Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/31 David Gasaway d...@gasaway.org On Sat, Jan 28, 2012 at 06:16, symphonick symphon...@gmail.com wrote: The general idea is to use the track title field to store the actual track title, normalized as any other title in mb (maybe a little bit more). Caps, delimiters etc. but not changing Adagio from Spring to: The four seasons: Spring: II. Adagio or something. This is what concerns me, and this is why I keep bringing up RFC-333. The important message of RFC-333 is that the Musicbrainz schema has no good place to store as-on-cover titles. If you try to use track titles to store this data, you lose import release context; track titles are the *only* place to put the recording in the context of the specific release. RFC-333 doesn't apply to classical. I'm sorry, but you lost me here. What is import release context? I suppose we could ask for a CSG language, if there are people who really really cannot be without old CSG-style track titles. What kind of titles would you like to see on your files when you tag a classical release? And where will these titles come from? -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org The titles printed on the cover (within context, also from cover/booklet) /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Tue, Jan 31, 2012 at 04:44, symphonick symphon...@gmail.com wrote: RFC-333 doesn't apply to classical. I'm sorry, but you lost me here. What is import release context? I know that RFC-333 does not apply to classical, and I think I made that clear long ago. The point I'm making is that the same issues that made RFC-333 necessary for popular music, should inform the choices we make for classical. A particular recording can be used on any number of releases. So release context is what gives the recording extra meaning in the context of the release. For popular, we have things like (live). For classical, the best example I can give is language. Say a recording is made and initially released in Russia. So, the recording title is entered in Russian. Now the same recording is released in the United States. Well, which titles do I use? Recording titles are useless for me, as they are not localizable. Track titles would be in English, since they put the recordings into the context of an English-language release. Make sense? The titles printed on the cover (within context, also from cover/booklet) In other words, you do not want full CSG titles, as we would build for the work titles. You want titles like those in the proposal. Unfortunately, I do not. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/31 David Gasaway d...@gasaway.org On Tue, Jan 31, 2012 at 04:44, symphonick symphon...@gmail.com wrote: RFC-333 doesn't apply to classical. I'm sorry, but you lost me here. What is import release context? I know that RFC-333 does not apply to classical, and I think I made that clear long ago. The point I'm making is that the same issues that made RFC-333 necessary for popular music, should inform the choices we make for classical. A particular recording can be used on any number of releases. So release context is what gives the recording extra meaning in the context of the release. For popular, we have things like (live). For classical, the best example I can give is language. Say a recording is made and initially released in Russia. So, the recording title is entered in Russian. Now the same recording is released in the United States. Well, which titles do I use? Recording titles are useless for me, as they are not localizable. Track titles would be in English, since they put the recordings into the context of an English-language release. Make sense? Yes, I can't use recording titles either, for the same reason. But IMO printed titles are fine in most cases, you should not be forced to use full blown CSG to just enter a classical release. If I'm not happy with a release, I could add a pseudo-release with CSG titles, it's easy now that we can start from a existing release. we could even mark the release [CSG] somewhere. The titles printed on the cover (within context, also from cover/booklet) In other words, you do not want full CSG titles, as we would build for the work titles. You want titles like those in the proposal. Unfortunately, I do not. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 16:04, David Hilton quercus.aeter...@gmail.com Regardless, I don't consider this the ideal solution. Putting the redundant information into the work hierarchy and then appending the unique parts is much more appealing. Definitely, and I'd be happy to revisit this when we get there. I think it's still likely to produce odd results with opera (though exactly how opera recording/work linking is going to work is still undecided, I think). I feel like you focused on less important parts of my message; perhaps I was unclear, or you agree with the rest of what I said? No, I think you were clear. It's just that I've been down this road already in my mind, and concluded it's unlikely to work well. :) -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: That I can agree with, some of them do not conform to old CSG (and some look too simplified for me too, tbh). But mostly I am asking: what would you, personally, see as