On January 5th, 2006, Disney filed a patent on a portable media player that would sense the user’s mood by testing their heart rate and body temperature, and use that to attempt to select appropriate music. Neat idea, right? Some friends and I had the same idea half a year ago, and discussed it on a public IRC server. The following conversation took place on July 30th, 2005. (Some unrelated bits of conversation have been removed for clarity.)
[07:30] <Ryan> Also, I had an idea for a mood-sensitive hardware mp3 player. To turn it on, you press your thumb on a temperature/heart-rate sensor, and then it plays songs and you tell it on a 1-5 scale how much you like that song at that time.
[07:37] <luser_> heh
[07:37] <luser_> i was thinking about a sort of mood-playlist generator
[07:37] <luser_> 1) get a bunch of data from somewhere about similarity of songs in your library
[07:37] <luser_> like same artist is similar, same album is very similar, same genre is similar, etc
[07:38] <luser_> 2) Start playing random songs
[07:38] <luser_> 3) If the user skips a song, lower that song’s score, and lower the scores of all songs connected to that song
[07:38] <luser_> 4) Try to pick a highly rated song for next in the list
[07:39] <luser_> assuming you can get useful similarity data, you should start to get similar songs pretty quickly
[07:39] <luser_> with very little user input
[07:40] <Ryan> You could use the Amazon API to find similar artists/albums.
[07:40] <luser_> ooh
[07:40] <luser_> that’d be hot
[07:40] <luser_> anyway, that’s mostly how we use the iPod in the car
[07:40] <luser_> put it on random, and skip stuff we don’t want to hear
[07:54] <Ryan> This will only really work if you can correlate it with mood somehow. Your algorithm didn’t actually take that into account.
[07:54] <luser_> eh
[07:54] <luser_> you determine the mood from what songs they skip
[07:55] <luser_> i’m looking for the lowest amount of user interaction
[07:55] <luser_> all you’re saying is basically “don’t play songs like that”
[07:55] <Ryan> Ah, I guess that works.
[07:57] <Ryan> I still like the idea of correlating it with body temperature and heart rate.
[08:02] <luser_> correlating with body temperature and heart rate would be crazy
[08:03] <luser_> the cool thing about my idea is that you could theoretically get a decent listening
experience out of a flash mp3 player with just “Play” and “Next Track” buttons
[13:21] <Vito`> luser…: just skipping songs you don’t like works if you already know the song.
[13:21] <Vito`> 1-5 rankings are dumb, should just be thumbs up/thumbs down
[13:21] <Vito`> but if it’s a lot of new music, I, at least, tend to listen to the entire track to at least give it a fair shake.
[13:22] <Vito`> but that’s me
[13:22] <Vito`> I like the idea of skipping a track to drop it in the ratings
[13:33] <luser_> i’m thinking where you already have a library of music
[13:33] <luser_> and you just want a playlist that fits your mood
[13:33] <Vito`> Yeah.
[13:33] <Vito`> No, I totally get it, just two buttons is great.
[13:33] <luser_> getting the data should be pretty easy
So does this count as prior art that could be used to challenge the Disney patent? I doubt it, but if anyone out there is a patent lawyer and wants to take on the case, be my guest. I’d love to see Disney get a little less money!
