"Name That Tune" Search Engines

Just launched in the UK, Shazam is a search engine that plays “Name That Tune.” Cell phone owners dial a number and play a 15-second song clip (presumably from the radio or a club) into the phone’s receiver. After comparing a hash of the clip against their database of 1.2 million songs, Shazam returns their best guess via SMS text message.

The audio recognition algorithm was developed by chief scientist and co-founder Avery Wang. His 1994 thesis on sound separation is available for download, which provides some clues into how they may be extracting music from voice and other background noise. (More information about the service from the BBC, The Guardian, and Red Herring.)

How long before someone (Google, maybe) creates a web-based version that allows you to upload sound clips for identification? And contribute properly-tagged MP3s from your own collection? A truly comprehensive database of music would help people like Alan Taylor and all these other poor souls.

Comments

    How long until you can call in and hum the song or mangle the words and it will tell you what you heard? That would be impressive.

    I say it’s in the distant future. We need proper metadata-ization of the mp3 files, but almost no-one gives a damn about properly tagging his MP3s. “A world of exhaustive, reliable metadata would be a utopia”, as Cory so well put it.

    (By the way, if something like that ever happens, it would also help yours truly.)

    How cool – it looks like it doesn’t need any kind of metadata to work, just hold the phone up to a speaker. It doesn’t seem like it would be too hard to do something like that online at all. Great find!

    For the record, I’m saying that the need for metadata is the answer to the second question. (“And contribute properly-tagged MP3s from your own collection?”).

    Heck, you wouldn’t even need to use ID3 tags in a situation like this. If an MP3 matches the sound clip, just return the file for download. (Of course, this whole enterprise would require a ridiculous amount of storage and bandwidth.)

Comments are closed.