good enough? Can you choose, from the list, those who are unacceptable to you, and say what is the minimum amount of normalisation you'd accept for each? (it's hard to compromise on vague grounds :) ). To be honest, it would be more or less the same as I expressed a few months ago in the work title thread. I have a feeling the core idea is getting lost in the discussion, however. My goal, in the end, is to have the same titles that I had before NGS. And in my mind, using track titles that are closer to whats on the release makes that impossible, or at least very difficult. So, yes, all the edits we did before for track titles, we would continue to do for track titles. I apologize if that's vague. First, I just want to make sure I've gotten the concept across. I don't necessarily have time to give a lot of detail aside from what input I've made in the past. Let me ask this. Suppose we implemented a system like that outlined by David Hilton. Do you believe that it could: a) produce consistent, high-quality titles, and b) retain relevant release context? -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Sat, Jan 28, 2012 at 06:16, symphonick symphon...@gmail.com wrote: The general idea is to use the track title field to store the actual track title, normalized as any other title in mb (maybe a little bit more). Caps, delimiters etc. but not changing Adagio from Spring to: The four seasons: Spring: II. Adagio or something. This is what concerns me, and this is why I keep bringing up RFC-333. The important message of RFC-333 is that the Musicbrainz schema has no good place to store as-on-cover titles. If you try to use track titles to store this data, you lose import release context; track titles are the *only* place to put the recording in the context of the specific release. I suppose we could ask for a CSG language, if there are people who really really cannot be without old CSG-style track titles. What kind of titles would you like to see on your files when you tag a classical release? And where will these titles come from? -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
If the recording titles are normalized and CSGified, why is it not possible to simply use those when tagging for those that want it (myself included) while those that want as on-cover could use the track titles. Sebastien On Mon, Jan 30, 2012 at 7:03 PM, David Gasaway d...@gasaway.org wrote: On Sat, Jan 28, 2012 at 06:16, symphonick symphon...@gmail.com wrote: The general idea is to use the track title field to store the actual track title, normalized as any other title in mb (maybe a little bit more). Caps, delimiters etc. but not changing Adagio from Spring to: The four seasons: Spring: II. Adagio or something. This is what concerns me, and this is why I keep bringing up RFC-333. The important message of RFC-333 is that the Musicbrainz schema has no good place to store as-on-cover titles. If you try to use track titles to store this data, you lose import release context; track titles are the *only* place to put the recording in the context of the specific release. I suppose we could ask for a CSG language, if there are people who really really cannot be without old CSG-style track titles. What kind of titles would you like to see on your files when you tag a classical release? And where will these titles come from? -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Mon, Jan 30, 2012 at 16:07, Lemire, Sebastien m...@benji99.ca wrote: If the recording titles are normalized and CSGified, why is it not possible to simply use those when tagging for those that want it (myself included) while those that want as on-cover could use the track titles. Well, as I said in my first post in this thread: a) They lack release context. b) Recording titles are not localizable (which I suppose you could include in (a) if you consider language a part of the release context). -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Mon, Jan 30, 2012 at 5:10 PM, David Gasaway d...@gasaway.org wrote: On Mon, Jan 30, 2012 at 16:07, Lemire, Sebastien m...@benji99.ca wrote: If the recording titles are normalized and CSGified, why is it not possible to simply use those when tagging for those that want it (myself included) while those that want as on-cover could use the track titles. Well, as I said in my first post in this thread: a) They lack release context. b) Recording titles are not localizable (which I suppose you could include in (a) if you consider language a part of the release context). What are you referring to by release context? I'm interpreting that as ARs, but tagging software should be able to attach any or all of the ARs for release, recording and work. That bit about recording titles not being localizable could be an issue, if you want to have a 3rd unique option to title the track with... David ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Mon, Jan 30, 2012 at 16:28, David Hilton quercus.aeter...@gmail.com What are you referring to by release context? I'm interpreting that as ARs, but tagging software should be able to attach any or all of the ARs for release, recording and work. Yes, well... I can think of no concrete classical-themed example off the top of my head. If I could, I'd have given it already. :) But in popular music, this would be referring to ETI, maybe some other things brought up in the RFC-333 discussion. http://musicbrainz.org/doc/Style/Titles/Extra_title_information -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/28 Nicolás Tamargo de Eguren reosare...