W3C home > Mailing lists > Public > semantic-web@w3.org > January 2006

Re: How about?

From: Oscar Celma <oscar.celma@iua.upf.es>
Date: Wed, 11 Jan 2006 18:30:37 +0100
To: love26@gorge.net
Cc: semantic-web@w3.org
Message-Id: <20060111183037.4013aa20.oscar.celma@iua.upf.es>

Here's another project that makes your dream come true.
Thus, "query-by-humming" problem its already solved (well, moreorless!) in the MIR (Music Information Retrieval) research field.

Another interesting problem is the "find similar music" from content-based attributes (the ones extracted from the audio itself). There's a nice demo at http://musicsurfer.iua.upf.edu that gives you similar audio,  given a track.
It's the next step for the music recommenders. That is, not only using collaborative-filtering or playlist coocurrences for music recommendation, but merging both worlds (context and content descriptors).

Cheers, Oscar.

On Tue, 10 Jan 2006 08:13:25 -0800
William Loughborough <love26@gorge.net> wrote:

> I can approach thousands of people and ask "what's the name of the tune 
> that goes 'dah-dah-dah-BOOM?'" and they will instantly respond 
> "Beethoven's Fifth".
> When will our machine indexing, etc. stuff allow me to do that with a 
> google-type search and a microphone?
> Love.
Received on Wednesday, 11 January 2006 17:31:43 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:44:55 UTC