it is a unique work because this is a subjective opinion.

How can a work "sound different" they don't have audio! They are intellectual and non audio things, aren't they? BTW, I see a work can be "a cover of" another work. How can this be?

There *are* a few remixes which get their own ISWC. But it's pretty rare I think.

Lixobix wrote:

Or the lyrics Bob Dylan writes on his typewriter, before they are ever sung.

Isn't it illegal to imagine Bob Dylan typing? I think we're supposed to picture him writing his lyrics in an old notebook he carries around in his pocket.  Or just making them up in his head and never writing them down at all; someone else can do that for him :-)

My point is that in real life, decisions about how to classify works are not always 100% rational and informed.  My personal preference is for MusicBrainz not to prescribe how we think those decisions should have been made in an ideal world, but to describe the reality of how they actually were made, even when it's messy.

ym wrote:

practik wrote:

Where do you see that?  It's not here anywhere that I can see:  http://musicbrainz.org/relationships/work-work

I meant a recording, http://musicbrainz.org/relationships/recording-work , which says that a recording can be a cover of work. how can you cover something not performed yet?

The term 'cover' means performance of material not written by the artist. But it is used in reference to artists who generally do write their own material, as opposed to those who usually perform material written by others. Either way, the work does not need to be performed by the writer first, but it generally is.

ym wrote:

practik wrote:

My point is that in real life, decisions about how to classify works are not always 100% rational and informed.  My personal preference is for MusicBrainz not to prescribe how we think those decisions should have been made in an ideal world, but to describe the reality of how they actually were made, even when it's messy.

Indeed, that's why I am objecting that subjective things shouldn't be entered in the database, thus, instead of having arguments about what is a work or not we should be following authorities and entering factual data. I don't thing acting like an authority is a wise thing to do.

By factual data, do you mean using ISWC/ISRC as definitive evidence?

http://www.iswc.org/en/faq.html wrote:

The descriptive metadata for an ISWC includes:
    the title of the work
    all composers, authors and arrangers of the work identified by their IPI numbers and role codes
    the work classification code (from the CIS standards list)
    in the case of 'versions', for example arrangements, identification of the work from which the version was made.

So we are talking about composers, authors and arrangers. The question remains, who is an arranger?

"...of a contemporary OR previously recorded [song]" http://en.wikipedia.org/wiki/Cover_version

A contemporary song need not have been performed.

If your definition of arranger is wide enough to ALWAYS include remixers, then you conflict with ISWC. Which means you have to either ignore ISWCs completely or you accept them as authoritative where they exist, which in many cases would mean remixers are not arrangers.

Which types of musical works can receive an ISWC?
An ISWC may be assigned to any musical work, published or unpublished, newly created or already existing as: Dramatico-musical work-Musical arrangement of a work Adaptation of the lyrics of a work-Translation of the lyrics of a work Recognised excerpt of a work-Medley-Potpourri etc...

Should , adaptations and translations get an ISWC ?
Musical arrangements, adaptations of lyrics and translations must receive their own unique ISWC numbers. These ISWC numbers are usually allocated by the agency which administrates the works of the arranger and/or adapter. The connection between the 'version' and the original work is indicated in the descriptive metadata of the ISWC.

May excerpts from other works be numbered too?
All works should be identified in their own rights; for example, an aria from an opera or a cadenza from a concerto can receive an ISWC. Similar to the 'versions', the relationship between excerpts should be indicated in the descriptive metadata of the ISWC.

May works or 'versions' which breach copyright obtain an ISWC?
Surprisingly enough, the answer is "yes". The non-authorised arrangements of a musical work can be identified, if only to ensure that they will be recognised at an international level as works infringing the copyrights of others.

'Contemporary' means 'of that time'. So the 'Mr Tambourine Man' was a contemporary song in the 1960's.

My point is that the 'or' in the definition means that the song has not necessarily been recorded. So you could cover an as-yet unrecorded work

'Contemporary' is slightly misleading IMO, because a work may have been written but never recorded by the author, but decades later another artist may record a cover version.

I think there are two approaches to this.

First, ISWC treats remixes and arrangements as different things. That's a clear contradiction to your statement.

