Chaotic AI Drum Brain • Touch • Feedback • Spectral Morphing

A Max for Live hybrid drum machine/sampler/AI-inspired sound engine combining:

• SOMA Enner (touch plates, chaos, feedback routing)
• Elektron Monomachine (trigs, parameter locks, FX lane)
• Elektron Digitakt II (sample workflow)
• SOMA Pulsar-23 (burst generators, pulses)
• AI/Spectral Processing (RAVE-inspired granular/spectral morphing)

Each track features:

• Dual voice architecture: Sample playback (groove~) + Synthetic voice (FM/noise/cycle~)
• Independent envelope (ADSR~)
• SVF filter (lowpass with cutoff/resonance)
• Parameter lockable controls: Pitch, Level, Decay, Filter, FX Sends
• Buffer~ sample loading (10 seconds per track)

• Step matrix (4 tracks × 16 steps)
• Parameter lock dictionary per step
• Trig types: Normal, Ghost, Trigless Lock, Retrig
• Conditional trigs: Probability, Fill, 1:2, 1:4, 2:4
• Microtiming nudge
• Velocity per step

8 interactive touch plates (live.slider objects):

• P1-P8: Modulation sources for chaos/routing/parameters
• Touch/drag to generate control signals
• Can modulate:
  • Track parameters
  • Feedback depth
  • Chaos amount
  • Spectral processing
  • FX routing

Enner-inspired feedback loop with:

• Dual delay lines (2000ms/2300ms L/R)
• Ring modulation (cycle~ @ 37Hz)
• Chaos injection (noise~ modulated by touch plates)
• SVF lowpass filter (800Hz, Q=0.9)
• Tanh~ saturation (analog-style drive)
• Feedback amount dial (0-0.95)
• Chaos amount dial (0-1.0)

Simulated RAVE-style processing:

• pfft~ spectral blur (2048 FFT size, 4× overlap)
• Granular delay (441/661 samples L/R for grain motion)
• Encoder Depth control (dry/wet mix)
• Spectral Morph (formant warping simulation)
• Can process all tracks or per-track sends

Per-step automation lane:

• FX A Send (delay/reverb)
• FX B Send (spectral/granular)
• Feedback Send (chaos injection)
• Chaos Burst (trigger instability)
• Stored in dict fx_lane_data

• pattrstorage object (syndrava)
• Saves with Ableton set:
  • Sequencer patterns
  • Parameter locks
  • Touch plate states
  • Sample references
  • Routing states

cp SYNDRAVA.amxd ~/Music/Ableton/User\ Library/Presets/Instruments/Max\ Instrument/

Or place in any Ableton User Library location.

1. Open Ableton Live 11+ (Suite or Standard with Max for Live)
2. Navigate to Browser → Max for Live → Max Instrument
3. Drag SYNDRAVA.amxd onto a MIDI track
4. Wait for Max to initialize (check bottom left corner)

• Click the Max window icon in SYNDRAVA
• Audio should be ON (green audio status)

• Drag audio files onto the device
• Samples load into buffer~ track1buf, track2buf, etc.
• Each buffer holds up to 10 seconds (44.1kHz)

• Click cells in the 16-step matrix
• Row = Track (1-4)
• Column = Step (1-16)
• Lit cells = active triggers

• P1-P8 sliders on the right side
• Drag to modulate:
  • Feedback depth
  • Chaos amount
  • Filter cutoff
  • Spectral morph

• Feedback dial: Amount of signal fed back (0-95%)
• Chaos dial: Random modulation depth (0-100%)
• Touch plates interact with feedback network

• AI Depth dial: Dry/wet spectral processing
• Morph dial: Spectral character/formant shift
• Creates RAVE-like textural transformations

• FX lane stores per-step routing changes
• Modulate sends to feedback/encoder busses
• Trigger chaos bursts on specific steps

Each step can lock parameters:

1. Select a step in the sequencer
2. Adjust track parameters (pitch, level, decay, filter)
3. Values stored in dict pattern_locks
4. Per-step override of global parameters

Normal: Standard trigger Ghost: Quiet trigger (50% velocity) Trigless Lock: Parameter change without trigger Retrig: Rapid repeating trigger (rate/count lockable)

Probability: 0-100% chance per step Fill: Only fires when Fill button is active 1:2: Every 2nd cycle 1:4: Every 4th cycle 2:4: 2 out of 4 cycles

Touch plates → Feedback Bus → Chaos Injection → Ring Mod → Filter → Saturation → Output

The feedback can:

• Self-oscillate (careful with feedback dial!)
• Generate rhythmic artifacts
• Create harmonic distortion
• Produce semi-random modulation

Encoder Bus signal flow:

1. Input → pfft~ spectral blur (frequency smearing)
2. Granular delay (time-domain microshifts)
3. Morph control (spectral tilt/formant shift)
4. Encoder Depth (dry/wet crossfade)

Simulates AI latent space interpolation through spectral manipulation.

