my fingerprints are very similar
runs off the same idea
use ten's method but replace the costume code with your list
Wow! That’s both incredibly smart AND incredibly geeky!
No wonder ... according to ChatGPT there is no such actual catalogue - that may be false, of course - but the concept is from ...
Gödel Escher Bach
(I must have that book somewhere - I'm not that organized either 
).
just re-made my system the list for a 1:30 song is 809,108 long. i dont think i can shorten it more. i just need to speed up my comparing code
Oh, my GEB is ready-to-hand, in the bookcase next to my desk!
ChatGPT doesn't know everything; I own a copy of the catalogue and eventually I'll find it and will tell you the publication details. I'm pretty sure that it was published before GEB, so if the idea is in there (I don't recall it, but there are a lot of things I don't recall) Hofstadter got it from this book I'm talking about.
You're not comparing the entire song you're given against entire songs from the database, are you? You should take a two-second sample of the song and look for it within the songs in the database. And if you don't find it, take a different two-second sample.
Back in the day, when I was a DJ (original meaning) at my college radio station, we used to quiz each other by playing the first note of a song and the other person had to say what song it was. There's a huge amount of information in half a second of a song! (Of course, unlike the system used in this book we're talking about, the information in the first note has nothing to do with melody, but is more about timbre: what instruments are playing, and how are the musicians playing them. And if someone starts singing on that first note, the voice tells you a lot about who's singing.)
that would be slower
Umm, okay, I don't see why, but it's your project!
make an example that deals with the same high numbers that mine deals with. then compare speeds
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.