How does the music identifying application Shazam work?

As I understand it, it’s similar to how scientists date trees based on their ring structure.

If you look at a tree’s rings, you will see a pattern of big rings and small rings. A tree grows a ring each year, so the size of the rings corresponds to how well the growing season went in a particular year.

So, let’s say you find a tree with the following pattern of rings:

B-S-S-S-B-S-B-S-B-B

If you find that same pattern of rings in another tree, chances are both sets of rings were formed in the same stretch of 10 years. In fact, if you find any tree in the area with that pattern of rings (BSSSBSBSBB), you can be pretty sure that those rings all grew in the same ten-year period. Thus you can date trees in an area relative to each other.

Now think about a song’s melody. At any given point in a song, the melody either goes up, or it goes down. Right? You don’t need to know the exact notes, or the rhythm, or the words or instruments or anything. You can create a pattern of any song simply by listing a pattern of “ups” and “downs” for its melody. And just like tree rings signifying the same stretch of years, these patterns turn out to be pretty unique indicator of a song. U-D-U-U-D-U-D-D-D-D-U-D is probably a unique pattern for a unique song.

I’m not 100% sure this is how it works but I’m fairly sure.

Perhaps each of those tracks contain identification data that’s readable by the application, even though that data isn’t ordinarily perceptible to human senses.

An approximate analogy to this is that Fluther sees your IP, even though you ordinarily don’t.

Shazam is crap, by the way, SoundHound is far superior at identifying all genres, not just new crap that comes out, and has a user database.

These apps just work by identifying different notes and tempos and matching them up, also with SoundHound you hav even hum or whistle it.

If you want to really nerd out read this