Top-level signal flow and UI presentation

• Step matrix (matrixctrl 4×16)
• Parameter lock dictionary
• Trig condition logic

Individual track voice engines:

• buffer~ sample storage
• groove~ sample playback
• cycle~ synthetic oscillator
• adsr~ envelope
• svf~ filter
• *~ VCA/mixer

Inlets:

1. Trigger (bang)
2. Pitch (0-127 MIDI / float multiplier)
3. Level (0-1)
4. Decay (ms)
5. Sample (buffer name)

Outlets:

1. Audio L
2. Audio R

8 live.slider touch plates (P1-P8)

Outlets: 8 float streams (0-1)

Chaos feedback network

Inlets:

1. Audio L
2. Audio R

Outlets:

1. Processed L
2. Processed R

Controls:

• Feedback amount (0-0.95)
• Chaos amount (0-1.0)

AI-inspired spectral processor

Inlets:

1. Audio L
2. Audio R

Outlets:

1. Spectral L
2. Spectral R

Controls:

• Encoder Depth (dry/wet)
• Spectral Morph (formant shift)

Per-step routing automation

Inlet: Step number (0-15)

Outlets:

1. FX A Send (0-1)
2. FX B Send (0-1)
3. Feedback Send (0-1)
4. Chaos Burst (0-1)

Final stereo bus mixer (8 inputs → 2 outputs)

┌─────────────────────────────────────────────────────────────────┐
│                    ⬡ S Y N D R A V A ⬡                          │
│      Chaotic AI Drum Brain • Touch • Feedback • Spectral       │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  SEQUENCER (16 steps × 4 tracks)                               │
│  ┌──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┬──┐           │
│  │01│02│03│04│05│06│07│08│09│10│11│12│13│14│15│16│  Track 1  │
│  ├──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┤           │
│  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  Track 2  │
│  ├──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┤           │
│  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  Track 3  │
│  ├──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┼──┤           │
│  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  │  Track 4  │
│  └──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┴──┘           │
│                                                                 │
├─────────────────────────────────────────────────────────────────┤
│  ENNER SURFACE                    CHAOS & AI CONTROLS          │
│  ┌──┬──┬──┬──┬──┬──┬──┬──┐       ┌────────┬────────┬────────┐│
│  │P1│P2│P3│P4│P5│P6│P7│P8│       │Feedback│ Chaos  │AI Depth││
│  │  │  │  │  │  │  │  │  │       │  Amt   │  Amt   │ / Morph││
│  │  │  │  │  │  │  │  │  │       └────────┴────────┴────────┘│
│  └──┴──┴──┴──┴──┴──┴──┴──┘                                    │
└─────────────────────────────────────────────────────────────────┘

Sample Rate: 44.1kHz / 48kHz (follows Ableton project) Bit Depth: 32-bit float (Max/MSP internal) FFT Size: 2048 (encoder bus) FFT Overlap: 4× Buffer Size: 10 seconds @ 44.1kHz per track Polyphony: 4 voices (1 per track) Latency: 5ms (pfft processing + buffer latency) CPU Usage: Moderate (depends on spectral processing depth)

SYNDRAVA is not a traditional drum machine. It is a ritualistic sound instrument where:

1. Chaos is intentional — Feedback and touch plates introduce controlled randomness
2. Routing is performance — FX lane changes the signal path per step
3. AI is simulated — Spectral processing mimics neural encoding without ML inference
4. Touch is expressive — Enner-style plates make static sequences dynamic
5. Sequencer is alive — Parameter locks and conditionals create evolving patterns

1. Program basic 16-step pattern
2. Add parameter locks to steps 4, 8, 12 (pitch/filter changes)
3. Set FX lane to boost encoder depth on steps 5, 13
4. Modulate feedback with touch plates in real-time
5. Add chaos bursts on steps 7, 15 for rhythmic artifacts

