Monday, September 3, 2012

Absolute pitch

The short story is, I don't have it.

Absolute pitch, formerly called perfect pitch, is the ability to name a pitch after hearing it. It's to be distinguished from relative pitch, which is the ability to name intervals after hearing two notes in succession, or at the same time; or copy melodies etc. There's different kinds: some people can do it in some contexts and not others. It's said to be rare: possessing "perfect pitch" is often included in the litany of inborn genius characteristics of the young prodigies, "one in ten thousand". People are studying it, though, and it's a little more complicated than just being a God-given gift to the chosen few. Learning music when being young helps, speaking a tonal language helps. Nearly everyone has some aspect: eg if you ask musically untrained members of the general public to sing a currently popular song, they start in the right key much more often than you would expect from chance. I've known people who had it, in a useful sense: my junior high orchestra director, for instance, would hear an airplane flying overhead, and announce that it was an E, but somewhat flat. And my son's AP is pretty good. I've never had the sense of having perfect pitch, but I feel like I do have some aspects of it. I can imagine playing any note of the bassoon, for instance, hearing the pitch and the timbre in my head, and imagining what my fingers and air column are doing. I'd be surprised if the pitches I'm imagining were that far off. And once when I was trying to learn intervals, with my wife playing them at the piano, she noted that I was always getting major and minor wrong, but was quite accurately calling white notes major and black notes minor. It's wrong, sure, but your brain can't make that kind of mistake without knowing, in some sense, what the absolute pitches are.

So I was intrigued when I read about an online study of AP. Here it is, out of UCSF. They are interested in the genetics, so there are survey questions about family background, musical training, and languages, before they hit you with an online pitch recognition test. I found the pace pretty quick, you get maybe one second for each tone, and I didn't feel like I was doing any better than guessing. About halfway through it switched from pure sine wave tones to piano notes, and I felt like I was doing even worse. I didn't care, I just wanted to finish. The results reflect that: I scored 14.75 on pure tones, and a 3 on piano tones; this is out of a best possible of 36. The average score in their test-taking population is a touch over 17, and random guessing would give you 7.5. Their cutoff for having AP is 24.5. There's a plot of where I fit in:

A couple of interesting things here: the responses cluster in two places: top right, where you mostly get them right, but with some errors; and middle lower left, around random guessing. So it does mostly look like a "you have it or you don't" binary proposition, though there are lots of intermediate people as well. There's a pretty good correlation between the scores for most participants: if you can do it for one type of tone, you can do it for the other. There are outliers, sure. Mostly these are above the line: people whose AP is pretty good when you play piano notes, not so good for sine waves. This could be a practice effect, people who get better at taking the survey during the survey itself (since the piano notes come second), but I think it's more likely people who grew up playing piano and can just recognize the notes better when they are played by a piano. People like me, who are better at the sine waves, are rarer. I'm not the furthest outlier in this direction, there's a dot way off in the lower right for someone who actually has perfect pitch on sine waves but was as bad as me at piano notes. He or she might have unusual characteristics, or perhaps they just didn't take the second part seriously. My personal results do look pretty funny: definitely better than guessing for sine waves, but worse than guessing for piano notes. Huh. I feel like if the survey had been bassoon notes, or if it has at least been notes in the bassoon range or my singing range, I would have done better. There were a lot of high notes, and I just feel totally lost up in the treble.

1 comment:

  1. I got sent to your blog by, believe it or not, the Botany Picture of the Day blog:

    However... absolute pitch. My daughter has it, or at least she used to, and she was part of the data set for that study you referred to.

    I don't have it, but if I hear music played on the clarinet (the instrument I learned as a child) I can often tell exactly what notes are being played. And my son said the same thing about the trumpet (his instrument). Which surprised me, because I couldn't do that at all with the trumpet.

    So I'm pretty sure that if you heard something played on the bassoon, you would know what notes were being played.