[Playlist] content resolution and musical/cultural aspects of a piece

[Lucas Gonze]

On 11/22/06, Bjorn Wijers <bjorn at simuze.nl> wrote:
> isn't the resolving harder to do with less popular or less wellknown
> songs? As in virtually impossible and therefor almost useless for lets
> say netlabel music? Or am I missing something?

That's a great set of issues to bring up -- what is the interaction
between content resolution and musical and cultural aspects of a
piece*?

In one sense content resolution depends on the distinctive features of
the music.  Is this song much longer or shorter than the average?
Does it use text which is uncommon?  For genres which rely on
minimizing the unusual and maximizing the formulaic, there isn't a lot
to work with.

At the same time, it's very reasonable for content resolvers to
incorporate popularity data to weight similar content, and this would
work to the advantage of more popular and well known songs.

There's also the issue of playlist context.  Let's say you have a
netlabel song which would normally lose out to a major label song, but
the containing playlist is all netlabel material rather than stuff
from major labels.

There's the issue of social context.  Let's say you have a netlabel
song which would normally lose out to a major label song, but the
netlabel song is more popular within the playlister's social circle.
For example, the playlister might subscribe to a MP3 blog which
mentioned the netlabel song.

Lastly, there are all the normal dynamics of web search.  Netlabel
material is out there competing for page rank, while major label
material is hidden away in darknets, so the netlabel stuff has an
advantage.

-L

* For myself, music is my technical focus and hence my default medium,
but this is a personal thing.  Please substitute "videoblogs" or
"claymation" or whatever else as you see fit.