@gmail.com On Sat, Jan 28, 2012 at 1:46 AM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Except that I explained the reasons why I still want full CSG on track titles, and why RFC-333 should apply to classical as well (meaning recording titles and track titles should match). You didn't address that at all. RFC-333 *never* said that recording titles and track titles should match - Lukas has repeated that several times. That's why my explanation included *both*. :) RFC-333 said that track titles get the same normalization (same as recording titles), but would include release context. Extrapolating that to classical, both would get CSG normalization. Because recording titles get CSG normalization, no? Even if the end result was to end CSG normalization for recording titles, I'm still right back where I started: I have no way, in the new scheme of things, to the get the same track titles that I know and love from the old system. It did say that for now, until changes were made if deemed interesting, we would use the same *guidelines* for recording titles and track titles (it also explicitly excluded classical, FWIW). Yes, it explicitly excluded classical so that it could be addressed in another discussion. Which is happening now. And now I'm making the argument that all the reasons for RFC-333 apply here. :) One of its goals was indeed not to blindly follow the cover, but I suspect symphonick is not trying to make anyone blindly follow the cover either: classical titles in this pre-RFC still retain a much stronger normalisation than any popular music titles do via the current titling guidelines. That's not my impression looking at the examples in the pre-RFC. That being said, what's full CSG for you? Is it using a standard track title, in the same form and language, for all appearances of a work regardless of liner language? No, in my view the form will not always be the same, thanks to release context, same as in popular music. Is it adding work titles at the beginning of the track, movement numbers, and colons to separate the full work title and the movement? (I tend to agree with that, but that's not that different from the pre-RFC). I've seen as many ideas of full CSG as classical editors, so it'd be useful to know what extent of CSG normalisation you consider basic. CSG as in old CSG. I know there's plenty of disagreement about what that does and does not include. But if you came across a classical release with track titles like those shown in the proposal, I'm fairly confident you would agree that those *do not* conform to old CSG. That I can agree with, some of them do not conform to old CSG (and some look too simplified for me too, tbh). But mostly I am asking: what would you, personally, see as good enough? Can you choose, from the list, those who are unacceptable to you, and say what is the minimum amount of normalisation you'd accept for each? (it's hard to compromise on vague grounds :) ). The general idea is to use the track title field to store the actual track title, normalized as any other title in mb (maybe a little bit more). Caps, delimiters etc. but not changing Adagio from Spring to: The four seasons: Spring: II. Adagio or something. The examples may be missing normalization now, I just copied them from the 2011 research page. Could (all of) you be more specific about what looks wrong in the examples? RFC-333 was not for classical: http://musicbrainz.1054305.n4.nabble.com/RFC-333-Unify-track-recording-guidelines-td3695823.html- I'm not sure I follow you there. I suppose we could ask for a CSG language, if there are people who really really cannot be without old CSG-style track titles. That would mean entering a pseudo-release with CSG in the language field then just choose CSG as language in Picard. A little bit more work for those who want CSG, but we would also be able to save many of the old tracklists by just changing the language field to CSG. Thoughts? /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 8:38 AM, David Gasaway d...@gasaway.org wrote: On Thu, Jan 26, 2012 at 06:46, symphonick symphon...@gmail.com wrote: Work titles are independent of track names. Work titles should ideally be sourced from the (best possible) score. Eventually most works will exist in the db, so it will be just a question of linking (the recordings) to the appropriate works. (Not replying to symphonick in particular here...) I'm at a loss to understand why the same arguments that applied to RFV-333 for popular music don't apply here. I, as a user, want fully normalized titles in my language, with release-specific ETI. So here are all the potential problems I see for not applying CSG to track titles: 1) Track titles: If I tag from these I get un-normalized garbage. 