Secondly, the two are treated as the same thing and assessed on the same criteria. In that case you still take each case on it's own merit. You judge both on the same criteria. I'm not sure what those criteria are. However, it seems that in practice a new arrangement is more likely to be considered a unique work than a remix, on a balance of numbers.

As a general point, I suggest that MB should accept what ISWC says, if the information is available. Where it isn't, I don't know how we can write better guidelines until I've seen their criteria.

ym wrote:

Lixobix wrote:

'Contemporary' means 'of that time'. So the 'Mr Tambourine Man' was a contemporary song in the 1960's.

My point is that the 'or' in the definition means that the song has not necessarily been recorded. So you could cover an as-yet unrecorded work

'Contemporary' is slightly misleading IMO, because a work may have been written but never recorded by the author, but decades later another artist may record a cover version.

Since it will be the first version, I don't understand how it can be a cover!

Because you stick to your definition of "cover=previously recorded". The "or" means that (according to wikipedia) if you consider the set of previously recorded works, some covers are not parts of that set. In other words, wikipedia disagrees with your definition. Or you misunderstood their definition. English is not my language, but I understand wikipedia's definition to mean: cover = {performance of contemporary song by performer who is not the author} + {performance of a song already recorded by someone else}

BTW, wikipedia's definition has a little problem: if a song is performed by A, then by B, then again by A, A's second performance would be considered as a cover according to this definition :-) Of course, a little common sense allows us to eliminate this kind of situation, but this shows that such definitions should not be taken too strictly.

a cover is a performance of a work not written by the artist. It does not matter whether the author has recorded it before the cover performance, so the first recording can be a cover.

Here's an example. The song (work) 'I'll Keep It with Mine' was written by Bob Dylan. The first recording was by Judy Collins, and this was a cover version. Dylan later recorded several versions of the song, although none were released until years later. Other cover versions were recorded by Nico, amongst others.

http://en.wikipedia.org/wiki/I%27ll_Keep_It_with_Mine

The song was clearly written by Bob Dylan, but the first recording was a cover by Judy Collins. So the first recording of a work can be a cover.

a cover is a performance of a work not written by the artist. It does not matter whether the author has recorded it before the cover performance, so the first recording can be a cover.

The main way in which a tribute band differs from a cover band that simply plays songs by other artists, is that it strives to capture every nuance of the imitated artist's actions and appearance for a perfect imitation.

I now realize that wikipedia means that if A composed a song, then B performed it, then A performed it, A would be doing a cover. This seems a little strange to me, but maybe it is right. A would be doing a cover of B's cover of A's song.

Lixobix wrote:

Here's an example. The song (work) 'I'll Keep It with Mine' was written by Bob Dylan. The first recording was by Judy Collins, and this was a cover version. Dylan later recorded several versions of the song, although none were released until years later. Other cover versions were recorded by Nico, amongst others.

http://en.wikipedia.org/wiki/I%27ll_Keep_It_with_Mine

The song was clearly written by Bob Dylan, but the first recording was a cover by Judy Collins. So the first recording of a work can be a cover.

I think you should read the "Dylan's Version" subtitle there.

Lixobix wrote:

a cover is a performance of a work not written by the artist. It does not matter whether the author has recorded it before the cover performance, so the first recording can be a cover.

Does this definition tells that there shouldn't be any recording-recording cover relationship?

Because it says that a cover does not care how or who performed it before. It just cares if the performer is the composer of the song or not! To be honest that is weird.

Anyways, for me, cover implies different style and I guess styles are not defined in compositions, or do they? IMO, if an artist is trying to perform close or the same performance, that is http://en.wikipedia.org/wiki/Tribute_act.

http://en.wikipedia.org/wiki/Tribute_act wrote:

The main way in which a tribute band differs from a cover band that simply plays songs by other artists, is that it strives to capture every nuance of the imitated artist's actions and appearance for a perfect imitation.

BTW, your Dylan example also quite weird because in that wikipedia page it is said as he was first recorded it and later released.

davitof wrote:

I now realize that wikipedia means that if A composed a song, then B performed it, then A performed it, A would be doing a cover. This seems a little strange to me, but maybe it is right. A would be doing a cover of B's cover of A's song.

I thin "A" is playing his own song in a different style later. Thats why later version should be called cover.