Subtitling and XML format (was: Karaoke)

Johan Pouwelse (pouwelse nospam at pds.twi.tudelft.nl)
Thu, 23 Sep 1999 10:55:01 +0200 (MET DST)

> > This is my first Email to this list, together with three different people
> > we are working on a system to extend cdindex with scrolling lyrics.
> That is indeed an useful extension of the present data.
The idea of karaoke in mp3 is not new, but using a central server for
download AND upload is.

> a kind of 'subtitling' for songs.
Wow, thanx that's a good name. Scrolling lyrics is to long, and karaoke
has the image of some druk japaneese in a obscure bar singing out of
tune..
table names: artists, album, track, subtitles..

> > We know this has some legal problems..
> What solution do you aim for minimizing troubles?
> I remember problems with a swiss site being closed down.
Yep, the audio industry is defending this stuff. Only if it is on a server
, not a small homepage with only lyrics of a single song.

> - Having an cdindex server with extended capabilities (thus having
> karoke fields in the database)?
> or
> - Lyrics data only on dedicated karaoke servers in secure countries
> (like Holland and Russia) that look up the CD ids and other
> stuff from available CD Index servers and add their karaoke information?
The first.
I'm currently thinking of using NEWS would be nice. Imagine a special
news group with (non-compressed) XML posts of subtitling data and the
reference to the track index. This would leave the message sender legally
responsible, caching news is not illegal.
example:

From: submittor nospam at hotmail.com
To: alt.audio.binairy.cdindex.submission
<subtitles CDIndex_Track_ID=100243>
<timestamp time="00:05.10" text="When"/>
<timestamp time="00:05.10" text="I'm"/>
<timestamp time="00:05.10" text="feeling"/>
<timestamp time="00:05.10" text="blue"/> ....
</subtitles>

This news group is a broadcast to which everybody can freely submit and
listen, can be used also to sync different servers. When messages are not
compressed, they can remain the the news format and can be keyword
searched.

XML format
In the current XML format the artist tag is duplicated for every entry. If
we want to include the subbmittor in the future we're stuck with the same
problem. If you want human read XML, that's what you want.
Would this format be an improvement? This format better models the
database format, yet can still be extended. The multiartist CD is no
exception anymore, but modeled as putting the <artist_ID="0001"/> tag
in the track or the album tag.
Any remarks on this?

example:
<CDIndex>
<artist Artist_ID="0001">
<name>The greatest of them all</name>
<Email>GodOfRock nospam at hotmail.com</Email>
</artist>
<artist Artist_ID="0002">
<name>RockBitch</name>
<official_URI>http://www.rockbitch.com</official_URI>
</artist>
....
<album Album_ID="0001">
<name>The greatest part II</name>
<artist_ID="0001"/>
<submittor_ID="0023"/>
</album>
<track album_ID="0003">
<name>Run baby Run.</name>
</track>
<subtitles track_ID="0003">
<timestamp....
.....
</subtitles>
</CDIndex>

> > I hope you guys support the idea.
> Van mijn kant zeker.
Great

Johan.