Skip to content

agent_plan.md

Latest commit

 

History

History
166 lines (150 loc) · 28.2 KB

agent_plan.md

File metadata and controls

166 lines (150 loc) · 28.2 KB

Web Sequencer: Living Roadmap

Mission: Build the ultimate browser-based DAW, fusing classic MIDI workflow (Cubase-style) with next-generation AI Vocal Synthesis (TTS Sampling).

Current Velocity: 1 Feature per Run Architectural Focus: Polyphony & Vocal Granularity

🚀 Active Backlog (Prioritized)

Domain A: Audio Engine (Synth & Sampler)

  • Refactor SingingVoice State: Expose alignment state setters in SingingVoice to avoid type casting hacks and improve multi-bank alignment handling.
  • TTS Slice Triggering: Implement a logic where a MIDI Note NoteOn event can trigger a specific slice or word from the TTS buffer (e.g., Note C3 = "Hello", Note D3 = "World").
  • Hybrid Polyphony: Finalize VoiceManager to handle 8-voice polyphony for synth-1 while keeping synth-2 strictly monophonic (legato priority).
  • Phoneme Elasticity: Connect rubberband.wasm to the sequencer steps. If a note is dragged longer, the specific phoneme should time-stretch to match the duration without altering pitch.
  • Polyphonic Singing: Implement SingingVoiceManager to allow polyphonic playback (chords) for the TTS engine, supporting multiple simultaneous phoneme streams.

Domain B: Editor Workflow (The "Cubase" Feel)

  • Per-Step Delay Send: Allow sequence steps to override the global delay send amount.

  • Step-Sequenced Vibrato: Allow individual sequencer steps to override the global Vibrato Depth, enabling natural vocal phrasing without affecting short rhythmic syllables.

  • Vocal Envelope Shaper: Upgraded the ExpressiveVoiceProcessor to use a full ADSR (Attack, Decay, Sustain, Release) envelope, exposing Decay and Sustain knobs to the SamplerPanel to allow granular shaping of TTS syllables into sharp plucks or smooth pads.

  • Dynamic Tremolo (AM) Effect: Expose the Tremolo effect from ExpressiveVoiceProcessor to the UI (Rate and Depth knobs) and map it to SingingVoice and SamplerPanel to make TTS playback more rhythmic.

  • Per-Step Freeze Amount: Implemented freeze parameter in NoteSelector and mapped it through useAudioEngine to SingingVoice to allow for rhythmic, per-step granular stutters and smears.

  • Step-Sequenced Formant Shifts: Allow users to pitch shift the formants of the TTS engine independently of the fundamental frequency per step.

  • Text-to-Drumkit: Implement logic to generate drum kit sequences by mapping short transient consonants to hats/snares and vowels to kicks based on TTS phrases.

  • Dynamic Phoneme Pitch Glide: Implemented pitch slide logic in SingingVoice.ts, useAudioEngine.ts and App.tsx to glide TTS pitch naturally when a step has slide enabled.

  • Slice Mode UI: Add a toggle in SamplerPanel to enable "Phoneme Slice Mode", allowing users to play slices via MIDI keyboard.

  • Visual Slice Feedback: Highlight the active phoneme slice in the UI during playback, visualizing real-time TTS articulation.

  • Rubber-Band Selection: Implement multi-note selection via mouse drag in Sequencer.tsx (and App.tsx main view).

  • Clipboard Operations: Standardize Ctrl+C / Ctrl+V logic to handle both Note Data and associated TTS Metadata (which word is attached to the note).

  • Per-Step Parameters: Create a UI to edit "Expression" or "Timbre" for individual sequencer steps (vital for humanizing TTS output).

  • Melodic Lyric Mode: Decouple "Slice Selection" from "Pitch" in the Sampler. Allow playing melodies (Pitch) while triggering specific phonemes (Slice Index) via LyricMapper.

  • Note Slice UI: Visualize the selected phoneme/slice index directly on the sequencer note (e.g., small text label) to improve TTS workflow.

  • Unify Sequencer Components: Refactor App.tsx to use the standalone Sequencer component or move App.tsx logic into a reusable view to reduce duplication.