2) Recording titles: I lose release-specific title variations. In the classical world, in particular, I'm thinking of language variations. Recording titles are not localizable, so there's no telling what language I get. Surely, we're not thinking of having a different recording for language it was released in. 3) Work titles: While they have can have localized aliases, they are not usable for track titles thanks to part-performance and multi-work-performance goodness. In other words, barring other changes, the only way to get what I have now would be to continue normalizing track titles per CSG as before. No? Nobody is asking to stop normalising titles - just to stop overdoing it. Right now, the idea is to always have, for example, a track title like Divertimento No. 2 for Flute, Oboe, Bassoon, 4 Horns Strings in D major, K. 131: I. Allegro, even if the release says Divertimento nº 2, K131: Allegro. The latter would seem like it's enough, though (especially IMO if you added the I. by default, but in general I see the point). Same for, for example, a release of clarinet concertos. It seems like overkill to add for clarinet to every track title if the release lists Concerto No. n, Op. n: I. Allegro molto. On the other hand, I do wonder what's symphonick's plan for releases which are like Allegro (from Divertimento nº 2, K131) - should the order for those be changed or kept? -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Nobody is asking to stop normalising titles - just to stop overdoing it. Right now, the idea is to always have, for example, a track title like Divertimento No. 2 for Flute, Oboe, Bassoon, 4 Horns Strings in D major, K. 131: I. Allegro, even if the release says Divertimento nº 2, K131: Allegro. The pre-RFC proposal disagrees with you. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 10:09 PM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Nobody is asking to stop normalising titles - just to stop overdoing it. Right now, the idea is to always have, for example, a track title like Divertimento No. 2 for Flute, Oboe, Bassoon, 4 Horns Strings in D major, K. 131: I. Allegro, even if the release says Divertimento nº 2, K131: Allegro. The pre-RFC proposal disagrees with you. I said *right now*, as in before the pre-RFC. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 02:00, David Hilton quercus.aeter...@gmail.com wrote: Sorry, I'm not clear on how part-performance and multi-work-performance stuff makes work titles unusable for track titles. Could you explain? Sure. If I have a track that spans multiple parts, then I want a track title that includes both parts. However, that recording will link to two works, one for each part. So, I have two work titles, how do I get from there to a single title with both parts? Similarly, if it's a partial performance, the work title would not properly reflect such. Theoretically, you could kludge it by slapping (partial) or (excerpt) if it's a partial performance AR, but static phrases like that might not always be appropriate. I would expect to instead use release/recording specific ETI. Again, using recording titles is problematic, so I'm back to track ETI for this information. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: On Fri, Jan 27, 2012 at 10:09 PM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Nobody is asking to stop normalising titles - just to stop overdoing it. Right now, the idea is to always have, for example, a track title like Divertimento No. 2 for Flute, Oboe, Bassoon, 4 Horns Strings in D major, K. 131: I. Allegro, even if the release says Divertimento nº 2, K131: Allegro. The pre-RFC proposal disagrees with you. I said *right now*, as in before the pre-RFC. I don't understand. You said nobody is asking to stop normalising titles, yet the pre-RFC says exactly that. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 10:21 PM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: On Fri, Jan 27, 2012 at 10:09 PM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Nobody is asking to stop normalising titles - just to stop overdoing it. Right now, the idea is to always have, for example, a track title like Divertimento No. 2 for Flute, Oboe, Bassoon, 4 Horns Strings in D major, K. 131: I. Allegro, even if the release says Divertimento nº 2, K131: Allegro. The pre-RFC proposal disagrees with you. I said *right now*, as in before the pre-RFC. I don't understand. You said nobody is asking to stop normalising titles, yet the pre-RFC says exactly that. It really depends on what you call normalising. Adding the name of the work to all track titles is normalising, and it's useful. The pre-RFC tells to keep doing it. Repeating the same stock title all the time so all the tracks in the DB for the same work have the same title is useless now that we can just relate the work to it to show they're all the same thing, so that's just overkill right now. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 1:16 PM, David Gasaway d...@gasaway.org wrote: On Fri, Jan 27, 2012 at 02:00, David Hilton quercus.aeter...