[ANN] iTunesTagger.rb 0.1 - a simple script to make use of the grouping field in iTunes

[Scott Parkerson]

"What this thingy will do is allow you to treat the arbitrary
"grouping" field as a comma-separated list of "tags", thus allowing
you to tag files as you would bookmarks at del.icio.us. This can aid
in creating smart playlists. Trust me."

More information over here on my blog:
http://www.smerpology.org/sprocket/article/957/itunestaggerrb-01-released

Requires iTunes 4.x on Windows (Mac OS version forthcoming) and Ruby 1.8.1.

Hopefully this is useful to someone.

Enjoy,
Scott

 

[Belorion]

"What this thingy will do is allow you to treat the arbitrary

"grouping" field as a comma-separated list of "tags", thus allowing
you to tag files as you would bookmarks at del.icio.us. This can aid
in creating smart playlists. Trust me."

This sounds very neat! Do I understand correctly that you can use
this tagger and the changes will be reflected next time you open
iTunes?

I was thinking of Ruby + music the other day when we had posts
announcing new classifiers (such as
http://rubyforge.org/projects/classifier/ and
http://rubyforge.org/projects/bishop/). I wonder if iTunesTagger + a
Ruby Classifier might be able to do automatic assignment of certain
tags. Maybe it would be possible to say "classify" the tempo of a
song as either being fast/slow to be able to distinguish between high
energy and slow music. Or tag anything with a certain level of bass,
etc.

 

[Florian Groß]

This sounds very neat! Do I understand correctly that you can use
this tagger and the changes will be reflected next time you open
iTunes?

I was thinking of Ruby + music the other day when we had posts
announcing new classifiers (such as
http://rubyforge.org/projects/classifier/ and
http://rubyforge.org/projects/bishop/). I wonder if iTunesTagger + a
Ruby Classifier might be able to do automatic assignment of certain
tags. Maybe it would be possible to say "classify" the tempo of a
song as either being fast/slow to be able to distinguish between high
energy and slow music. Or tag anything with a certain level of bass,
etc.

Hm, this is interesting. So basically you would need a way of converting
song characteristics into English words (perhaps the algorithms could
accept a stream of symbols as well?) which a Bayesian algorithm could
then work on. I wonder if there is any more complex algorithms for that.
(Tempo is a nice initial classification aspect, but I think that tempo
alone is not all that makes pieces of music similar or different.)

Interesting stuff.

 

[Scott Parkerson]

This sounds very neat! Do I understand correctly that you can use
this tagger and the changes will be reflected next time you open
iTunes?

Actually, you have to have iTunes open while running the program (it
interacts with the COM interface). Changes made with my program are
made instantly (i.e. the music database is updated on the fly).

I was thinking of Ruby + music the other day when we had posts
announcing new classifiers (such as
http://rubyforge.org/projects/classifier/ and
http://rubyforge.org/projects/bishop/). I wonder if iTunesTagger + a
Ruby Classifier might be able to do automatic assignment of certain
tags. Maybe it would be possible to say "classify" the tempo of a
song as either being fast/slow to be able to distinguish between high
energy and slow music. Or tag anything with a certain level of bass,
etc.

You could certainly do that if you know audio processing. I'm not sure
the COM interface would make sense to use; there may be a
visualization plugin API (haven't looked) by which you can obtain
access to the decoded bitstream.