  • Glitch Mode: Add a probability knob that randomly retriggers/stutters the start of a TTS sample (granular synthesis) in the UI.

  • Cleanup Legacy Code: Identify and remove unused components like src/components/Sequencer.tsx to streamline maintenance.

  • Refactor NoteSelector Focus Management: Verified NoteSelector robustness with new test suite src/components/__tests__/NoteSelector.test.tsx mocking requestAnimationFrame.

  • Step-Sequenced Drive: Allow individual sequencer steps to override the global Drive/Distortion amount for aggressive, rhythmic vocal accents.

  • Glissando/Portamento Curves: Allow users to draw custom pitch curves or select between Linear and Exponential glide types between steps. (Implemented Exponential Glide in SingingVoice.ts!)

  • Per-Step Breath Intensity: Allow sequence steps to override global breathiness for rhythmic breathing and whisper effects. (Implemented in useAudioEngine.ts!)

  • Custom Sample Slicing UI: Add a waveform view to SamplerPanel that allows users to manually add, move, and remove transient markers for slicing a custom WAV file instead of just auto-slicing by phoneme.

  • Voice Layering / Chorus per Voice: Allow a detune spread parameter per step for instant vocal thickening.

  • Gesture Controls: Implement pinch-to-zoom for the sequencer timeline to handle longer patterns or finer steps.

  • Custom Waveform LFO: Allow users to draw custom LFO shapes for formant and freeze modulation. (Implemented in FormantShifter and SamplerPanel using DrawableLFO and PeriodicWave via DFT!)

  • Gesture Controls: Implement pinch-to-zoom for the sequencer timeline to handle longer patterns or finer steps.

  • Custom Waveform LFO: Allow users to draw custom LFO shapes for formant and freeze modulation.

  • Vocal Formant Envelope: Add a dedicated Attack/Decay envelope specifically for formants to create plucky or sweeping vocal filter effects on every note.

  • Step-Sequenced Reverb Types: Allow sequence steps to override the global reverb type (e.g., switch from a small room to a massive hall on a specific snare or vocal slice).

  • Per-Step Delay Send: Allow sequence steps to override the global delay send amount to create rhythmic echoes for TTS or synth sequences.

Domain C: Accessibility & Mobile

  • Touch Targets: Audit Sequencer.tsx click listeners to ensure mobile drag-to-create works smoothly.
  • A11y Colors: Verify high-contrast separation between synth-1 (Chords) and synth-2 (Lead) notes.
  • Formant Automation: Implemented continuous vowel morphing by drawing curves for formant shift over time, interpolating parameters globally.
  • [2026-04-25] - Implemented Step-Sequenced Reverb Types: Added reverbType parameter to Note interface and updated NoteSelector UI to include a space dropdown (Room, Plate, Hall). Refactored useAudioEngine.ts to instantiate all three convolution spaces simultaneously to prevent pop artifacts on hot-swapping and updated audioPlayback.ts routing to send signals to the correct active reverbNodesRef based on the sequence step.

🧠 Innovation Lab (The "Dream" Log)

  • Idea: "Spectral Granulator" - Add a granular synthesis mode to the sampler that uses FFT to freeze and smear TTS phonemes over time. (Implemented in Sampler and RubberBandProcessor via a 100ms looping Hann window!)

  • Idea: "Chord Evolving" - Allow drawing automation curves for the chord inversions or voicings used by VoiceManager in Polyphonic Synth A. (Implemented via automation track logic in useStepHandler.ts and App.tsx!)

  • Idea: "Step-Sequenced Formant Shifts" - Allow users to pitch shift the formants of the TTS engine independently of the fundamental frequency per step. These are concepts to be fleshed out by the agent during "Architect Mode".

  • Idea: "Voice Layering / Chorus per Voice" - Allow a detune spread parameter per step for instant vocal thickening. (Implemented per-step choir in NoteSelector/useAudioEngine!)

  • Idea: "Lyric Track" - A global text input that automatically distributes syllables across selected MIDI notes. (Implemented via global Lyric Track lane and sliceIndex auto-mapping!)

  • Idea: "Choir Stack" - Using Polyphony to detune the TTS voice slightly on 3 channels to create a chorus effect. (Implemented via Polyphonic Singing update!)

  • Idea: "Gesture Controls" - Implement pinch-to-zoom for the sequencer timeline to handle longer patterns or finer steps.

  • Idea: "Per-Step Reverse" - Allow reversing the TTS sample on a per-step basis for creative rhythmic effects. (Implemented in Sampler phoneme slicing!)

  • Idea: "Dynamic Phoneme Pitch Bends" - Allow drawing pitch bend automation within a single phoneme slice (e.g. going up a fifth on a single vowel). (Implemented as Slide/Glide property!)

  • Idea: "Text-to-Drumkit" - Auto-generate a drum kit from a TTS phrase by mapping short transient consonants (t, k, p) to hats/snares and vowels to kicks/toms.

  • Idea: "Vocal Envelope Shaper" - Add granular attack/decay ADSR shaping explicitly for TTS syllables to create sharp plucks or smooth pads from any word. (Implemented in ExpressiveVoiceProcessor and exposed to SamplerPanel!)

  • Idea: "Dynamic Tremolo" - Expose Tremolo Rate and Depth to UI to pulse vocals. (Implemented!)

  • Idea: "LFO to Freeze Amount" - Automate the Freeze parameter with an LFO to create rhythmic pulsing granular clouds. (Implemented in RubberBandProcessor and exposed to SamplerPanel!)

  • Idea: "Per-Step Filter & Resonance" - Allow sequence steps to override cutoff and resonance for rhythmic acid-style filtering of TTS samples. (Implemented!)

  • Idea: "Filter Envelope Mod" - Allow the sequence steps to have an envelope mod amount that specifically shapes the filter envelope per step. (Implemented!)

  • Idea: "Vocal Formant LFO" - Introduce a Formant LFO with rate and depth controls for dynamic rhythmic Wah-Wah effects on TTS vowels. (Implemented in FormantShifter and SamplerPanel!)

  • Idea: "Step-Sequenced Formant LFO" - Allow individual steps to override the global Formant LFO rate and depth for highly articulated rhythmic sequences.

  • Idea: "Custom Waveform LFO" - Allow users to draw custom LFO shapes for formant and freeze modulation. (Implemented in FormantShifter and SamplerPanel using DrawableLFO and PeriodicWave via DFT!)

  • Idea: "LFO Rate Sync to Tempo" - Allow LFO rates (like Formant LFO or Freeze LFO) to sync to the sequencer tempo. (Implemented in NoteSelector, SamplerPanel, and useAudioEngine!)

  • Idea: "Microtonal Scale Mapping" - Allow specifying custom microtonal scales rather than the standard 12-TET for Synth and Sampler parts.

  • Idea: "Custom Waveform LFO" - Allow users to draw custom LFO shapes for formant and freeze modulation. (Implemented via DrawableLFO and FormantShifter Fourier Transform!)

  • Idea: "Glissando/Portamento Curve Drawing" - Allow users to draw custom pitch curves between steps, rather than just a linear glide. (Implemented!)

  • Idea: "Phoneme-Aware Velocity" - Automatically adjust the amplitude envelope attack/decay based on the phoneme type (e.g., plosives get faster attack, vowels get smoother attack). (Implemented via dynamic envelope overrides in SingingVoice.ts!)

  • Idea: "Spectral Morphing" - Implement functionality to morph spectrally between two different TTS phonemes or samples over a sequence of steps. (Implemented via FormantShifter Voice Character Morphing and step-sequenced automation!)

  • Idea: "Dynamic Reverb" - Allow users to draw automation curves for reverb send per step. (Implemented!)

  • Idea: "Global Saturation / Tape Warmth" - Add a master channel saturation unit to glue the mix together. (Implemented via WaveShaperNode!)

  • Idea: "Auto-Slice by Transients" - Use energy-based analysis to automatically detect and place slice markers at drum hits or clear transients when a custom sample is loaded. (Implemented via AUTO-SLICE button logic in WaveformDisplay and SamplerPanel!)

  • Idea: "Per-Step Delay Send" - Allow sequence steps to override the global delay send amount to create rhythmic echoes for TTS or synth sequences. (Implemented!)

  • Idea: "Vocal Formant Envelope" - Add a dedicated Attack/Decay envelope specifically for formants to create plucky or sweeping vocal filter effects on every note. (Implemented in FormantShifter and SamplerPanel!)

  • Idea: "Step-Sequenced Reverb Types" - Allow sequence steps to override the global reverb type (e.g., switch from a small room to a massive hall on a specific snare or vocal slice). (Implemented via three separate global ConvolverNodes inside useAudioEngine allowing glitchless dynamic per-step routing!)

  • Idea: "AI Auto-Mix Assistant" - Automatically adjusts levels, panning, and EQ based on track content to maintain a balanced mix.

  • Idea: "Real-time Convolution Reverb for Vocal Spaces" - Enhance the dynamic reverb by allowing users to select impulse response types. (Implemented!)

  • Idea: "Per-Step Breath Intensity" - Allow sequence steps to override global breathiness for rhythmic breathing and whisper effects. (Implemented!)

  • Idea: "Advanced Slice Tuning" - Allow users to fine-tune slice sensitivity and adjust the threshold parameters for Auto-Slice. (Implemented in SamplerPanel and WaveformDisplay!)

  • Idea: "Per-Step Delay Send" - Allow sequence steps to override the global delay send amount. (Implemented!)

  • Idea: "Master Bus Compressor" - Add a DynamicsCompressorNode to the master output for gluing the mix together. (Implemented via a fixed "Glue Compressor" block before the Master Gain!)

  • Idea: "Sidechain Compression" - Allow routing the kick drum to a dedicated sidechain bus to dynamically duck the bass/synths. (Implemented via scheduled GainNode automation on kick hits!)

  • Idea: "Granular Envelope Follower" - Allow mapping the amplitude envelope of the voice sample to control granular parameters like grain size or freeze amount.

