forum

[Proposal - Metadata] Clarify which YouTube uploads valid as primary metadata

posted
Total Posts
23
Topic Starter
Ryu Sei
As you know, there are three kinds of YouTube uploads that may exist at the same time:
  1. Artist or label's official YouTube upload
  2. YouTube Music
  3. Auto-generated YouTube uploads (i.e artist's topic channel)
Only the first one counts as primary metadata source. Current description doesn't clearly convey which YouTube uploads counts as official, so I propose a clarification regarding that.

Current entry

An artist's website, their label's website, CD scans, and official uploads to websites like Bandcamp, YouTube, and SoundCloud are considered primary sources for beatmap metadata. Third party websites such as wikis, databases (vndb, vgmdb, etc.), and music services (Spotify, iTunes, Amazon, etc.) do not always reflect the artist's intention and should not be treated as primary sources.

Proposed changes

An artist's website, their label's website, CD scans, and official uploads to websites like Bandcamp, YouTube, and SoundCloud are considered primary sources for beatmap metadata. Third party websites such as wikis, databases (vndb, vgmdb, etc.), and music services (Spotify, iTunes, Amazon, YouTube Music, artist's YouTube topic channel, etc.) do not always reflect the artist's intention and should not be treated as primary sources.
This should detach these two YouTube uploads from primary metadata source.

I'm open for more rewording.
TheKingHenry
the existing wording already says "official Youtube uploads", and neither Youtube music nor topic channels etc. are official, so I don't personally see the point in the redundant addition

on additional note,

Ryu Sei wrote:

label's official YouTube upload

RC wrote:

their label's website
I always disliked using label sources to begin with, they get things wrong all the time either by accident of "intentionally" for marketing purposes (all caps usage and such is a common one). They represent themselves so to say, not necessarily the artist as the artist intents. I'd totally advocate for treating labels as a secondary source compared to anything directly handled by the artist themselves.

On the other hand to be fair though, similar marketing effect "distorting" metadata can also be seen relatively frequently in the first party official socials too. As crude cut example "official music video" or similar are technically there in a lot of Youtube material, but should obviously be ignored. Descriptions tend to be more reliable, but I would just recommend against using Youtube (and similar sources) for metadata regardless, and go for better sources that are almost always there

anyhow in my opinion the rules seem alright, and which sources among the technically valid should be prioritized can be left up to common sense like it has been. Though I suppose common sense isn't always terribly common ¯\_(ツ)_/¯
Topic Starter
Ryu Sei
Not everyone knows that YouTube Music and YouTube topic uploads are not the primary metadata source. More clarity is always better.

There is this clause on most YouTube auto-generated topic uploads, which is "Provided to YouTube by (label name)". For most people, they assume it's the official ones because "the label provides it to YouTube", while in the truth the label only provide YouTube the song, and let YouTube themselves curate the metadata.
SaltyLucario
while henry is correct that this is handled by rc already, i'm all for making this clearer with how many people use auto-generated topic videos as their only metadata source
h3oCharles
pretty sure auto-generated YT uploads are made for YT Music
and YT Music should be treated as a music service

agree
-White
Seems unneeded. Bns will continue to make sure correct Metadata exists
Topic Starter
Ryu Sei
There will be always a hindsight. I've seen some maps using YouTube Music as primary metadata.
FuJu
There will always be people using wrong metadata for their sets. F.e Spotify is mentioned in the RC and I have seen people use it as their primary source before aswell. Current wording is clear enough so doesn't need a change imo.
Topic Starter
Ryu Sei
As I mentioned before, YouTube case is specific. People who don't know the differences between official and curated uploads will definitely assume the YouTube Music uploads (we'll use that term for both YouTube Music itself and auto-generated channels from now on) is official because "YouTube counts as official".

It is technically an "official upload" by the artist or label, but the metadata given by that is the same treatment as the secondary, so it's unreliable at its best.
Protastic101
Maybe rather than change the RC ruling itself, providing some examples of primary sources and examples of non-primary sources might be the way to go? That way people are able to see what a primary source might look like on any of the aforementioned platforms.

The examples could be for some popular existing FA artists too so it might even help people as a starting point to find metadata in the future (thinking like Camellia's official YT for example)
FuJu
But that is the point. Its not an official upload by the artist, which is also clearly noted in the description of topic uploads.
-Flashlight-
I've seen people mistaking "Topic" uploads as official/primary sources for metadata a few times, so it's nice to clarify this point. +1
Drum-Hitnormal
video uploaded by artist on youtube could be auto translated and not intended english title
Topic Starter
Ryu Sei
Sure, here's the example:



The YouTube topic page shows incorrect metadata that we can explain one by one:
  1. There is no trailing space between character name, : symbol
  2. Character names written in FN LN, not standardised as LN FN
Serizawa Haruki

Drum-Hitnormal wrote:

video uploaded by artist on youtube could be auto translated and not intended english title
This is a bigger problem I think because it's not as noticeable as Youtube Topic uploads (but clarifying
that doesn't hurt), maybe it could be mentioned somewhere too? If I remember correctly niconico also displays automatic translations sometimes.
Topic Starter
Ryu Sei
Perhaps we can put it to one umbrella term for auto-generated videos/metadata?

Third party websites [...] and music services (Spotify, iTunes, Amazon, auto-generated YouTube entries and similar, etc.) do not always reflect the artist's intention and should not be treated as primary sources.
Topic Starter
Ryu Sei

Drum-Hitnormal wrote:

video uploaded by artist on youtube could be auto translated and not intended english title
Oh... right. I remembered that YouTube allows uploaders to set their own title for audiences using certain display language. For example:


On my Indonesian setting, it shows "THE SIXTH SENSE/THE SIXTH SENSE", which definitely sounds ambiguous and incorrect than the intended original title "第六感", and official translation "THE SIXTH SENSE". Do note that when searching on YouTube, sometimes title and description are auto-translated instead of using artist's preferred translation, which is another problem.

To reiterate, auto-generated uploads and descriptions should not convey the artist's intention, so we need to detach these from official YouTube uploads.
aceticke
No thanks, share same sentiment as FuJu here. The Primary metadata source wiki page already details that YouTube videos must be official. If anything, the wording should be clarified in this article, and not in the Ranking Critera.
Topic Starter
Ryu Sei
I thought we're going to clarify the wording of that article, not adding more guidelines?
aceticke
This is definitely the wrong place for this then, having it in the Ranking Criteria subforum is misleading
Topic Starter
Ryu Sei
I'm not sure where to post it, so I posted here instead.
radar
if a change would impact something to do with the ranking process, then i see no reason why it cant be on the ranking criteria subforum. there have been other posts placed here about other mapping ecosystem related proposals

edit: that being said, after actually reading through the proposal it seems like what you're looking for is less of a proposal and more of a clarification to the preexisting wiki article. this is probably a topic more suited for osu!wiki people, so will forward it there
Okoayu
i think we should just write an article about what sources we consider to be primary and what the caveats are around each type; draft currently says primary metadata soruces and links a stub article which i think we can just expand instead of putting it on the RC itself.

I think this is a good todo to have so I added this to the agenda :D/ archiving this for now
Please sign in to reply.

New reply