[okfn-discuss] open film metadata database

Sat Apr 15 14:45:49 UTC 2006

On Sat, Apr 15, 2006 at 12:31:23PM +0100, Julian Priest wrote:
> but every conceivable player VHS betamax dvd dat etc. The media were
> stored in a huge carousel and when you requested a file through the
> web interface it found the media, put it in thr right player digitised
> it cached it and streamed it to you as mpeg2. There used to be one of
> these monsters on the 20th floor of centre point linked wirelessly to
> post production houses in Soho as offsite backup.

WOW. I want one.

I've done a *lot* of reasearch in the last few days into a specific
aspect of this issue: how to digitize, annotate, search and store video
which may include dvd rips, self-recorded material, transcriptions,
subtitles etc in a way that focuses on compatability and openess of
format, metadata and frameworks, protocols, and compliance with web
standards etc.

I don't know if this directly relates to Rufus' 'musicbrainz for DVDs',
but I were writing a piece of software were to be developed that chose a
format, generated metadata and sought a way to share the material this
project would form a large part of my plans:

http://annodex.net

The main website is misleadingly shallow and crap. I've been hanging out
on irc.freenode.net #annodex for the last 48 hours, asking lots of
questions and getting good response from what seem like a *really nice*
and capable bunch of mostly aussie developers. I like to think I'm
taking a leaf out of Jo's book in terms of testing the water of lots of
projects before diving in :)

check:

http://annodex.org (the annodex foundation!)
http://beta.annodex.org - the web2.uh-oh type website, 1 year into 'beta'
http://www.annodex.net/software/cmmlwiki/ - wiki video annotation!

and my personal favorite so far:

http://svn.annodex.net/scripts/trunk/ircCMMLBot.py - a bot for irc =>
cmml annotation of live streams.

This is a pretty well rounded solution, they've got a firefox plugin
media player as well as a couple of stand-alone efforts, not properly
working cross-platforms yet, but apparently stable and using VLC on
windows, and experimental on linux and macOS - which seems like a
credible solution that might yield results soon. Development is active.

They've got a server-side component; 
http://www.annodex.net/software/mod_annodex/ which seems very straight
forward..

And everything uses Ogg Theora and open protocols and standards
throughout. 

Commercial endeavors in the web2.0 sphere include such trite offerings
as http://eyespot.com and http://jumpcut.com - which are very poor and,
I think, destined to fail offerings for del.icio.us-like services using
low-res video.

Basically, server-side flash components can be made to do 'scratching'
across timelines, splicing and pseudo-editing, but flash is all there is
currently apart from annodex that does this. I have looked very hard and
have found nothing else. Perhaps I've not looked hard enough.

I've got about 1000GBP to sink into making this work for Talkaoke, The
People Speak and a few other projects... we are already digitizing 10
years of our conversational archive, so maybe I have a few small scale
bounties to put up already on this front.

Is anyone else interested, or know of interested parties looking for an
annotated media solution who might come in on a focused development
project?

X

S.