musical thoughts

Ligeti’s Atmosphères is a piece that had a profound impact on me. I actually studied the score before hearing it, and heard it outside the film score context where most people encountered it. I think I was in my mid teens at the time. I had access to a university library and a really hip professor that turned us on to some great modern compositions.

I’m interested especially in the way the usual musical devices were forsaken in this piece in favor of big blocks of sound, and the impact that had on the creation and release of tension. I’ve been looking for a way to do my own experimentation along these lines. I found a fairly satisfactory set of tools, and here’s my first effort at trying to adopt this sort of musical vocabulary.

In my piece, meter isn’t a factor. Individual voices enter and exit in a way that builds to a climax by manipulation mainly of pitch and texture.

Each voice is a sound that occupies a bit wider spectrum than a pure pitch might–you could look at them as bandlimited noise. Right now I won’t describe how I created the voices, but one of my bigger challenges in doing this type of project is that a stack of synth voices sounds fairly crappy, and lacks the impact of a section of real instruments.

Other than the scaling on the pitch axis in this spectrogram (linear, unlike the equal octave scale on the piano roll above), you can see the spectral shapes I created get rendered pretty well. The constant band running from 3.5kHz to 4.5kHz isn’t shown on the piano roll–it’s a separate track.

I think you can spot the climax. I wouldn’t say my use of this idiom is particularly subtle. However, the volume associated with that climax is not dramatically louder than the buildup–it’s a massive change in texture though, along with a peak in register.

The cricket sound is the same “voice” as the entire rest of the piece. This was all developed from a single sample (not a cricket sample either, it was a synthetic source).