Icicle Harvest revisited

Icicle Harvest is a 60-minute composition. While this might sound like it’s some kind of “magnum opus,” I think of it as just a long-form experiment. 

It’s interesting how people equate duration with quality and conceptual weight. An hour-long composition must have taken years to write! And yet I see many prolific ambient composers, churning out album after album. 

I’m sure some people consider this type of music to be lazy, as it often contains simple and repetitive material. But I think that composers in this genre are deeply invested in the sounds, subtle transformations, and the concept of time, even if they aren’t fussing over nested tuplets and minute gradations of dynamics in written scores. 


For this piece, I did not fuss over the score at all. In fact, I just wrote simple lead-sheets for each instrument. The material is quite elementary and I used parts of it in my Orff Ensemble (plus ukulele and guitar) piece Save the Castle! 

This screenshot is a spreadsheet showing some basic musical choices I made. The start time for each section is based roughly on “golden ratio” numbers–in other words, cutting sections by ⅔ and repeating to get different durations. I chose tempo in a similar way, then shuffled the order. I chose key areas close to C and c-minor for the main tempo changes. Then within those large sections I randomly chose more keys close to those “main” keys (by “random,” I mean I went to random.org and shuffled the lists).

“pre-composition” of Icicle Harvest. Scare quotes because it’s still “composition”

I composed several related (and very simple) rhythmic motives, then again used random.org to choose the order of motives, and which notes in the scale would be used. In each section, each instrument plays in the same key center, but they play different motives and different parts of the scale. I wrote out the notes for each instrument, like this:

piano “part” for Icicle Harvest

As you can see, this is very, very sketchy.


One of my main goals with this project was to use real instruments. It’s cheap and easy nowadays to make music with samples (see my “Restore” project which used a digital piano). I really wanted this to sound “authentic,” embracing the extraneous noises and imperfect tuning of the instruments.

I recorded this over winter break at East Carolina University. This was a great time to record, since there was very little activity on campus and because the AC wasn’t running constantly! I recorded the vibraphone, marimba, and celeste backstage at the main auditorium. I recorded the toy piano in a quiet lounge room. I recorded the piano in the recital hall. Unfortunately, the piano hadn’t been tuned since the last week of classes. But I went with it anyway–it fits with the out-of-tune toy piano.

Then I recorded long takes on each instrument. For each motive, I repeated throughout the pre-specified duration. To add variety in the way the motives overlap, I let myself change the amount of rest between each repeat.

Once I had these long takes, I did some editing to get rid of flubs and improve some of the timing. But mainly, I left it true to the performance, so it sounds spontaneous.

Is this ambient, and what next?

I set out to write an ambient piece, but I’m still not sure it quite fits that category. There are sections that certainly fit the mold–long sustained tones with sparse melodies. But some parts, like the opening, are very rhythmic. It almost sounds like Santa’s toy factory at the beginning, before going into more traditionally-ambient sections. 

Ambient music is an extremely wide genre, so I do hope this piece finds a place within it. I also would love to eventually see a live performance of this someday. I can imagine something with either the instruments around the audience, or vice-versa. If you want to do it, hit me up.

Leave a Reply

Your email address will not be published. Required fields are marked *