1. Load noisy/percussive samples
2. Set high feedback (0.7-0.9)
3. Increase chaos amount (0.5+)
4. Modulate touch plates rapidly
5. Use encoder bus morph to smear frequencies

1. Use retrig trig type on steps
2. Set high retrig rate (8-16 repeats)
3. Short decay envelopes (<50ms)
4. Modulate pitch per retrig with parameter locks
5. Add chaos burst in FX lane

1. Drag drum hits onto tracks
2. Use groove~ start/length for slicing
3. Pitch parameter locks for melodic sequences
4. Encoder bus for spectral mangling
5. Feedback bus for rhythmic echoes

This is a Minimum Viable Product demonstrating core concepts. Future enhancements:

• Full 8-track expansion
• Visual step editor with parameter lock UI
• More trig conditions (euclidean, ratcheting)
• Per-track FX sends (currently global)
• MIDI CC mapping for hardware controllers
• Preset management system
• Advanced spectral effects (freeze, bin scramble)
• Polybuffer~ for multi-sample layers
• Sample slice auto-detection
• Live recording into buffers

1. Open SYNDRAVA.amxd in Max for Live
2. Navigate to p encoder_bus
3. Replace p_spectral_blur pfft~ patch with custom FFT processing
4. Example processors:
   • fft~ + cartopol~ + frequency bin scrambling
   • spectdelay~ for spectral freezing
   • frameaccum~ for smearing

1. Open p track1 (or any track)
2. Replace cycle~ with:
   • saw~ (aliased saw)
   • rect~ (pulse/PWM)
   • pink~ (noise generator)
   • rand~ (sample & hold chaos)
   • Custom poly~ voice

1. Insert between p mixer and p feedback_bus
2. Recommended Max effects:
   • tapin~/tapout~ (delay)
   • comb~ (resonator)
   • freqshift~ (frequency shifter)
   • teeth~ (bit crusher)

1. Click floppy disk icon (pattrstorage UI)
2. Type preset name
3. Stores:
   • Sequencer pattern
   • Parameter locks
   • Touch plate states
   • Routing settings

1. Click preset number in pattrstorage
2. All parameters restore instantly

• Use dict pattern_locks object
• Export to JSON via dict inspector
• Import to other SYNDRAVA instances

Steps: 1,5,9,13 (kick)
       3,7,11,15 (snare)
       Every step (hihat with 50% probability)

P-Locks: Step 5 (pitch +7 semitones)
         Step 13 (pitch +12 semitones)

FX Lane: Step 8 (feedback burst)
         Step 16 (encoder depth 100%)

Steps: 1 (long sample, decay 2000ms)
       9 (retrig ×8)

P-Locks: Step 1 (filter 200Hz)
         Step 9 (filter 5000Hz, pitch -12)

Touch: P1→Feedback, P3→Chaos, P5→Morph
FX Lane: All steps (encoder 80%, feedback 60%)

Steps: Random pattern with 70% probability per step

P-Locks: Every step (random pitch -24 to +24 semitones)
         Alternating filter cutoff (100Hz / 8000Hz)

Touch: Rapid P2/P4/P6 modulation
FX Lane: Chaos bursts every 4 steps

Max for Live Documentation: https://docs.cycling74.com/max8/vignettes/m4l_intro Buffer~ Reference: https://docs.cycling74.com/max8/refpages/buffer~ pfft~ Spectral Processing: https://docs.cycling74.com/max8/vignettes/fftspectral pattrstorage Presets: https://docs.cycling74.com/max8/vignettes/presets

Concept: Hybrid of SOMA Enner, Elektron Monomachine, Digitakt II, Pulsar-23 Architecture: Semi-modular chaotic drum brain with AI-inspired processing Built with: Max/MSP 9, Max for Live Designed for: Ableton Live 11+ Suite

This Max for Live device is provided as-is for educational and creative use.

Feel free to:

• Modify and extend
• Use in commercial productions
• Share with others (with attribution)

Do not:

• Resell as a standalone product
• Claim original authorship

• Install SYNDRAVA.amxd in Ableton User Library
• Load on MIDI track in Ableton Live
• Verify Max DSP is running (bottom left)
• Load 4 drum samples (kick, snare, hihat, perc)
• Click steps in sequencer matrix
• Press Play in Ableton
• Adjust Feedback/Chaos dials
• Modulate with touch plates (P1-P8)
• Enable Encoder Depth for spectral morphing
• Save your first preset!

⬡ Welcome to SYNDRAVA — where chaos meets rhythm ⬡