Cricket Music

And always let your pitch pipe be your guide.
– Milton, Paradise Lost

Everybody, and their buddies, it seems, is linking to this recording which claims to be simply cricket chirps slowed down a whole bunch. It’s kind of pretty.

I’m not buying it – ‘it’ being that it’s just cricket chirps slowed way down without any other treatment. One big problem: unlike just about any other natural sound – say, bird songs or whale songs, for example – the cricket chirps contain only notes in the diatonic scale. What this means is that we hear a melody and harmony, while in all other natural sound we get the full array of pitches *between* the notes of the scale. Not probable. Natural sounds are just a lot messier than Western music. Happy to be proved wrong, of course.

There are an infinite number of pitches between any two pitches, say, C and C#. The untrained human ear can typically distinguish at least quartertones – tones halfway between C and C#. Highly trained or gifted musicians can hear a lot finer gradation than that. Yet here, all we hear are bugs remarkably on pitch. It’s just too tidy.

Ornithologists of a musical bent have long tried to jot down bird songs using something like Western notation. Sometimes, such a musical score is close enough to remind one of the actual bird song. But recordings of bird song analyzed closely reveal that those darn birds just don’t limit themselves to the notes of the scale, but rather slip and slide through all sorts of pitches with no obvious preference for the ones that make sense to us.

This sounds more like what you’d get if you ran the slowed-down chirps through something like Autotune (a gadget that takes the pitches of ‘pitchy’ singers and digitally moves them to the nearest ‘real’ pitch). It even more sounds like Phillip Glass, if he wrote for a Russian Orthodox choir.

A musician could actually notate this music very closely – there are easily identifiable melody fragments and clear pedal points and an internal ‘alto’ harmony drone-ish part, as well as the occasional high counter melody. (I’d do it, but I’m betting better musicians than I are already on it.) That this could be fairly cleanly notated in standard Western notation would make it, as far as I know, absolutely unique among complex natural sounds.

So, somebody prove me wrong, already!

Author: Joseph Moore

Enough with the smarty-pants Dante quote. Just some opinionated blogger dude.

4 thoughts on “Cricket Music”

  1. First of all:

    With Bonnie Jo Hunt:

    The NPR Transcript:
    The Documentary:
    Robert Wilson:

    The soundcloud doesn’t appear to be either Jim’s or Robert’s. But that first article doesn’t seem to be exactly correct. The soundcloud version is “digitally remixed and remastered.”

    I’ve done some analysis on it. It repeats 4 times (the woman’s voice returns near the beginning.) Within each repeat, the same “verse & chorus” are repeated several times. Of all of that, I’m highly confident that we’ve got about 2.5 minutes of still very self-similar audio.

    At ~400% speed I’m hearing lots of hum and tape noise. That, reverb, and ambient sound could account for alot of the base harmonics. Much of the detuning would be explained by slips and stretching in the tape while being dubbed at slow speeds.

    Depending on how the audio was slowed down, it could lead to more harmonics in the transients (overtones in the waveform). I’m not convinced there aren’t human voices on the soundcloud version, but the audio as it stands could probably be produced by layering sound. I’m not sure what equipment an audio engineer would have had access to in 1994 (I first saw a personal computer that year – this is not done digitally though I hear mp3 artifacting as well), but I do know he would have had a multitrack tape machine. I think what we have here is multiple loops of the same crickets being played overtop of each other at various speeds, giving us the melodic content. That said, I can’t quite account for all of the sounds.

    So the question is: what is the actual original audio, and does the NPR audio differ? I’ll need to spend some more time on that!

      1. Thanks for the research – I couldn’t find anything in the time I had to spend on it. And speeding it up is a good test – that you got tape noise and hum suggests it could not be *just* slowed down crickets – unless you slowed it down and recorded the results, then slowed down the second recording and recorded that, rinse and repeat until something interesting happens….

        Still think the craziest part is the diatonic scale,and the I-IV-I harmony – why would anything in nature correspond to something artful as that? Even the music of most human cultures isn’t that harmonically sophisticated.

        It occurred to me that the if the process of slowing it down dropped the original notes below audible level, and if the overtone series was just right, one could conceivably get some interesting diatonic-sounding overtones – but that doesn’t explain why they would be arpeggiated instead of stacked. And then there’s artifacts of the process – I’d find it easier to believe that that’s what we’re hearing – a loose ball bearing in the capstan, or a jitter in the play-back head.

        It’s still a mystery to me. Thanks for doing the research. Please let me know if you find out anything more.

      2. I’m tempted to buy the CD. I have a feeling that the NPR section sampled in the youtube video is just a part of a longer song, just as the 2:30 of actual audio from soundcloud is sampled from a different part. I’m pretty sure the arpeggiation is just samples being layered in at different speeds… I might try to find an email address for someone involved!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s