Midomi launched their music recognition service today - hum into your computer's microphone and the service will try to recognize the song.  Add words to the tune and it does even better.  People have been having mixed results, which is no surprise considering both the complexity of the problem and the fact that many people are somewhat tone deaf.  I haven't done so hot myself.

The blogosphere is full of comments about the lack of business model, how could a company like this get funding, the bubble is back etc.  I'm not so sure.  First, it would depend on how much money they got (anyone know - it's not on Venture Source yet). 

Second, maybe it is just me but there have been numerous times when I would have paid to have a song identified.  Sometimes, a tune gets stuck in my head and I desperately have to know what it is to buy it.

Third - if this works well, it is the way all music search should be conducted. Even if you know the name of the song and the artist, singing a 10-second snippet would probably be an easier way to bring it up.  With all the music stores trying to differentiate themselves  - wouldn't acquiring this technology be the perfect way?

Last, if the people at Midomi manage to solve this problem, what other speech recognition problems can they solve?