forum

[Proposal] General metadata ruleset draft

posted
Total Posts
34
Topic Starter
IamKwaN
This paper explains a proposal discussed among Metadata Division members from the United Beat-Knights of Ranking Criteria to update the obsolete metadata ranking criteria. We would like to have your views on the proposal on or before 31 May 2017.

Notice this is NOT the final result, we need your precious opinion before any official implementation.

Glossary


  1. Romanised Fields: Collective term for Romanised Title and Romanised Artist, its usage is also shared by translated title, preferred romanised title and artist.
  2. Modified Hepburn Romanisation: Refer to this page for information.
  3. Word-by-word Romanisation: Each character must be romanised into a single, capitalised, separated word. Refer to this thread for examples and supplementary information.
  4. Cyrillic Romanisation: Use BGN/PCGN system for Russian/Cyrillic. "Е" and "е" should be romanised as "ye" if it stands alone or after "a", "e", "ё", "и", "о", "у", "ы", "э", "ю", "я", "й", "ъ", "ь". In other cases, it should be romanised as "e". Rule 1 also applies to ё, however, use "yo" or "o" to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. If Cyrillic is used to represent a word from another language, use the form from that language. For most of the other characters, refer to the first page of this document.

Rules

All rules are exactly that: RULES. They are NOT guidelines and may NOT be broken under ANY circumstance.

  1. Metadata must be consistent across all difficulties. This is to ensure that searching is consistent and clean across a mapset.
  2. Songs with multi-lingual adaptions must use the metadata in its original form in the unicode fields and have their translated counterparts in the romanised fields. Only translated metadata is used unless translated version is different from the original version.
  3. Do not alter the song's title. This includes placing official preferred romanisation at the highest priority and not adding any marker to describe cut of the song such as "Short Ver.", "TV Size", "Long Version" and so forth unless those markers are part of the official song title as listed by a reputable source. If possible, copy and paste the exact title directly from an official source and use official markers as they are when they are present. However, songs with official metadata exceeding any field limits should be subject to reasonable modifications under a case by case basis. Please refer to the supplementary information section for some examples.
  4. Where multiple official titles exist for a song, the mapper may decide which to use as all are considered equally acceptable. Mappers have the right to choose based on their preference. Please refer to the next section for more information defining official sources.
  5. Minor edits of the song, including timing adjustments or adding beats require no change in the title. A description showing what has been edited must be shown in the creator's words.
  6. The creator is entitled to name the song with their major edits, including compilation, nightcore or remixing, in a reasonable way indicating the song is no longer official. Some common renames are "Artist - Songs Compilation", "A & B - C vs D", "Artist - Title (Nightcore Mix)" and "Artist - Title (Speed Up Ver.)".
  7. Do not alter the song's artist. The exact alias for the song on the date published should be used. Eastern, including Chinese, Japanese and Korean, artists must be written in the proper 'surname firstname' format, unless official preferred romanisation is available. Eastern artists consisting of several persons must use preferred romanisation with the "surname firstname" format, unless there is information regarding official preferred romanisation for the whole group.
  8. Priority on the artist is applied on Vocaloid tracks in the following order: circle name / CD alias, composer and lyricist. If one or more staff member(s) for the song does not belong to the same circle, prioritize composer and lyricist instead.
  9. Different priority on the artist is adopted depending on the existence of a vocalist for a doujin. Circle name / CD alias overrides the vocalist for songs with vocals while circle name / CD alias overrides the composer for instrumental songs. If one or more staff member(s) for the song does not belong to the same circle, prioritize vocalist and composer instead.
  10. Only use the Source field if the song, including different versions and remixes, comes from an animation, video game, movie, or series. Website names, album names and format names are not acceptable sources.
  11. Do not alter the song's source. Both official romanisation and translation are allowed. Although using a romanised source is acceptable when there is no official translation available, using the original and the most accurate source is encouraged.
  12. Information of multiple mapset contributors must be provided in the mapset, if there is any guest mapper or storyboarder. This might be in the creator's words, via a storyboard or via naming the guest difficulties appropriately. You only need to provide information of guest mappers and corresponding guest parts.
  13. Guest mappers and storyboarders must be added to the tags of a mapset. This helps others know if the map uploader is the main contributor of the mapset and who else contributed to the given mapset.
  14. Songs with Japanese metadata must use the Modified Hepburn Romanisation method in romanised fields when there is no romanisation or translation information listed by a reputable source. The same applies to the "Source" field if a romanised Source is preferred by the mapper. As a non-unicode field, long vowels such as "おう" and "うう" should be romanised into "ou" and "uu" to avoid macrons. Loan words should be expressed using romanisation from the original language.
  15. Songs with Chinese metadata must be romanised in Word-by-word Romanisation method in romanised fields when there is no romanisation or translation information listed by a reputable source. The same applies to the "Source" field if a romanised Source is preferred by the mapper. Loan words should be expressed using romanisation from the original language.
  16. Songs with Russian metadata must be romanised using the Cyrillic Romanisation method in romanised fields when there is no romanisation or translation information listed by a reputable source. The same applies to the "Source" field if a romanised Source is preferred by the mapper.
  17. Songs with German umlauts ("ü", "ö", "ä" and "ß") must be romanised into two-letter combination ("ue", "oe", "ae" and "ss").

