Ligeti's Method

The most important difference between Xenakis' conditional random processes and Ligeti's micropolyphony lies in the micro-structural level.

While random processes as well as deterministic ones deal with the overall density, sound-color and density aspects of the sound-cloud, all musical elements: pitch, rhythm, range, timbre and so on are partly (in random processes) or totally (in deterministic processes) under the control of the composer.

However, when each part is considered alone horizontally, stochastic processes do not show any kind of micro-structure.

In Ligeti's micropolyphony, by contrast, each part has a structure of its own.

Figure: G. Ligeti Atmosphères, Violin I-1, Bar:44. Despite being only one component of the 48-voice polyphony, this line of the number one of the first violins still shows a clear melodic structure.

19#19

It is striking that every ``horizontal slice'', as well as every single part, of the music will show a supremely crafted polyphony.

The cello section of the same bars, together with the violas are actually playing a mirror canon of the violin parts. When this cello section is examined on its own, it reveals the coherent micro-structure of the polyphony.

Figure: Cello section, 1 to 10 Bar:44

20#20

However one may ask the question ``is this audible and if so in what sense?''. It is obivous that the structure shown, one part of the forty-eight voice polyphony, is literallt buried and would never be recognized in the mass of sound.

The point in crafting each part of the micropolyphony to such an extend is to control the shape and evolution of the ``soundscape''. Individual notes and parts, altough not discernable in their own, are creating the characteristics of the ``colour'', ``shape'' and ``inner activity'' of the soundscape.

Range, rhythm, dynamics and intervals for each part are not meant to be heard as such but rather to be a part of the overall design.