@gmail.com wrote: Sorry, I'm not clear on how part-performance and multi-work-performance stuff makes work titles unusable for track titles. Could you explain? Sure. If I have a track that spans multiple parts, then I want a track title that includes both parts. However, that recording will link to two works, one for each part. So, I have two work titles, how do I get from there to a single title with both parts? Work1: Subwork1 Work1: Subwork2 It wouldn't be difficult to have Picard recognize that a recording represents multiple works and remove the starting duplicate info: Work1: Subwork1 / Subwork2 That said, conceptually I prefer something like what follows. Das klagende Lied - I. Waldmärchen - II. Der Spielmann - III. Hochzeitsstück Each entry only has its relevant info, and can easily be combined into the full work name. Similarly, if it's a partial performance, the work title would not properly reflect such. Theoretically, you could kludge it by slapping (partial) or (excerpt) if it's a partial performance AR, but static phrases like that might not always be appropriate. I would expect to instead use release/recording specific ETI. Again, using recording titles is problematic, so I'm back to track ETI for this information. It was suggested a while back that where common divisions take place sub-sub works could be created; take for example the 4th movement of Beethoven's 9th, where it's not uncommon for each part of the 4th movement to have its own track. In these cases, the tagger could grab all relevant works and combine them as desired: W1 - S1 -- SS1 -- SS2 -- SS3 W1: SS1 - SS2 W1: SS3 With this approach, consistent naming schemes could be generated for those who care, and track titles could contain what was listed from the cover. If a recording marked as a partial performance, the tagger could use the recording/track title. At the same time, there is nothing against your tagging software saving both the work info and the track title... What issues are there with this particular approach? David ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: It really depends on what you call normalising. Adding the name of the work to all track titles is normalising, and it's useful. The pre-RFC tells to keep doing it. Repeating the same stock title all the time so all the tracks in the DB for the same work have the same title is useless now that we can just relate the work to it to show they're all the same thing, so that's just overkill right now. Except that I explained the reasons why I still want full CSG on track titles, and why RFC-333 should apply to classical as well (meaning recording titles and track titles should match). You didn't address that at all. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 15:10, David Hilton quercus.aeter...@gmail.com It wouldn't be difficult to have Picard recognize that a recording represents multiple works and remove the starting duplicate info: Work1: Subwork1 / Subwork2 You're making a lot of assumptions about what the subworks would be, and whether edits like this would be appropriate. With opera, in particular, things get really hairy. Work titles only have structure in that they will conform to an as-yet-WIP CSG and can and will change. Works, themselves, are used to represent a multitude of entities. Bottom line, I have no confidence the the output of such a process would not be total garbage. It was suggested a while back that where common divisions take place sub-sub works could be created; They *could*, but maybe not, maybe it will have both, maybe it will be case-by-case, that change changes from release to release. That's the whole point. By using works instead of tracks for tagging you *lose all release context*. It's the same argument made months ago in RFV-333. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Sat, Jan 28, 2012 at 1:13 AM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: It really depends on what you call normalising. Adding the name of the work to all track titles is normalising, and it's useful. The pre-RFC tells to keep doing it. Repeating the same stock title all the time so all the tracks in the DB for the same work have the same title is useless now that we can just relate the work to it to show they're all the same thing, so that's just overkill right now. Except that I explained the reasons why I still want full CSG on track titles, and why RFC-333 should apply to classical as well (meaning recording titles and track titles should match). You didn't address that at all. RFC-333 *never* said that recording titles and track titles should match - Lukas has repeated that several times. It did say that for now, until changes were made if deemed interesting, we would use the same *guidelines* for recording titles and track titles (it also explicitly excluded classical, FWIW). That's why we have, for example, a guideline to take (live) out of recording titles and into the disambiguation comment, while keeping it on track titles when it applies, or why we can have track titles in several languages linking to the same recording. One of its goals was indeed not to blindly follow the cover, but I suspect symphonick is not trying to make anyone blindly follow the cover either: classical titles in this pre-RFC still retain a much stronger normalisation than any popular music titles do via the current titling guidelines. That being said, what's full CSG for you? Is it using a standard track title, in the same form and language, for all appearances of a work regardless of liner language? (I consider that a big overkill) Is it adding work titles at the beginning of the track, movement numbers, and colons to separate the full work title and the movement? (I tend to agree with that, but that's not that different from the pre-RFC). I've seen as many ideas of full CSG as classical editors, so it'd be useful to know what extent of CSG normalisation you consider basic. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Except that I explained the reasons why I still want full CSG on track titles, and why RFC-333 should apply to classical as well (meaning recording titles and track titles should match). You didn't address that at all. RFC-333 *never* said that recording titles and track titles should match - Lukas has repeated that several times. That's why my explanation included *both*. :) RFC-333 said that track titles get the same normalization (same as recording titles), but would include release context. Extrapolating that to classical, both would get CSG normalization. Because recording titles get CSG normalization, no? Even if the end result was to end CSG normalization for recording titles, I'm still right back where I started: I have no way, in the new scheme of things, to the get the same track titles that I know and love from the old system. It did say that for now, until changes were made if deemed interesting, we would use the same *guidelines* for recording titles and track titles (it also explicitly excluded classical, FWIW). Yes, it explicitly excluded classical so that it could be addressed in another discussion. Which is happening now. And now I'm making the argument that all the reasons for RFC-333 apply here. :) One of its goals was indeed not to blindly follow the cover, but I suspect symphonick is not trying to make anyone blindly follow the cover either: classical titles in this pre-RFC still retain a much stronger normalisation than any popular music titles do via the current titling guidelines. That's not my impression looking at the examples in the pre-RFC. That being said, what's full CSG for you? Is it using a standard track title, in the same form and language, for all appearances of a work regardless of liner language? No, in my view the form will not always be the same, thanks to release context, same as in popular music. Is it adding work titles at the beginning of the track, movement numbers, and colons to separate the full work title and the movement? (I tend to agree with that, but that's not that different from the pre-RFC). I've seen as many ideas of full CSG as classical editors, so it'd be useful to know what extent of CSG normalisation you consider basic. CSG as in old CSG. I know there's plenty of disagreement about what that does and does not include. But if you came across a classical release with track titles like those shown in the proposal, I'm fairly confident you would agree that those *do not* conform to old CSG. -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Fri, Jan 27, 2012 at 4:24 PM, David Gasaway d...@gasaway.org wrote: On Fri, Jan 27, 2012 at 15:10, David Hilton quercus.aeter...@gmail.com It wouldn't be difficult to have Picard recognize that a recording represents multiple works and remove the starting duplicate info: Work1: Subwork1 / Subwork2 You're making a lot of assumptions about what the subworks would be, and whether edits like this would be appropriate. With opera, in particular, things get really hairy. Work titles only have structure in that they will conform to an as-yet-WIP CSG and can and will change. Works, themselves, are used to represent a multitude of entities. Bottom line, I have no confidence the the output of such a process would not be total garbage. Yes, there are limitations to this approach; without having consistent delimiters everything has to match to, this could produce garbage. Stuff that we already use like ':' would be ok. Regardless, I don't consider this the ideal solution. Putting the redundant information into the work hierarchy and then appending the unique parts is much more appealing. It was suggested a while back that where common divisions take place sub-sub works could be created; They *could*, but maybe not, maybe it will have both, maybe it will be case-by-case, that change changes from release to release. That's the whole point. By using works instead of tracks for tagging you *lose all release context*. It's the same argument made months ago in RFV-333. I'm not suggesting that release track titles should contain work titles. I'm trying to propose that track titles should contain what was literally on the release (as much as possible), and that the work information should be sufficient to automatically replicate the old CSG track titles, if desired (ideally using the work hierarchy to reduce duplicate text). Users could then specify how they want these things saved; 'save work names (merging work+subwork names) as track titles' instead of 'save work names as work names and track titles as track titles' or either 'save work names as track titles' or ' save track titles as track titles'. I feel like you focused on less important parts of my message; perhaps I was unclear, or you agree with the rest of what I said? David ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Sat, Jan 28, 2012 at 1:46 AM, David Gasaway d...@gasaway.org wrote: 2012/1/27 Nicolás Tamargo de Eguren reosare...