Supplementary Information

This section serves as a guideline and will be posted somewhere in the forum
when the above draft is officially implemented.

  1. Russian romanisation
    Examples:
    1. каникулы - kanikuly
    2. признание - priznaniye
  2. Official metadata exceeding field limits
    Title examples:
    1. nao - とある小国のお姫様が...
      Original Title: とある小国のお姫様が、会ったこともない兄の家にやってきて住みついた事情。
      Final Title: とある小国のお姫様が...
    Artist examples:
    1. 中村繪里子、清水 愛、青葉りんご - 星風のホロスコープ(#12 ver.)
      Original Artist: 中津川 初(CV:中村繪里子)、神本円佳(CV:清水 愛)、日向伊吹(CV:青葉りんご)
      Final Artist: 中村繪里子、清水 愛、青葉りんご
    2. 小岩井ことり、村川梨衣、佐倉綾音、阿澄佳奈 - おかえり
      Original Artist: 宮内れんげ(小岩井ことり)、一条蛍(村川梨衣)、越谷夏海(佐倉綾音)、越谷小鞠(阿澄佳奈)
      Final Artist: 小岩井ことり、村川梨衣、佐倉綾音、阿澄佳奈
  3. Official references
    Examples (without priority):
    1. Artist website
    2. Bandcamp, Soundcloud or YouTube channel of the artist
    3. CD cover or scan
    4. Label website
    5. Source (e.g. Anime) website with a proper credit to the artist
Topic Starter
IamKwaN
On top of the draft, we'd love to have your opinions particularly on:

  1. Is BMS acceptable as a Source?
  2. Any ideas on how to standardize or formalize Korean romanisation method?
Natsu
Sonnyc made a post about korean romanization here: p/4530570
-Atri-
Few questions:
1. If a song compilation is mapped from an album, Then is it okay to use the name of the album as the title of the map?
ex: https://osu.ppy.sh/s/448488
2. If a song itself have an official English title, then should the English Title/Artist should be used or using the original Japanese name as Romanized Title/Artist?
ex: https://osu.ppy.sh/s/586841 , The song supposed romanized as "Aozora no Rhapsody"
3. If the song's title is different depending the localized version of the game/anime, The should question 2 applies in this situation (Official, in this case)
ex: The ending of Atelier Firis, The Original japanese name is 光ノ軌跡, which was romanized as "Hikari no Kiseki"
But in the Chinese/Worldwide release of the game, it was named as "Trail of Light"
Electoz

IamKwaN wrote:

  1. Guest mappers and storyboarders must be added to the tags of a mapset. This helps others know if the map uploader is the main contributor of the mapset and who else contributed to the given mapset.
Is including storyboarders in a mapset a necessity? True that they should be respected for what they have done, but there's also some cases like:
  1. The storyboard is just tag names for collaboration (something like https://puu.sh/vTNX1/384adce02a.PNG ) so it wouldn't make much sense to include storyboarders in tags when the storyboard itself is minor as it can be.
  2. There's also some cases like https://osu.ppy.sh/s/442327 where people did minor SB elements and all of them ended up in the tags, I mean it would be okay-ish to put them in tags but it's not a necessity either since they didn't majorly impact the mapset with what they did with their storyboarding.
Topic Starter
IamKwaN

Firis Mistlud wrote:

Few questions:
1. If a song compilation is mapped from an album, Then is it okay to use the name of the album as the title of the map?
ex: https://osu.ppy.sh/s/448488
Yes, as stated.
> The creator is entitled to name the song with their major edits, including compilation, nightcore or remixing, in a reasonable way indicating the song is no longer official.

Firis Mistlud wrote:

2. If a song itself have an official English title, then should the English Title/Artist should be used or using the original Japanese name as Romanized Title/Artist?
ex: https://osu.ppy.sh/s/586841 , The song supposed romanized as "Aozora no Rhapsody"
According to rule no.2, Aozora no Rhapsody as Romanised Title and 青空のラプソディ as Unicode Title.
> Songs with multi-lingual adaptions must use the metadata in its original form in the unicode fields and have their translated counterparts in the romanised fields.

Firis Mistlud wrote:

3. If the song's title is different depending the localized version of the game/anime, The should question 2 applies in this situation (Official, in this case)
ex: The ending of Atelier Firis, The Original japanese name is 光ノ軌跡, which was romanized as "Hikari no Kiseki"
But in the Chinese/Worldwide release of the game, it was named as "Trail of Light"
If "Hikari no Kiseki" is not officially stated, then 光ノ軌跡 should be as Unicode Title while Trail of Light as Romanised Title, same answer as the above question.
Raveille
So, on the BMS topic, I'm just gonna post this on behalf of what #osumania has discussed in the modding association server.

Hydria - Today at 8:15 PM
"Unless the song was specifically made for the game (e.g. anything from BMSOF for BMS), it should not be a creditable source."
We have been discussing that as long as songs are made exclusively for a certain game aka original songs (such as some plot songs like Lachryma or HE4VEN for SDVX), it should be fine, hence SDVX is a citable source.

However, when it comes to BMS,

XeoStyle - Today at 8:24 PM
BMS is an emulator, how it can have anything "official" in it lol

Raveille - Today at 8:29 PM
so actually
BMS to IIDX
as for kshoot to SDVX
am I right

XeoStyle - Today at 8:29 PM
yup
So, the current verdict as it stands is...

BMS should not be a citable source.

HOWEVER, songs created in vsrg-related song-creation based tournaments, such as BMS Of Fighters (http://www.bmsoffighters.net/) and Shooter's Festival 2015 (http://ksmeve.wixsite.com/shooters-festival/sf2015) can be citable sources as the music is created purely for said tournaments and most artists credit their music with these tournament titles on soundcloud/youtube etc as seen here:


(edit thanks to Hydria)
Mint
A lot of western artists tend to link services like Spotify and iTunes when looking up their discography. Should this be considered as official then?

Also, say for instance we take the song Kimi no Shiranai Monogatari by supercell.
As far as I know, there IS a preferred AND official (according to this ruleset draft) romanization. https://youtu.be/eLPs_w-FepA
This was used as a joke before on Monstrata's map thread, but according to this, should we take the romanization from the YouTube link instead?

I don't remember which mapset (probably a SAO map) , but recently one got nominated using Recochoku, an online music store comparable to iTunes. Sometimes TV Sizes and "Radio Versions" are only released via this medium. Since stuff like iTunes is not considered official, should we then take this as a source?

Sometimes Hi-Res audio versions (24/96 FLACs) are released via OTOTOY and mora, also online music stores. These studio version can usually only be obtained from these sources. Same question, should we also consider these stores as a source?

For Korean songs, a lot of songs have an English Title (like BTS' FIRE, Girls' Generation's Genie). How do we determine to either use translated, or romanized?

Perhaps, you could also add a rule to specify whether "東方project" or the whole Cherry Blossom title etc.
Natsu
Korean songs mostly uses translated titles as official ones, But there are a few cases without translated title such as https://osu.ppy.sh/s/519665
MadBricktree

Raveille wrote:

So, on the BMS topic, I'm just gonna post this on behalf of what #osumania has discussed in the modding association server.

Hydria - Today at 8:15 PM
"Unless the song was specifically made for the game (e.g. anything from BMSOF for BMS), it should not be a creditable source."
We have been discussing that as long as songs are made exclusively for a certain game aka original songs (such as some plot songs like Lachryma or HE4VEN for SDVX), it should be fine, hence SDVX is a citable source.

However, when it comes to BMS,

XeoStyle - Today at 8:24 PM
BMS is an emulator, how it can have anything "official" in it lol

Raveille - Today at 8:29 PM
so actually
BMS to IIDX
as for kshoot to SDVX
am I right

XeoStyle - Today at 8:29 PM
yup
To elaborate, BMS is a file format, made for an emulator called BM98.
So BMS isn't even an emulator like stepmania or kshootmania.
This should be even more of a reason to not have "BMS" in source.



As for romanization of Korean, I suggest the Revised Romanization of Korean made by the Korean Ministry of Culture and Tourism.
Hydria

TheKwaninator wrote:

5. Minor edits of the song, including timing adjustments or adding beats require no change in the title. A description showing what has been edited must be shown in the creator's words.
I think a description showing that it has been edited would suffice more than what has been edited.

"Reol - No Title [Short Ver]" would be a better fit than "Reol - No Title - 00:00:000 -> 01:19:197" or "Reol - No Title (until after the first chorus)", plus not all maps in a mapset finish at the same time.

If it was just intending to be a [Short Ver] in the map's description then wording needs to be fixed as it is misleading imo.
XeoStyle
A lot of western artists tend to link services like Spotify and iTunes when looking up their discography. Should this be considered as official then?
You're missunderstanding what is "source" used for. Let's take some more simple example than BMS (wich is kinda more complicated) : a game OST.

Let's say you're charting BFG Division by Mick Gordon from the DOOM soundtrack, would you put DOOM or Spotify as a source ? Yeah ofc DOOM.

If you're song is not related to some kind of media (film, serie, game...), just let the source empty. iTunes, Bandcamp... are delivery platform, not direct source.
Kroytz

IamKwaN wrote:

Website names, album names and format names are not acceptable sources. Is BMS acceptable as a Source?
I think BMS is acceptable as a Source if and only if the song was intended to be released with it as a default song in the game or contest within the game. Iterations of the song that are not the original song belonging to its respective source should not be listed as part of that source.

Example 1:
xi - Halcyon (mapset by gowww) is the original adaptation of the song released in 2010 for BMS as part of the BoF2010 event whereas xi - Halcyon (mapset by Natsu) is the 2011 album released version (2nd adaptation) not intended to be affiliated with BMS. On a side note Natsu, the title for the full version of Halcyon is shown > here < as well as the backside of Parousia's cover. So I don't know where you got - Long Version - from zz. Therefore, only the first original version of the song should be cited as BMS and not the latter.

Example 2:
xi - Ascension to Heaven by Shiirn should be BMS whereas xi - Ascension to Heaven by Kroytz should not be.

However, in the case of it being a contest song such as Sakuzyo - Altale or xi - Aragami, I'm starting to wonder now if it would it be better to have it listed as BMS or the contest from which it was submitted to? For the two listed above, it's technical original Source would be: G2R2014 "GO BACK 2 YOUR ROOTS" -THE BMS OF FIGHTERS 2004 RELOADED- according to the the name of the contest but the platform it was released on is BMS.

Similarly, for songs belonging to other games such as SDVX, do contest songs get listed by the game platform or the contest?
As an example: Does iLLness LiLin get listed under "SDVX IV HE4VENLY HAVEN" because it was first debut'd on it or does it get Sourced under The 6th KAC Original Song Contest (The 6th KAC オリジナル楽曲コンテスト) because that was where it was first submitted to originally?

Personally, I believe for simplicity's sake that contest songs should be Sourced under their respective game platform as the original released version of the song and not part of the contest it belongs to only because these contests are part of these game platforms and in most cases designed for them - in this case, The 6th KAC Original Song Contest winners were designed specially for SDVX IV HE4VENLY HAVEN. A person looking through SDVX IV songs will only find six default songs followed by a plethora of contest song winners, licensed songs, and arrangements among others.
Mint

XeoStyle wrote:

A lot of western artists tend to link services like Spotify and iTunes when looking up their discography. Should this be considered as official then?
You're missunderstanding what is "source" used for. Let's take some more simple example than BMS (wich is kinda more complicated) : a game OST.