📜 Changelog

  • [2026-06-29] - Implemented Sidechain Compression: Added sidechainBus between the master saturation and master compressor to duck Synths/Basses when the kick drum triggers. Implemented triggerSidechainDuck in audioPlayback.ts to use deterministic native browser parameter scheduling (linearRampToValueAtTime / exponentialRampToValueAtTime) on the new sidechain gain node to avoid CPU overhead and achieve a classic EDM pump. Fulfills the "Sidechain Compression" Innovation Lab idea.
  • [2026-06-28] - Implemented LFO Rate Sync to Tempo: Added toggles and subdivision dropdowns to SamplerPanel and NoteSelector for freezeLfoRate and formantLfoRate. Updated useAudioEngine.ts to calculate real-time Hz values from BPM and note subdivisions when sync mode is enabled. Fulfills the "LFO Rate Sync to Tempo" Innovation Lab idea. Added new idea: "Granular Envelope Follower".
  • [2026-06-27] - Implemented Master Bus Compressor: Inserted a default DynamicsCompressorNode configured as a gentle "Glue Compressor" (Threshold: -15dB, Ratio: 4:1, Attack: 30ms, Release: 250ms) into the master audio chain (masterSaturation -> masterCompressor -> masterGain -> masterPanner). Updated useAudioEngine.ts to reflect the new master bus input point and updated PlaybackRefs.
  • [Date] - Implemented Custom Waveform LFO: Added DrawableLFO component to allow users to draw custom LFO shapes for formant modulation. Integrated with FormantShifter.ts to convert the drawn array into a PeriodicWave using a Discrete Fourier Transform (DFT), passing formantLfoShape from the SamplerPanel through SingingVoice and useAudioEngine. Added new ideas: "LFO Rate Sync to Tempo" and "Microtonal Scale Mapping".
  • [2026-06-27] - Implemented Vocal Formant Envelope: Added Formant Envelope parameters to SamplerBankParams and Note interfaces, allowing users to modulate formant detune amounts via an attack/decay envelope on ConstantSourceNode and GainNode. Added global and per-step controls to SamplerPanel and NoteSelector.
  • [2026-06-27] - Verified Gesture Controls: Confirmed pinch-to-zoom is fully implemented in MainSequencer.tsx and marked as complete.
  • [2026-06-27] - Implemented Per-Step Delay Send: Added delaySend parameter to Note interface and updated NoteSelector UI to include a slider. Plumbed the parameter through App.tsx and updated useAudioEngine.ts to allow both Sampler (playSamplerVoice) and Synth (VoiceManager) engines to override the global delay send bus per-step. Fulfills the "Per-Step Delay Send" Innovation Lab idea.
  • [2026-06-26] - Implemented Custom Waveform LFO: Added a new DrawableLFO canvas component to SamplerPanel allowing users to draw an arbitrary LFO shape. Added a Discrete Fourier Transform (DFT) engine inside FormantShifter.ts to convert the normalized shape into a PeriodicWave used by the internal Web Audio OscillatorNode.
  • [2026-06-25] - Implemented Step-Sequenced Formant Shifts: Added formantShift parameter to Note interface and NoteSelector UI. Wired it through useAudioEngine and useStepHandler to allow per-step overrides of the global Formant Shift parameter, fulfilling the "Step-Sequenced Formant Shifts" Innovation Lab idea.
  • [2026-06-24] - Implemented Real-time Convolution Reverb for Vocal Spaces: Added ReverbType ('room', 'plate', 'hall') to types.ts and updated useAudioEngine.ts to dynamically generate and swap the convolution impulse response based on the selected type using createReverbImpulseResponse. Added a dropdown selector to App.tsx next to the Master Saturation control to allow global reverb type switching.
  • [2026-06-23] - Implemented Advanced Slice Tuning: Added an auto-slice sensitivity slider to the WaveformDisplay component. Plumbed the sensitivity state through SamplerPanel down to PhonemeAligner.detectSegmentBoundaries via a new thresholdMultiplier parameter, allowing users to fine-tune transient detection for different sample types.
  • [2026-06-22] - Implemented Auto-Slice by Transients: Added an "AUTO-SLICE" button to the WaveformDisplay component that uses energy-based peak detection in PhonemeAligner.ts to automatically detect drum hits and transients and populate custom slice markers.
  • [2026-06-21] - Implemented Custom Sample Slicing UI: Enhanced WaveformDisplay.tsx to support interactive mousedown, drag, and double-click events, allowing users to manually slice custom WAV files directly on the canvas. Connected to the AudioEngine via setAlignment.
  • [2026-06-20] - Implemented Glissando/Portamento Curves & Per-Step Breath Intensity: Added slideType parameter (Linear/Exponential) to allow musical variations of pitch glides in TTS, and allowed individual steps to override global breath noise via breathIntensity. Added Custom Sample Slicing UI to Active Backlog.