@gmail.com: Except that I explained the reasons why I still want full CSG on track titles, and why RFC-333 should apply to classical as well (meaning recording titles and track titles should match). You didn't address that at all. RFC-333 *never* said that recording titles and track titles should match - Lukas has repeated that several times. That's why my explanation included *both*. :) RFC-333 said that track titles get the same normalization (same as recording titles), but would include release context. Extrapolating that to classical, both would get CSG normalization. Because recording titles get CSG normalization, no? Even if the end result was to end CSG normalization for recording titles, I'm still right back where I started: I have no way, in the new scheme of things, to the get the same track titles that I know and love from the old system. It did say that for now, until changes were made if deemed interesting, we would use the same *guidelines* for recording titles and track titles (it also explicitly excluded classical, FWIW). Yes, it explicitly excluded classical so that it could be addressed in another discussion. Which is happening now. And now I'm making the argument that all the reasons for RFC-333 apply here. :) One of its goals was indeed not to blindly follow the cover, but I suspect symphonick is not trying to make anyone blindly follow the cover either: classical titles in this pre-RFC still retain a much stronger normalisation than any popular music titles do via the current titling guidelines. That's not my impression looking at the examples in the pre-RFC. That being said, what's full CSG for you? Is it using a standard track title, in the same form and language, for all appearances of a work regardless of liner language? No, in my view the form will not always be the same, thanks to release context, same as in popular music. Is it adding work titles at the beginning of the track, movement numbers, and colons to separate the full work title and the movement? (I tend to agree with that, but that's not that different from the pre-RFC). I've seen as many ideas of full CSG as classical editors, so it'd be useful to know what extent of CSG normalisation you consider basic. CSG as in old CSG. I know there's plenty of disagreement about what that does and does not include. But if you came across a classical release with track titles like those shown in the proposal, I'm fairly confident you would agree that those *do not* conform to old CSG. That I can agree with, some of them do not conform to old CSG (and some look too simplified for me too, tbh). But mostly I am asking: what would you, personally, see as good enough? Can you choose, from the list, those who are unacceptable to you, and say what is the minimum amount of normalisation you'd accept for each? (it's hard to compromise on vague grounds :) ). -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
[mb-style] Pre-RFC: CSG for track titles
I thought I should seize the moment to get back to CSGv2. Please have a look att this pre-RFC: http://wiki.musicbrainz.org/User:symphonick/tracknames_preRFC For those interested in the research leading to this: http://wiki.musicbrainz.org/User:symphonick/Unofficial_CSG_track_names -- /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Thu, Jan 26, 2012 at 3:15 PM, symphonick symphon...@gmail.com wrote: I thought I should seize the moment to get back to CSGv2. Please have a look att this pre-RFC: http://wiki.musicbrainz.org/User:symphonick/tracknames_preRFC For those interested in the research leading to this: http://wiki.musicbrainz.org/User:symphonick/Unofficial_CSG_track_names A small complaint: not adding the I. II. III. etc to titles makes it a pain to create works from them. -- /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/26 Nicolás Tamargo de Eguren reosare...@gmail.com On Thu, Jan 26, 2012 at 3:15 PM, symphonick symphon...@gmail.com wrote: I thought I should seize the moment to get back to CSGv2. Please have a look att this pre-RFC: http://wiki.musicbrainz.org/User:symphonick/tracknames_preRFC For those interested in the research leading to this: http://wiki.musicbrainz.org/User:symphonick/Unofficial_CSG_track_names A small complaint: not adding the I. II. III. etc to titles makes it a pain to create works from them. Work titles are independent of track names. Work titles should ideally be sourced from the (best possible) score. Eventually most works will exist in the db, so it will be just a question of linking (the recordings) to the appropriate works. /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
Regarding the the roman numerals. I used to be of the opinion several years ago that almost all parts or movements should have roman numerals for the order but recently I've changed my mind. When I source work parts if I see all or most sources not putting roman numerals, I don't put them either. Should we set a rule that they apply by default to Symphonies, Sonatas and Concertos and on a case-by-case basis for other work-types? There should definitely be no Roman numerals for Operas, Song-cycles, collections, etc... Sébastien On Thu, Jan 26, 2012 at 9:46 AM, symphonick symphon...@gmail.com wrote: 2012/1/26 Nicolás Tamargo de Eguren reosare...@gmail.com On Thu, Jan 26, 2012 at 3:15 PM, symphonick symphon...@gmail.com wrote: I thought I should seize the moment to get back to CSGv2. Please have a look att this pre-RFC: http://wiki.musicbrainz.org/User:symphonick/tracknames_preRFC For those interested in the research leading to this: http://wiki.musicbrainz.org/User:symphonick/Unofficial_CSG_track_names A small complaint: not adding the I. II. III. etc to titles makes it a pain to create works from them. Work titles are independent of track names. Work titles should ideally be sourced from the (best possible) score. Eventually most works will exist in the db, so it will be just a question of linking (the recordings) to the appropriate works. /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
