- Map A=Train, B=Clear Examples, X=Randomize, LB=Thumbs Down, RB=Thumbs Up
- Add periodic gamepad polling fallback for environments where
gamepadconnected event doesn't fire reliably (Steam Deck, some Linux browsers)
- Show "Press any gamepad button to connect" hint on page load
When synth controls (volume, tempo, etc.) have focus after dragging,
keyboard shortcuts 1/2 for thumbs down/up were intercepted by the
range input. Now all four UIs skip shortcut handling when an input,
select, or textarea element has focus.
Relocated the output mode toggle from a floating top-right position into
the bottom floating bar where it's more discoverable. Removed the
duplicate toggle from the expanded sheet. Compact pill-toggle-sm style
fits the toolbar layout.
Merge Examples/RL tabs into single unified toolbar across all four UIs.
Users can now freely mix supervised learning (Add Example + Train) with
RL feedback (thumbs up/down) without switching modes. Param bar dragging
is always enabled.
Add Web MIDI input module (js/synth/midi-input.js) for external MIDI
controllers — routes note on/off to C15 synth, CC 1/2 to joystick.
Add gamepad module (js/ui/gamepad.js) with auto-detection, deadzone,
and axis normalization across all UIs.
Fix synth parameter sync: routeOutputs() now called after trainModel()
in onThumbsUp and loadState to prevent stale params on first joystick
move. Add separate Clear Examples button (dataset only, keeps weights).
- Fix RL buttons stuck high on desktop: media query was overriding
bottom to 196px, now matches mobile 92px
- Add floating Visual/Synth toggle at top-right of immersive view,
synced with the existing sheet toggle
Add ?spread=0-1 URL param that controls weight initialization scaling,
RL noise scaling per layer, noise cap, and weight decay to prevent
sigmoid output saturation. At spread=0 (original behavior) weights are
uniform [-1,1] and outputs polarise near 0/1. At spread=1 weights use
Xavier scaling (1/sqrt(fan_in)), noise is proportionally reduced, and
10% weight decay per thumbs-down prevents unbounded magnitude drift.
Also fix randomise to re-inject current joystick position and re-run
inference before routing outputs, eliminating the jump on first
joystick move after randomise.
Defaults: tame=1, spread=0.6 across all app variants.
When switching from synth back to visual mode, the FlowFieldVisualizer
canvas had stale dimensions from being hidden. Add requestAnimationFrame
resize call after toggling canvas visibility so the visualizer picks up
correct dimensions and renders properly.
Three redesigns of the playground UI, each exploring a different
interaction philosophy for the NISPS interactive ML engine:
A) Immersive — fullscreen canvas with floating glassmorphism overlays
B) Workbench — dashboard with 2D mapping heatmap as centerpiece
C) Journey — phased UI that dissolves (Explore → Teach → Perform)
All share the same 126-output MLP ([32,48,64] hidden layers), C15
synth bridge, and flow field visualizer. Each adds:
- SynthVisualizer (126 grouped vertical bars) replacing particles in synth mode
- RL as default learning mode with keyboard shortcuts (1/2)
- Follow mode for trackpad exploration
- ?tame= URL param for safe-range constraining
- localStorage persistence (random on first boot)
- Live iframe previews on designs.html homepage
See devlog/2026-03-21-ui-redesign-explorations.md and meml-2c3 for
remaining refinements identified during browser testing.
- Insert DynamicsCompressorNode as always-on brick-wall limiter after
master gain (threshold -6dB, ratio 20:1, 2ms attack)
- Add devmode tame parameter (0-1) controlling range constraining:
- ?tame=0.7 URL param (default) or window.setTameLevel(n)
- 0 = no mitigation, 1 = strongest constraining
- Constrain 20 volume/buildup-sensitive params with safeMin/safeMax:
- Envelope release/decay2 times capped to prevent infinite ringing
- Envelope gains kept in moderate range
- All 5 drive stages (ShpA/B, FB, Out, Cabinet) capped
- Output mixer levels given floor and ceiling
- Echo feedback and reverb size limited to prevent wash buildup
- Fix Comb Filter group count (8, not 9) in comments and color generator
Control all sonically meaningful C15 parameters through NISPS ML outputs.
Excludes hardware routing, macros, scale/tuning, key tracking, velocity,
envelope mod depths, discrete switches, and dangerous volume/pitch params.
- Expand param-map.js with 126 curated params grouped by synthesis section
- Widen MLP architecture from [3,10,10,14,20] to [3,32,48,64,126]
- Add ParamDisplay.rebuild() for mode-dependent bar count (20 visual, 126 synth)
- Add scrollable compact layout for synth mode param bars
- Pad visual presets and old saves to 126 outputs for backward compatibility
Adds a secondary Synth mode (toggled via collapsible left side panel)
where the NISPS ML engine's 20 outputs control curated C15 synthesizer
parameters (oscillator PM, shapers, filters, reverb, echo, flanger,
output mixer) through a WASM AudioWorklet bridge.
Includes a chord progression arpeggiator with controls for tempo (BPM),
octave range, octave offset, and 4 selectable progressions. The C15
engine runs in an AudioWorklet with lock-free SharedArrayBuffer ring
buffer communication.
New files: c15/ (WASM assets), js/synth/ (bridge, arpeggiator, param map),
serve-coop.py (COOP/COEP headers for SharedArrayBuffer support).
Faithful JS port of nisps-core MLP + IML engine with a touch-friendly
UI for exploring neural parameter mapping visually. Two learning modes:
example-based (set slider targets) and RL feedback (thumbs up/down with
exploration noise). Flow field particle system controlled by 8 MLP outputs.
- Remove platform-specific code (ARM_MATH_CM33, XMOS __XS3A__, std::printf)
- Add set_output()/set_outputs()/add_example() API for programmatic training
- Fix release build crash: side effect inside assert() for loss function init
- Replace fake smoke test with real convergence tests (5 tests, all pass)
- Rewrite example to demonstrate actual training with real output
- Update README, CHANGELOG, and extraction plan to match reality
Extract the interactive machine learning engine from MEMLNaut-NISPS
firmware into a standalone, platform-agnostic C++20 header-only library.
What is nisps-core?
-------------------
NISPS (Neural Interactive Shaping of Parameter Spaces) core is a
parameter mapping engine. It takes N input parameters (joystick,
sensors, audio features) and maps them to M output parameters through
an interactively-trained neural network.
Use it to control: synthesizers, effects, lights, robots, game
parameters, or anything that responds to continuous control data.
Key Features
------------
- Header-only: No compilation needed, just include and use
- Platform-agnostic: Pure C++20, works anywhere
- Zero dependencies: Only standard library
- Interactive learning: Train by demonstration
- Lightweight: ~3,500 lines of optimized neural network code
- Flexible: Map 1-100 inputs to 1-100 outputs
Architecture
------------
Core components:
- IML: High-level interactive ML interface
- MLP: Multi-layer perceptron (feedforward neural network)
- Dataset: Training data management with replay memory
- Layer/Node: Neural network building blocks
- Loss: MSE and categorical cross-entropy functions
- Utils: Activation functions (sigmoid, ReLU, tanh, etc.)
Transformations Applied
-----------------------
✅ Removed Arduino/RP2040 dependencies (Serial, SD, Pico SDK)
✅ Removed audio synthesis code (nisps-core is control-only)
✅ Added nisps namespace to all code
✅ Converted to header-only library with _impl.hpp pattern
✅ Updated to C++20 (required for std::span)
✅ Removed platform-specific serialization
✅ Replaced debug macros with no-op stubs
✅ Added comprehensive documentation and examples
Files Added
-----------
- nisps-core/README.md: Complete documentation and API reference
- nisps-core/CHANGELOG.md: Version history and migration guide
- nisps-core/include/nisps/*.hpp: 13 header files (~3,500 lines)
- nisps-core/test/main.cpp: XOR test demonstrating basic usage
- nisps-core/examples/simple_mapping.cpp: Interactive demo
- nisps-core/CMakeLists.txt: Build system for tests
Testing
-------
✅ Compiles with GCC 14.2 (C++20)
✅ All tests passing
✅ Successfully instantiates networks and runs inference
Performance
-----------
- Inference: 1-10 µs for small networks (2-10-10-4)
- Training: 10-100 ms for 100 examples, 1000 iterations
- Memory: ~1 KB per hidden neuron
Migration from Embedded IMLInterface
------------------------------------
Old (embedded):
IMLInterface iml(n_inputs, n_outputs);
New (nisps-core):
nisps::IML<float> iml(n_inputs, n_outputs);
All method names remain the same, just add the namespace.
Related
-------
- Implements: NISPS_CORE_EXTRACTION_PLAN.md
- Task graph: NISPS_CORE_TASKS.md
- Origin: MEMLNaut-NISPS firmware
- Docs: https://musicallyembodiedml.github.io/memlnaut/
Co-authored-by: Claude Code <claude@anthropic.com>
