Name That iTune

iTunesThe current generation of iPods can be had with a hard drive as big as 40 Gigabytes, and the iPod Photo can be had in a 60 Gigabyte model. If you don’t limit yourself to the storage available in a portable digital music player, you can have an exponentially more capacious warehouse for all of your MP3s on your computer’s hard disk — a desktop computer with a 250 Gigabyte internal drive is not uncommon these days. That’s a lot of music. So it occurred to me this morning, when the chorus of some half-remembered song popped into my head inexplicably and then haunted me all the way to the office, that iTunes and iPods — or whatever substitutes you care to name — still don’t allow you to find that one song that goes something like “Doo de dum de doo…” You know the one I’m talkin’ about?

That is, all of these digital music management systems allow you to search for a song by meta data in text form — artist name, song name, album name, genre, date, etc. — but you still can’t find a song that you remember for its most fundamental characteristic: what it sounds like. Maybe it’s impractical to ask a computer to recognize the few bars of a tune that you can hum, but there must be some other way to identify a track you have locked in your head (I remember seeing plot-finders that allow you to find similarly remembered/forgotten movies a few years ago, but they seem to have gone away) — and anyway, why can’t you just hum into your computer? With speech recognition continually ‘just around the corner,’ this would seem to be an ideal use for it. Wouldn’t that be incredibly handy? Moreover, wouldn’t that be an effective way to sell more songs on the iTunes Music Store?



  1. Musicbrainz already does this somewhat. It uses sound signatures to identify songs you already have for tagging. It works really well most of the time.

    If the song isn’t in the database though, you can sometimes get a match for something that sounds close to it. Evidently, some of the music from “The Sims” sounds like Cream. Who would have thought?

  2. Here in the UK, Shazam – – offered a service where you could dial their number on your mobile phone, hold the phone up to the music and they would then text you a note of the song.

    I don’t know what kind of system they used at the other end, but it was obviously capable of recognising the song. Presumably it would need a fairly big database to have any sort of success rate though.

  3. MIT is developing a search engine for music to where you can hum the tune of the song you are looking for and it will find the closet match. I don’t have a link but I saw it on the History Channel.

Thank you! Your remarks have been sent to Khoi.