  • [2026-06-19] - Implemented Global Saturation: Added a master channel WaveShaperNode with a variable distortion curve mapped to a "Warmth" (Saturation) slider in the top utility UI. Routed the entire master mix through it to add glue and presence. Added new idea: "AI Auto-Mix Assistant".
  • [2026-06-18] - Implemented Dynamic Reverb: Added a ConvolverNode hooked up to the master output with a generated exponential decay noise impulse response. Mapped reverbSend from individual sequence steps in NoteSelector to send audio from the TTS SingingVoice into the new global reverb bus. Added new idea: "Global Saturation / Tape Warmth".
  • [2026-06-17] - Implemented Spectral Morphing: Added characterMorph and morphTarget to allow smooth interpolation between voice characters (e.g., male to female) per sequence step using FormantShifter.ts. Added Morph knob and target selector to SamplerPanel.tsx and override controls to NoteSelector.tsx.
  • [2026-06-16] - Implemented Phoneme-Aware Velocity: Modified triggerSlice in SingingVoice.ts to dynamically scale the amplitude envelope attack and decay based on the phoneme.category parameter (e.g., extremely fast attack for plosives, smoother attack for vowels). Added new Innovation Lab idea: Spectral Morphing.
  • [2026-03-31] - Implemented Phoneme-Aware Velocity: Modified SingingVoice.ts and triggerSlice to dynamically scale the envelope attack and decay based on the phoneme.category properties (e.g. extremely fast attacks for plosives, smooth for vowels) for more natural articulated vocal rendering. Added new idea: Voice Layering / Chorus per Voice.
  • [2026-06-15] - Implemented Step-Sequenced Vibrato: Added vibratoDepth controls to the NoteSelector component and wired them into App.tsx and useAudioEngine.ts to allow per-step overrides of the global Vibrato Depth for natural phrasing. Added new ideas: "Glissando/Portamento Curve Drawing" and "Phoneme-Aware Velocity".
  • [2026-06-14] - Implemented Step-Sequenced Formant LFO: Added formantLfoRate and formantLfoDepth controls to the NoteSelector component and wired them into App.tsx and useAudioEngine.ts to allow per-step overrides of the global Formant LFO settings for dynamic, rhythmic Wah-Wah effects. Added new idea: Custom Waveform LFO.
  • [2026-06-13] - Implemented Vocal Formant LFO: Added formantLfoRate and formantLfoDepth to FormantShifter.ts, routing an internal oscillator to the detune parameters of the peaking filters to create a rhythmic Wah-Wah effect. Added UI controls to SamplerPanel.tsx. Fulfills the "Vocal Formant LFO" Innovation Lab idea. Added new idea: Step-Sequenced Formant LFO.
  • [2026-06-12] - Implemented Chord Evolving: Added 'Chord Inversion' automation parameter allowing users to draw curves that dynamically shift the inversions of chords per step for synth and bass tracks, fulfilling the Chord Evolving Innovation Lab idea.
  • [2026-06-11] - Implemented Global Lyric Track: Replaced LyricMapper dialog with a dedicated LyricTrack component in App.tsx. Text generation now automatically maps syllables/words to consecutive active notes in the sequence using sliceIndex, fulfilling the Lyric Track Innovation Lab idea.
  • [Date] - Implemented Dynamic Tremolo Effect: Exposed Tremolo Rate and Depth knobs to the SamplerPanel UI and mapped them through useAudioEngine.ts and SingingVoice.ts down to the RubberBandProcessor AudioWorklet to allow dynamic amplitude modulation for rhythmic pulsing effects on vocals.
  • [2026-06-10] - Implemented Vocal Envelope Shaper: Upgraded ExpressiveVoiceProcessor to support a full ADSR (Attack, Decay, Sustain, Release) envelope for TTS samples. Added Decay and Sustain knobs to the SamplerPanel UI and mapped parameters via SingingVoice and useAudioEngine. Added a new Innovation Lab task for Tremolo Effect.
  • [2026-06-09] - Implemented Filter Envelope Mod: Added envMod control to NoteSelector and wired it through App.tsx and audioPlayback.ts to allow step-sequenced filter envelope modulation amounts.
  • [2026-06-08] - Implemented Per-Step Filter & Resonance: Mapped noteParams.filterCutoff and noteParams.filterResonance in useAudioEngine.ts to process per-step filter configurations correctly on both stretch (TTS SingingVoice) and standard buffer sampler playback nodes.
  • [2026-06-07] - Implemented LFO to Freeze Amount: Added freezeLfoRate and freezeLfoDepth to RubberBandProcessor parameter descriptors and processing logic to allow rhythmic granular cloud pulsing. Integrated UI controls in SamplerPanel.
  • [2026-06-07] - Implemented Per-Step Freeze Amount: Added freeze parameter to NoteSelector UI and mapped to noteParams in useAudioEngine and SingingVoice.ts for per-step granular synthesis control.
