data.matrix Glitch Grid (Binary Cellular Automaton + Beat Faders)

Math essence

state[i,j] ∈ {0,1}^G×G. neighbour_count = Σ state shifted in 8 directions. next = (state ∧ (n ∈ {2,3})) ∨ (¬state ∧ (n == 3)). glitch_mask = random < glitch_rate; flip_mask = random < 0.5; next = where(glitch_mask, flip_mask, next). if random < invert_rate: next = 1 - next. fader_bar[i] = |sin(2π*(0.4 + 0.13 i) * t + 0.7 i) + 0.6 beat|. canvas[y0:y1, x0:x1] = fg_color * state[r, c]; faint noise floor += random * noise_floor * 255.

Parameters

Reference

First 60 lines of reference.py. Full source: library_external/audio_reactive/IkedaGridMatrix/reference.py

from __future__ import annotations

from typing import ClassVar

import numpy as np


class IkedaGridMatrix:
    """Ryoji Ikeda's grid-matrix aesthetic.

    A high-contrast binary cell grid (default 32×32) plus a thin
    "frame" of fader bars around the edges. Per step, the grid
    evolves via cellular-automaton-like state transitions with
    random glitch inversions, then the active cells are binarised
    (white-on-black) and the canvas is washed with the glitch
    white-noise floor that Ikeda uses in *datamatics* and
    *data.matrix*. The faint per-row numerical readout (binary
    strings) is a direct nod to the *matrix [for rooms]* CD
    track listing (0000000001, 0000000010, ...) — Ikeda's piece
    is literally a binary matrix.
    """

    family: ClassVar[str] = "audio_reactive"
    sketch: ClassVar[str] = "ikedagridmatrix"

    def __init__(
        self,
        grid_size: int = 32,
        cell_px: int = 14,
        glitch_rate: float = 0.04,
        invert_rate: float = 0.02,
        fader_bars: int = 8,
        fader_height: int = 10,
        noise_floor: float = 0.04,
        beat_freq: tuple[float, ...] = (0.7, 1.4, 2.1, 3.0),
        beat_amp: tuple[float, ...] = (0.20, 0.18, 0.14, 0.10),
        fg_color: tuple[int, int, int] = (240, 240, 240),
        bg_color: tuple[int, int, int] = (0, 0, 0),
        seed: int | None = None,
    ) -> None:
        self.grid_size = grid_size
        self.cell_px = cell_px
        self.glitch_rate = glitch_rate
        self.invert_rate = invert_rate
        self.fader_bars = fader_bars
        self.fader_height = fader_height
        self.noise_floor = noise_floor
        self.beat_freq = np.array(beat_freq, dtype=np.float32)
        self.beat_amp = np.array(beat_amp, dtype=np.float32)
        self.fg_color = np.array(fg_color, dtype=np.float32)
        self.bg_color = np.array(bg_color, dtype=np.float32)
        self._seed = seed
        self._t = 0.0
        self._tick = 0
        self._last_beat = 0.0
        self._phase = np.array([], dtype=np.float32)
        self.reset()

    def reset(self, seed: int | None = None) -> None:
        if seed is not None: