machine learning - Music analysis software -

greetings

I might have thought this but no one knows what Last.fm first Any form used was to determine the same music to analyze on the music open source project.

As it has now been taken to a paid version, I would like to do something that can add known music to my playlist. (I hate to scan my computer for music similar to myself)

Do not believe - does anyone know that I can use it to change it? Ideally, I need some form of API / source code that I can use to automate the entire process in batch jobs.

Thanks,

[edit] Ideally I was looking for something else on the lines of content matching. I am the type of person who throws all his music into an unorganized position. Then being lazy I was hoping to make a playlist ideally, which gave me the same music type playlist.

Uses Last.fm - It provides access to your database via API.

[/ edit]

Music parallelism is not an easy problem.

Approach 1. Throw data to this problem. This approach takes the last FM and pandora. It is basically a large database maintained by any community or group of experts. Note that in order to use this approach you will need a clean metadata or some kind of audio fingerprinting solution. Once you have a feature database, you can use algorithms like to find similar items.

Approach 2 Throw algorithms to the problem. In particular, computer audition algorithms. This means that you calculate the different characteristics of the vectors in which one song is involved and using the nerve trap and other techniques you can find other songs with the same vectors. This approach has been used successfully and

If you are looking for open source software for music analysis, then what commercial goods can do, a lot can be done. On its brain child and its web site, you can find many papers about the situation of the computer audition.