  • [2026-03-14] - Implemented Step-Sequenced Formant Shifts: Updated SingingVoice.ts to utilize FormantShifter.ts and enabled formant shifting in useAudioEngine.ts to process per-step timbre offsets correctly via Biquad filters.
  • [2026-06-06] - Implemented Spectral Granulator: Added freeze parameter to RubberBandProcessor to continuously loop a ~100ms Hann-windowed grain when activated. Wired to SamplerPanel and NoteSelector for global and per-step granular smearing.
  • [2026-06-06] - Implemented Formant Automation: Updated FormantShifter.ts, SingingVoice.ts, and useAudioEngine.ts to support continuous interpolation (linearRampToValueAtTime) of formant shifting over step durations during automation playback.
  • [2026-06-06] - Implemented Spectral Granulator (Freeze): Integrated granular freeze into rubberband-processor.ts by halting the sample read pointer and looping a ~100ms grain, smearing TTS phonemes. Added Freeze knob to SamplerPanel.
  • [2026-06-05] - Implemented Text-to-Drumkit Feature: Added a 'TEXT TO DRUMS' action in the Lyric Mapper that automatically populates the Kick, Snare, and Hi-Hat tracks by rhythmically mapping TTS phonemes based on their phonetic properties (vowels to kicks, fricatives/plosives to hats and snares).
  • [2026-06-04] - Implemented Dynamic Phoneme Pitch Glide: Enhanced SingingVoice and sequencer to slide between TTS phonemes smoothly when slide is active.
  • [2026-06-03] - Implemented Per-Step Reverse: Updated SingingVoice.ts and useAudioEngine.ts to allow reversing individual TTS phoneme slices on a per-step basis, matching the "Per-Step Reverse" Innovation Lab idea.
  • [2026-03-03] - Implemented A11y Colors: Updated getNoteColor logic to apply complementary hue shifts and lightness contrast for synth-2 (Bass/partB) notes vs synth-1 (Lead/partA), significantly improving visual separation in the sequencer.
  • [Date] - Roadmap re-initialized for long-term recursion.
  • [Date] - Implemented Vocal Envelope: Added Attack/Release envelope controls to ExpressiveVoiceProcessor and exposed them to SamplerPanel to allow softer vocal attacks and prevent clicks on rapid retriggering. Improved SingingVoice.ts precise noteOff timing handling via absolute target times.
  • [Date] - Implemented Glitch Mode UI: Added probability knob for random TTS retriggers.
  • [2026-06-02] - Implemented Polyphonic Singing: Added SingingVoiceManager to handle a pool of TTS voices, enabling chord playback and multi-voice "Choir" effects in the Sampler.
  • [2026-06-01] - Wired up Melodic Lyric Mode in App.tsx, connecting state to SamplerPanel toggle and MainSequencer rendering. Verified and tested NoteSelector accessibility.
  • [2026-05-30] - Implemented Note Slice UI: Visualized TTS phonemes directly on sequencer steps in MainSequencer.tsx. Verified absence of legacy Sequencer.tsx.
  • [2026-02-05] - Implemented Phoneme Elasticity in Sampler Engine.
  • [2026-05-21] - Refactored SingingVoice state management to eliminate type casting hacks and improve multi-bank alignment handling.
  • [2026-05-22] - Implemented TTS Slice Triggering (Phoneme Mode) in Audio Engine.
  • [2026-05-23] - Implemented Slice Mode UI in SamplerPanel, enabling Phoneme Slice Mode triggering via MIDI keys.
  • [2026-05-24] - Implemented Rubber-Band Selection (Shift+Drag) in the main sequencer view (App.tsx), enabling multi-step selection and bulk deletion.
  • [2026-05-25] - Implemented Hybrid Polyphony using VoiceManager, enabling 8-voice polyphony for Synth A and legato monophony for Synth B.
  • [2026-05-26] - Implemented Visual Slice Feedback in SamplerPanel using canvas-based WaveformDisplay and imperative playback highlighting.
  • [2026-05-27] - Implemented Clipboard Operations (Ctrl+C/V) and Drag-to-Edit (Painting) in the main sequencer view (App.tsx).
  • [2026-05-28] - Implemented Per-Step Parameters (Timbre, Probability, Microtiming) in Audio Engine and NoteSelector UI.
  • [2026-05-29] - Implemented Phoneme-Aware Time Stretching DSP in RubberBandProcessor, enabling dynamic vowel stretching during playback.
  • [2026-05-29] - Implemented Melodic Lyric Mode: Added sliceIndex to Note data, allowing independent pitch control and slice triggering for "Singing" TTS. Updated LyricMapper to preserve pitch.
  • [2026-05-30] - Refactored Sequencer UI: Extracted Sequencer, SequencerRow, SvgStep into dedicated components, removed legacy code, and centralized sequencer constants.

2026-06-26 - Implemented AI Auto-Mix Assistant: Added a deterministic heuristic system triggered via the '✨ AUTO-MIX' button in the master utilities UI. The assistant automatically sets levels and spreads panning across Synth A, Synth B, TB-303 Bass, Drum Kits, and 8-channel Sampler tracks. Connected StereoPannerNodes natively into Open303Manager, VoiceManager, and AudioPlayback to process 'pan' properly on the audio thread.