Audio/Acoustic Fingerprinting
-
Hey everyone, Question... is anyone aware of an SDK or a library that will do audio/acoustic fingerprinting? If not; is anyone willing to help write one? *grin* In case you're not aware of what I'm talking about - it's the same technology that you see nowadays in the iPhone or what have you ... you hold your phone up to the TV or radio while you listen to a tune that you like; but you're not sure who the song is by... give your phone about 30 seconds, and it returns back to you the name of the song and the artist. Is there any open source library or SDK that anyone is aware of that could get us coders started on some projects? I'm not interested in creating another "what song was that?" recognition application; but rather something different. Anyways, having such a library would be awesome. Or, if anyone is willing; could someone give me a few pointers on what it would take to write such a library? I'd imagine there is some sort of hard math involved to figure out patters and peaks; etc to a song. Does anyone know of the research that it would take to make it happen? Thanks for reading - Brandon.
-
Hey everyone, Question... is anyone aware of an SDK or a library that will do audio/acoustic fingerprinting? If not; is anyone willing to help write one? *grin* In case you're not aware of what I'm talking about - it's the same technology that you see nowadays in the iPhone or what have you ... you hold your phone up to the TV or radio while you listen to a tune that you like; but you're not sure who the song is by... give your phone about 30 seconds, and it returns back to you the name of the song and the artist. Is there any open source library or SDK that anyone is aware of that could get us coders started on some projects? I'm not interested in creating another "what song was that?" recognition application; but rather something different. Anyways, having such a library would be awesome. Or, if anyone is willing; could someone give me a few pointers on what it would take to write such a library? I'd imagine there is some sort of hard math involved to figure out patters and peaks; etc to a song. Does anyone know of the research that it would take to make it happen? Thanks for reading - Brandon.
My business partner implemented one to sort identify duplicates and classify images/music. The main issue you will face is that this area, practically, has the hell patented out of it. So don't expect to be making a commercial release ;) As for open source... the Neuros digital audio computer had an Hear It Save It feature, similar to the iPhone (but this was around 10 years ago), and open source firmware (not sure if this algorithm was a binary blob or not). You might want to look there. Yes, the math is hard. You will be needing a good understanding of audio transforms, some black magic in recording energy transistions and compressing this all down into a shortish key you can easily index while still doing partial matches. The iPhone (like the Neuros) will probably use a third party service. I imagine they will make the client fingerprinting libraries avaliable for little cost, and charge a small amount per lookup.
Mark Churchill Director, Dunn & Churchill Pty Ltd Free Download: Diamond Binding: The simple, powerful, reliable, and effective data layer toolkit for Visual Studio.
Alpha release: Entanglar: Transparant multiplayer framework for .Net games.