2025-06-04

variation

defines both stateless and stateful transformations to motifs, sequences, or parameter sets. as part of the interplay layer, variation spans quick, one-time changes as well as gradual, deterministic evolution - all without runtime reactivity. each form is fully resolved during synthesis planning, preserving repeatability and clarity.

introduction

variation includes both:

  • stateless variation: one-pass, non-accumulating modifications (e.g., jitter, filter)
  • stateful mutation: deterministic evolution across iterations (e.g., drift, reset) all behaviors are implemented through compact parametric forms, use no internal runtime logic, and operate in a domain-pure fashion - affecting only their target structure.

overview

stateless forms

  • rule-based

    • behavior: applies a fixed transformation rule to all elements
    • analogy: genetic mutation with known effect
    • a: selects transformation type (e.g., rotation, inversion)
    • b: controls depth or intensity

  • jitter

    • behavior: applies deterministic pseudo-random offsets
    • analogy: humanized performance
    • a: max perturbation magnitude
    • b: density of affected events

  • filter

    • behavior: probabilistically retains/removes events via seeded thresholding

    • analogy: sieve with adjustable holes

    • a: inclusion probability

    • b: clustering bias (uniform → grouped)

stateful forms

  • accumulated drift

    • behavior: small additive perturbations on each repetition
    • analogy: detuning or temporal wobble
    • a: mutation step size
    • b: maximum drift clamp

  • rotating sequence

    • behavior: cyclically shifts the sequence over iterations
    • analogy: conveyor loop
    • a: number of elements shifted
    • b: direction bias (0 = forward, 1 = backward)

  • segment swap

    • behavior: exchanges subsections at regular intervals
    • analogy: shuffling cards
    • a: swap size (fraction of total)
    • b: swap frequency

  • value mapping

    • behavior: remaps values via deterministic function
    • analogy: applying a static lens
    • a: mapping type
    • b: intensity (blend amount)

  • reset clock

    • behavior: periodically returns to original template
    • analogy: return to home state
    • a: reset interval
    • b: reset depth

  • template switch

    • behavior: changes underlying source template

    • analogy: change of strategy or tone

    • a: switching rate

    • b: degree of morph between old and new

parameter behavior summary

stateless

  • rule-based

    • a: transformation type
    • b: intensity

  • jitter

    • a: perturbation size
    • b: percentage of affected values

  • filter

    • a: base retention probability
    • b: grouping of passes/fails

stateful

  • accumulated drift

    • a: step magnitude
    • b: drift bound

  • rotating sequence

    • a: shift amount
    • b: directional weight

  • segment swap

    • a: swap size
    • b: frequency

  • value mapping

    • a: mapping index
    • b: interpolation amount

  • reset clock

    • a: reset interval
    • b: reset extent

  • template switch

    • a: rate of switching

    • b: morphing degree

why these were chosen

  • unified abstraction: covers both transient and evolving variation in one domain
  • irreducibility: each form defines a unique mode of structural change
  • perceptual relevance: all parameters correspond to musical sensations (e.g., wobble, decay, return)
  • compactness: 9 forms span a large expressive space
  • determinism: even pseudo-random behaviors are fully precomputed with fixed seeds

what is not included

  • feedback, runtime adaptation, or reactive logic
  • probabilistic models with history (e.g., markov chains)
  • onset-level variation - this is covered by offset, pressure, role, and activation
  • timbre, spatial, or amplitude envelope changes (defined in their own domains)
  • unbounded or entropic systems - all mutation is limited and bounded

conclusion

the variation domain provides a complete and structured framework for altering musical material in both immediate and progressive ways. stateless transformations offer instant diversity, while mutation forms yield long-range, non-repetitive motion. together, they form a central creative axis for structured evolution without sacrificing clarity, determinism, or modularity.