2025-06-04

field

defines non-metrical onset distributions: stochastic clouds, chaotic intervals, continuous-time drifts. used for naturalistic textures (rain, fire, random impulses) and non-cyclic materials. this domain generates timestamps only - i.e., when things happen. it does not define the sound itself. domains like noise, amplitudes, and envelopes shape what occurs at those times. each form outputs an explicit list of onset timestamps (in beats) over a finite window, resolved entirely at compile time. this domain is complementary to grid; together they span the full onset space.

overview

  • 0 explicit list

    • description: user-supplied onset times, verbatim
    • real-world analogy: hand-annotated score, foley cue sheet
    • notes: reference / bypass form

  • 1 poisson cloud

    • description: random events with constant average density
    • real-world analogy: light rain, geiger counter clicks

    • notes:

      • a controls rate
      • b controls window length normalisation

  • 2 logistic chaos

    • description: chaotic inter-event intervals derived from logistic map
    • real-world analogy: fluttering flames, unstable machinery

    • notes:

      • a sets chaos parameter r
      • b scales average spacing

  • 3 gaussian density curve

    • description: dense centre tapering to sparse edges (bell-shaped density)
    • real-world analogy: thunder burst, crowd applause swell

    • notes:

      • a peak position
      • b spread

  • 4 drift walk

    • description: random-walk deviations around a nominal period, accumulating drift

    • real-world analogy: human rubato, tape-loop wow & flutter

    • notes:

      • a nominal period

      • b drift strength

parameter behavior summary

  • 0 explicit list

    • a: unused
    • b: unused

  • 1 poisson cloud

    • a: mean rate λ (events per beat)
    • b: window proportion (0 → short, 1 → full segment)

  • 2 logistic chaos

    • a: chaos strength (r from 3.7 to 4.0)
    • b: mean spacing scaler

  • 3 gaussian density curve

    • a: centre position (0 = start, 1 = end)
    • b: width (small → narrow burst, large → wide spread)

  • 4 drift walk

    • a: base period (beats)
    • b: drift coefficient (0 = stable, 1 = high drift)
  • all parameters: a, b ∈ [0, 1].

why these were chosen

  • explicit list: passthrough for externally generated cues
  • poisson cloud: canonical memory-less random process
  • logistic chaos: deterministic yet aperiodic spacing
  • gaussian density: shaped bursts common in natural phenomena
  • drift walk: captures slow, accumulative timing variation
  • each form is perceptually distinct and cannot be reproduced by the others with only two parameters.

what is not included

  • continuous random noise at audio rate (belongs in synthesis layer)
  • feedback-responsive timing (forbidden by static-schedule rule)
  • infinite processes (all outputs are finite lists)

distinction from noise

  • field defines when events occur - the timing structure of impulses, bursts, or drifts
  • noise defines what those events sound like - the spectral character or broadband content

  • natural textures (e.g. rain, fire) are modeled using both domains:

    • field sets the timing of impulses
    • noise shapes the energy spectrum at each impulse
  • this separation enables clean composability between onset and timbre layers

conclusion

field fills the onset gaps left by grid, providing deterministic but non-metrical timing generators. with both domains active, the composer can author anything from strict drum grooves to chaotic rainfall - all within the project's static, parametric framework.