Making Musical Mood Metadata - M4

The M4 project is about making music collections easier to navigate and analyse by using the mood or emotion of the music. The primary application area is in TV and Radio production, but other fields such as domestic use would also apply. The main aspects and challenges of the project are:
* Extracting audio features from music to generate mood-related metadata to allow music tracks to be matched according to mood.
* Use existing mood metadata in production music to improve the matching process.
* Design a tool that returns a selection of mood-matched music tracks based on matches from a user's chosen 'seed' track.
* Adding visualisation of the music using colours and shapes to represent the mood to aid navigation of lists of music.
These developments will be incorporated into the Desktop Jukebox (a music database I Like Music provides for the BBC), which contains over a million commercial
and production music tracks. Programme makers will benefit from being able to find music based on mood thus reducing search times, increasing the variety of choice,
and making better selections. Tools for domestic use may be developed, such as creating playlists from personal music collections.