Instead of trying to scrawl down lyrics, doing a search on Google (only to be lead to some bizarre lyric website repository that may or may not work, all the while offering me free ringtones and ceaseless pop-up ads), with Shazam's app, I've actually found songs by, well, listening to them. Shazam is slick. You hear a song, start the app, hold your iPhone's mic up to the sound source and viola! It's been 'tagged' and identified like a wild bird you caught on Animal Planet. Another app by Midomieven lets you hum a tune in, search by title, artist and even lets you listen to samples.
Sending the sample for off-site sonic analysis, querying the proper artist remotely, album art and sending a response back to your iPhone in less than 10 seconds is pretty dang slick. The sample time itself is only about 12 seconds! Once you get your result, you can bookmark it, and, if you're in a WiFi hotspot, launch iTunes to buy the song. Most popular songs are on these services. Shazam doesn't work in loud places like clubs, bars or restaurants, but works well in cars or at home.
What these services haven't been able to do, however, is to analyze classical music. I've tried a few times. Shazam says the Beethoven Fifth Symhony is "unrecognized." What would Ludwig say? After humming the piece into Midomi, I got the strangest country folk song in response. This isn't surprising. There are very long phrases in classical music and it makes even die-hard fans puzzle as to "what was that piece?" Having these services decipher classical music presents a lot of challenges. First, recordings of pieces are almost nearly indistinguishable especially if you only had a 12-second sample of them. Also, unlike pop music, where there is one artist performing one song (sure there are cover versions), with classical you have hundreds of ensembles, conductors and performers spanning 50 years of audio recording doing the same 'song' over and over again. For example, there are more than 200recordings of Beethoven's Fifth Symphony alone!
When (or if) classical music can be deciphered by these types of services it will probably be a great engineering feat. Progress, if this is ever attempted, would probably be incremental. First, identifying a particular symphony, concerto or quartet is, in itself, very difficult. Next, identifying the particular recording will be staggering. All those ensembles and conductors perform pieces many times over leading to various versions and editions that would be a morass of data to sort through. While there's probably not a market demand for this type of service or application, it may serve as Shazam or Midomi's Everest.