Let's say you're charting BFG Division by Mick Gordon from the DOOM soundtrack, would you put DOOM or Spotify as a source ? Yeah ofc DOOM.

If you're song is not related to some kind of media (film, serie, game...), just let the source empty. iTunes, Bandcamp... are delivery platform, not direct source.
I meant that these type of media can be USED a source, where you can take metadata like the Artist and Title, not being used in the "Source" field.
Enon
In case it's a korean songs, I think we have to follow the korean romanization method, which is Sonnyc posted, that is used commonly in korea or somewhere when we do romazination the korean words.
(the case of this, the metadata of the korean songs)

It's exist why avoiding any issue to standardize as the romazination the korean words imo.

Yeah, I mean I agree with Natsu in here >_>
Sieg
Cyrillic Romanisation: Use BGN/PCGN system for Russian/Cyrillic. "Е" and "е" should be romanised as "ye" if it stands alone or after "a", "e", "ё", "и", "о", "у", "ы", "э", "ю", "я", "й", "ъ", "ь". In other cases, it should be romanised as "e". Rule 1 also applies to ё, however, use "yo" or "o" to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. If Cyrillic is used to represent a word from another language, use the form from that language. For most of the other characters, refer to the first page of this document.
This is messed up, how about romanization of Cyrillic that is not on Russian, they should use Russian system too? For reference you may compare romazition of different Cyrillic based languages, there are differences.
https://www.gov.uk/government/uploads/s ... USSIAN.PDF
https://www.gov.uk/government/uploads/s ... RUSIAN.PDF
https://www.gov.uk/government/uploads/s ... AINIAN.pdf
etc
marshallracer

IamKwaN wrote:

  1. Website names, album names and format names are not acceptable sources. Is BMS acceptable as a Source?

I know that Wikipedia isn't a reputable source for scientific research but according to the BMS article I would see it as an acceptable source since they were made for this specific purpose like countless bemani songs have been made for their specific games.


Also, what bothers me since eternities and what I wish you'll hopefully consider adding to the draft

Please don't allow tags that have zero relevance to artist, song, album, mapper or anything else for that matter.
Saying you're "helping players find harder maps by adding an irrelevant tag to the map" is a pathetic excuse for littering your tags with memes


Despite my rant though, I see this as an issue, even if it doesn't appear to happen often (or rather rarely), it happens noticeably (e.g. there was a map which had it's whole lyrics in the tags iirc)
Monstrata
You wouldn't use wikipedia as a metadata source, so why refer people to wikipedia for metadata information? If you want an actual guide to the modified hepburn romanization system for Japanese, here's a much better site that's actually official: https://www.loc.gov/catdir/cpso/romaniz ... panese.pdf courtesy of the US Library of Congress.

In case you weren't sure, that document is the Modified Hepburn Romanization system. It doesn't specify it's the modified hepburn system, but it is, since the modified hepburn system is the default romanization system for Japanese according to the ALA-LC.

As for Korean romanization, also refer to: https://www.loc.gov/catdir/cpso/romanization/korean.pdf for the official standard.
Monstrata
Also for:

Cyrillic Romanisation: Use BGN/PCGN system for Russian/Cyrillic. "Е" and "е" should be romanised as "ye" if it stands alone or after "a", "e", "ё", "и", "о", "у", "ы", "э", "ю", "я", "й", "ъ", "ь". In other cases, it should be romanised as "e". Rule 1 also applies to ё, however, use "yo" or "o" to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. If Cyrillic is used to represent a word from another language, use the form from that language. For most of the other characters, refer to the first page of this document.

I wouldn't recommend BGN/PCGN system since the system is intended for geographical naming. BGN = US Board of Geographic Names, PCGN = Permanent Council on Geographic Names. The BGN/PCGN system is more applicable to geographic naming systems. Actually, there are specific rules for handling geospatial locations, GIS etc... For a site like osu! it's better to just use the romanization standard set by the LOC since this is the system Information Management professions and systems (Archivists, Librarians, databases) use. For the Russian/Cyrillic script, the romanization system is quite simple: https://www.loc.gov/catdir/cpso/romaniz ... ussian.pdf
rawrneru
This must be because of the whole Alveryn vs Asuka thing isn't it
show more
Please sign in to reply.

New reply