In this repository, I'll be revisiting and restoring patches that I've been programming since 2003. Some of them were originally created in Pure Data and are now being rebuilt in Max to keep them functional and accessible.

These are primarily tools for sound generation, visuals, and control systems that have accompanied my productions over the years. Together, they form a personal archive of instruments, experiments, and working methods that have shaped projects across different periods of my practice.

I'll be gradually making them available here, preserving not only the tools themselves but also some of the ideas, processes, and obsessions that led to their creation.

A generative noise instrument built in Max/MSP. Four rhythmically triggered voices and a continuous drone feed into a shared effects chain and stereo panning system, producing evolving textural soundscapes.

Metro --> Euclidean Sequencer (4 channels)
            |        |        |        |
          Voice1   Voice2   Voice3   Voice4
            |        |        |        |
            +--------+--------+--------+--> Stereo Panning --> dac~
            |        |        |        |
            +--[FX gates]--+-----------+--> Effects Chain --^
                                                |
          Drone --------------------------------+--> Panning

The master metro drives a 4-channel Euclidean sequencer. Each channel fires rhythmic triggers into its voice. Every voice produces a short, enveloped burst with randomized parameters on each trigger. The drone runs continuously underneath. All audio passes through a stereo panning module with per-voice random positioning, and optionally through a shared effects chain controlled by per-voice FX send toggles.

A 4-channel Bjorklund/Euclidean rhythm generator. A single counter drives all channels. Each channel has independent pulses and steps parameters that determine its rhythmic pattern through a * -> % -> < -> sel 1 chain. Outputs a bang per channel when the pattern activates.

Randomized cycle~ and rect~ oscillators through an svf~ bandpass filter. Each trigger randomizes frequency, modulation index, filter cutoff, and envelope shape (attack/decay via line~ 0 0). Produces sharp, clicky tonal bursts inspired by Ryoji Ikeda's micro-sound aesthetics.

noise~ sampled by sah~ (sample-and-hold) with a click~ trigger, creating stepped random waveforms. Filtered through svf~ with randomized cutoff and envelope. The rect~ frequency is randomized per trigger for varying timbral density. Produces metallic, ring-modulated textures.

saw~ through degrade~ (bit-crushing and sample-rate reduction) into svf~ filtering. Randomized frequency, filter cutoff, and envelope per trigger. Produces gritty, lo-fi digital artifacts with a raw, aggressive character.

Four switchable noise types routed through selector~ 4:

Selected noise passes through svf~ bandpass with randomized cutoff per trigger, shaped by the same attack/decay envelope as voices 1-3.

Continuous tone generator using detuned saw~ oscillators (32 Hz and 32.1 Hz) with slow cycle~ LFO modulation, summed and filtered through svf~ lowpass at 120 Hz. Produces a deep, slowly evolving bass foundation.

Slow random parameter modulation for the drone. Two independent metro clocks (8s and 3s) drive random values that periodically shift the drone's pitch and filter cutoff, keeping the drone texture alive over long durations.

Shared effects processor fed by per-voice FX send gates (*~ controlled by toggles). Signal path: tapin~/tapout~ delay into comb~ resonator with feedback (*~ 0.6 loop), through tanh~ soft saturation and clip~ limiting, scaled by *~ 0.25 for return level. Creates dense, reverberant feedback textures.

4-channel stereo panning with per-voice random pan positioning. On each Euclidean trigger, a new random pan position is generated, scaled by a global Spread parameter, and smoothed with line~. Equal-power panning coefficients (+~ 1 -> *~ 0.5 for R, *~ -1 -> +~ 1 -> *~ 0.5 for L) multiply each voice's audio. All channels sum through a cascade of +~ bus objects into clip~ -1 1 limiting before dac~. The drone enters at center (mono *~ 0.5 to both channels). The effects return is scaled by times~ 0.4 before the final mix.

• Max 9 (uses pink~, degrade~, svf~)