2012/1/26 lorenz pressler l...@gmx.at example: String Quartet no. 12 in F major, op. 96 American: Finale: Vivace, ma non troppo and recording titles should follow track titles. so then the old CSG formatting guidelines are completely obsolete? how i interpret your draft is: follow the release-liners completely for track and recordings. Nr X. NoX noX no X op OPUS opus ... i actually liked the way musicbrainz was giving the titles a consistent way of formatting throughout different releases; i would not care if the tracktitles follow the liners if the recordings are normalized and there would be the feature re-introduced in picard to use recording names for tagging. like this it's kind of a downer for me. Yes, tracks ( recordings) will be more like normal mb. Old CSG will be mostly for works. /symphonick ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
lorenz pressler l...@gmx.at writes: i actually liked the way musicbrainz was giving the titles a consistent way of formatting throughout different releases; i would not care if the tracktitles follow the liners if the recordings are normalized and there would be the feature re-introduced in picard to use recording names for tagging. like this it's kind of a downer for me. Ah! That's a use for recording titles! That hadn't occurred to me. Hmm, I prefer not propagating crap from the CD sleeve into the database, but I guess I could live with this too. Rupert pgpzmjLDnnQ9P.pgp Description: PGP signature ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Thu, Jan 26, 2012 at 11:40 PM, Rupert Swarbrick rswarbr...@gmail.com wrote: lorenz pressler l...@gmx.at writes: i actually liked the way musicbrainz was giving the titles a consistent way of formatting throughout different releases; i would not care if the tracktitles follow the liners if the recordings are normalized and there would be the feature re-introduced in picard to use recording names for tagging. like this it's kind of a downer for me. Ah! That's a use for recording titles! That hadn't occurred to me. Hmm, I prefer not propagating crap from the CD sleeve into the database, but I guess I could live with this too. But for that, wouldn't it be simpler to tag with work names? (not possible now either, but both need changes to Picard / a plugin anyway) Rupert ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style -- Nicolás Tamargo de Eguren ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
Am 26.01.2012, 22:56 Uhr, schrieb Nicolás Tamargo de Eguren reosare...@gmail.com: But for that, wouldn't it be simpler to tag with work names? (not possible now either, but both need changes to Picard / a plugin anyway) not all recordings have a work attached (but all tracks have a recording) some recordings have multiple works attached some releases have one work for multiple recordings but yes, having to maintain all these title-lvls with different formatting rules will be quite bothersome. -- lorenz pressler PGP 0x92E9551A ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style
Re: [mb-style] Pre-RFC: CSG for track titles
On Thu, Jan 26, 2012 at 06:46, symphonick symphon...@gmail.com wrote: Work titles are independent of track names. Work titles should ideally be sourced from the (best possible) score. Eventually most works will exist in the db, so it will be just a question of linking (the recordings) to the appropriate works. (Not replying to symphonick in particular here...) I'm at a loss to understand why the same arguments that applied to RFV-333 for popular music don't apply here. I, as a user, want fully normalized titles in my language, with release-specific ETI. So here are all the potential problems I see for not applying CSG to track titles: 1) Track titles: If I tag from these I get un-normalized garbage. 2) Recording titles: I lose release-specific title variations. In the classical world, in particular, I'm thinking of language variations. Recording titles are not localizable, so there's no telling what language I get. Surely, we're not thinking of having a different recording for language it was released in. 3) Work titles: While they have can have localized aliases, they are not usable for track titles thanks to part-performance and multi-work-performance goodness. In other words, barring other changes, the only way to get what I have now would be to continue normalizing track titles per CSG as before. No? -- -:-:- David K. Gasaway -:-:- Email: d...@gasaway.org ___ MusicBrainz-style mailing list MusicBrainz-style@lists.musicbrainz.org http://lists.musicbrainz.org/mailman/listinfo/musicbrainz-style