Demoscene music and MusicBrainz

Hello,

  Part of our digital heritage are recordings produced by musicians
vaguely affiliated with the Demoscene, a primarily european digital
art subculture around realtime multimedia artwork. As far as I can
tell, there are various projects dedicated to the preserval and, in
part, organization of them, but as far as I am aware, there is no
simple method that would let you take a sample of them and organize
them by a trivial thing like "artist" like you might drag and drop
some files into MusicBrainz Picard, hit scan, save, and are done.

Part of the problem is that much of the music comes as module files,
containers for samples and instructions how to play them, and the
formats often do not have appropriate meta data fields to carry in-
formation like "artist" in a easily recognizable form. A common way
to include credit information is to encode various comments in the
sample titles, so reading such information is hard and storing it in
the file without altering it too much is difficult aswell.

A related problem is that support for these formats is not as wide-
spread as one might wish, software might not exist or it might be
too much of a drain for the battery without proper hardware support,
so there are incentives to convert the recordings to formats like MP3
which would generally amplify the problem (you could store the origi-
nal as meta data, but as far as I am aware, there are no tools that
support this in a meaningful way).

So, recently I saw MusicBrainz intends to switch to EchoPrint for
music identification, and I am wondering if that could help curating
the hundreds of thousands of recordings that archives generally hold,
if it works well enough to at least provide a reasonably stable ID for
music without much additional meta data, that would at least solve the
problems around format choice and meta data storage to some degree. It
would then be possible to add meta data like artist and titles in some
distributed and semi-automated fashion, possibly by pulling data from
existing efforts together.

At this point I am interested in some brainstorming about requirements,
possibilities, problems, who to engage, and how to organize, and also
existing efforts that I may be unaware of.

A thing I would like to do is this: turn all related music I have into
MP3s, drop them into Picard which will then find recording and artist
IDs and titles where this process is reproducible so that it's easy to
locate the recording in its original form later. I note that this is
somewhat possible already, I checked Skaven's "Catch That Goblin!!" as
http://musicbrainz.org/recording/27c11a26-34b1-4778-988d-0720d27f7cee
simple test case. It's there already even with a PUID, and Picard does
find a PUID for my OGG-version of the S3M original, but it's different.
I could submit mine, but in the grand scheme of things, this is unlike-
ly to scale very well.

Also, from a demoscene perspective, the more interesting meta data for
this track in particular would be that Skaven published this when with
the demogroup "Future Crew" and that this recording was an entry into
the Multichannel Music Competition at the Assembly demo party in 1995
where it made the first place. These things do not fit very well into
the existing meta data model for MusicBrainz. That's not that much of
a problem though, if there is a recording id, linking would be easy. I
do wonder however, how well suited the MusicBrainz schema and community
would be to hold a rather large data set that's a poor fit with much of
the rest of the database, if this would indeed lead to the addition of
hundreds of thousands of recordings (realistically I would expect tens
of thousands, the top of the crop where people bother to do some work).

Thoughts welcome,
-- 
Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: +49(0)160/4415681 · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/

Received on Saturday, 22 October 2011 20:26:46 UTC