Monday, September 17, 2012

Criticism and Computers

This relates slightly to my post on the aesthetics of popular music. In a kind of case study, some Spanish researchers did a computer analysis of over 460,000 popular songs from 1955 to 2010. Here is the article in the New York Times. And here is a link to the original paper which has some colorful charts. Here are some of their conclusions:
we ... found three significant trends over time. Again using pitch code words, we counted the different transitions between note combinations and found that this number decreased over the decades. Our analysis also indicated that pop music’s variety of timbre has been decreasing since the 1960s, meaning that artists and composers tend to stick to the same sound qualities — in other words, instruments playing the same notes sound more similar than they once did. Finally, we found that recording levels had consistently increased since 1955, confirming a so-called race toward louder music.
 So, louder and with less variety. That's what it sounds like to me too. But the really interesting thing about this article in the NYT is the headline, probably written by an editor, not the authors:

The Computer as Music Critic

Of course that is precisely what is not happening here. In fact, that's the problem and it comes right out of what I was discussing in my post on aesthetics. You cannot dodge the problem of aesthetic quality when you are talking about music. Some songs, symphonies and sonatas are just better than others. Some are a lot better and some are a lot worse. Probably 90% of the music written in the Baroque era sits dusty on library shelves because it is not interesting enough, charming enough or expressive enough to be worth playing. This is a nice instance of Sturgeon's Law: 90% of everything is crap. What these researchers did was dump everything in the hopper, the 10% alongside the 90%. Without knowing a single thing about their methodology, I know that they will come up with no aesthetically interesting conclusions. We will learn nothing about good music because it has been put on a par with bad music. Computers don't sort these things out and I suspect they can't. They sure can count notes, though. It's just that that doesn't tell us much.

Here is a Lennon song they choose as an example:

No comments: