Euterpe - Music Theory, Audio Production, and Generative Music
Library-only domain providing a comprehensive music technology platform covering music primitives (notes, intervals, scales, chords, rhythm, keys, form, dynamics, MIDI), advanced music theory (harmony, counterpoint, melody, set theory, orchestration), AI-powered generative music, a DAW (Digital Audio Workstation) engine, audio mastering, spatial audio (Dolby Atmos, ambisonics, binaural), and voice technology (cloning, TTS, singing synthesis). Named after the Greek muse of music and lyric poetry.
Overview#
The Euterpe domain implements a complete music technology stack spanning from fundamental pitch and rhythm representation through to AI-powered composition, professional audio production, and immersive spatial sound. The platform is built on a rigorous music theory foundation that supports not only Western tonal music but also microtonal systems, Indian ragas, Arabic maqamat, Byzantine neumes, and other global traditions. Multiple tuning systems are first-class citizens: standard A440, A432, Baroque A415, scientific pitch, just intonation, and Pythagorean tuning. The theory engine goes beyond simple scale and chord lookup to provide deep harmonic analysis, species counterpoint, melody generation, chord progression analysis, groove quantization, pitch class set theory, and orchestration guidance. The generative layer uses state-of-the-art AI models for text-to-music generation, stem separation, voice generation, melody generation, arrangement, audio inpainting, style transfer, and conditional generation. The studio layer provides a complete DAW engine with timeline, clip editing, MIDI editing, mixer, effects processing, virtual instruments, automation, and project management. Audio mastering provides professional mastering chains, mix analysis, stem mastering, format conversion, and AI-assisted mix engineering. The spatial audio module implements Dolby Atmos authoring, binaural rendering, higher-order ambisonics, VR/AR audio spatialization, and AI-powered spatial upmixing. Voice technology covers voice cloning, voice conversion, text-to-speech, text-to- singing synthesis, vocal processing, choir simulation, and voice analysis.
Architecture#
@euterpe/studio
(DAW engine, timeline, mixer,
effects, instruments, automation)
|
+---------------+---------------+
| | |
@euterpe/master @euterpe/spatial @euterpe/voice
(mastering (Atmos, binaural, (cloning,
chain, mix ambisonics, conversion,
analysis, VR audio, TTS, singing,
stem master, spatial upmix) choir, analysis)
format) | |
| | |
+-------+-------+-------+-------+
|
@euterpe/genesis
(text-to-music, stems,
melody gen, arrangement,
inpainting, style transfer,
conditional, quality)
|
@euterpe/theory
(harmony, counterpoint,
melody, progressions,
groove, set theory,
orchestration, genre,
ear training, notation)
|
@euterpe/core
(notes, intervals, scales,
chords, rhythm, keys, form,
dynamics, audio, MIDI)
Layer Summary#
- Core - Music primitives: pitch classes, notes, tuning systems, intervals, scales (including ragas and maqamat), chords and voicings, rhythm and time signatures, keys and tonality, musical form, dynamics and expression, audio buffer/format handling, and MIDI protocol implementation
- Theory - Music intelligence: harmonic analysis, species counterpoint, melody analysis and generation, chord progression patterns, groove quantization, pitch class set theory, orchestration, genre classification, ear training, and notation
- Generation - AI music creation: text-to-music, stem separation, voice generation, melody generation, arrangement, inpainting, style transfer, conditional generation, quality assessment, and training infrastructure
- Production - DAW engine with timeline, clip/MIDI editing, mixer, effects, instruments, automation, project management, and AI-assisted features; mastering chain with mix analysis, stem mastering, and format conversion; spatial audio with Dolby Atmos, ambisonics, binaural, and VR audio; voice technology with cloning, TTS, and singing synthesis
Libraries (7)#
@euterpe/core - Core Music Primitives and Foundations#
Comprehensive music theory and audio primitives library providing the foundational building blocks for the entire platform. Ten sub-modules cover all fundamental music concepts.
Notes and pitch implements pitch classes and notes with full accidental
support, enharmonic equivalence, and frequency calculation. Supports multiple
tuning systems as first-class constructs: STANDARD_TUNING (A440),
A432_TUNING, BAROQUE_TUNING (A415), SCIENTIFIC_PITCH_TUNING,
JUST_INTONATION_TUNING, and PYTHAGOREAN_TUNING. MIDI-to-frequency and
frequency-to-MIDI conversion with cent-level precision. Note creation from MIDI
numbers, scientific notation (e.g., "C4"), and raw frequencies. Microtonal note
support with arbitrary cent offsets. Multiple notation formats: scientific
pitch, Helmholtz, solfege, Nashville number system, Indian sargam syllables, and
Byzantine neumes.
Intervals provides interval construction, inversion, quality classification (major, minor, perfect, augmented, diminished), semitone distance calculation, and compound interval handling.
Scales and modes implements Western scales (major, minor, harmonic minor, melodic minor, chromatic, whole-tone, pentatonic, blues) plus world scales including Indian ragas and Arabic maqamat. Mode rotation and scale degree naming.
Chords and voicings covers chord construction from root and quality, chord inversion, voicing generation, chord symbol parsing, and extended chords (7ths, 9ths, 11ths, 13ths, altered dominants, polychords).
Rhythm and time defines time signatures (simple, compound, complex, additive), note durations, dotted and tuplet divisions, tempo with BPM, and rhythmic pattern representation.
Keys and tonality provides key signature management, relative/parallel key relationships, key detection from note collections, and modulation path finding.
Musical form defines structural units (phrase, period, section, movement) and common formal templates (sonata, rondo, binary, ternary, strophic, through-composed).
Dynamics and expression represents dynamic markings (ppp through fff), hairpins (crescendo, diminuendo), articulations (staccato, legato, accent, tenuto, marcato), and tempo modifiers (ritardando, accelerando, fermata).
Audio handles audio buffer creation and manipulation, sample rate conversion, format detection, and waveform representation.
MIDI implements MIDI message types (note on/off, control change, program change, pitch bend, system exclusive), channel management, and MIDI file parsing/generation.
Dependencies: none (zero runtime dependencies)
@euterpe/theory - Music Theory Intelligence Engine#
Deep music understanding and analysis with ten sub-modules.
Harmony provides harmonic analysis including Roman numeral analysis, chord function identification (tonic, dominant, subdominant, pre-dominant), secondary dominants, borrowed chords, Neapolitan and augmented sixth chords, pivot chord modulation detection, and harmonic rhythm analysis.
Counterpoint implements species counterpoint rules (first through fifth species), voice leading validation, parallel fifths/octaves detection, proper dissonance treatment, melodic interval constraints, and cantus firmus generation.
Melody analyzes melodic contour, range, intervallic content, climax identification, motivic patterns, sequence detection, and conjunct/disjunct motion. Generates melodies constrained by harmonic context.
Progressions catalogs common chord progression patterns by genre and style, provides progression generation with voice leading optimization, and analyzes existing progressions for functional harmony.
Groove handles rhythmic quantization (straight, swing, shuffle), groove template application, humanization algorithms that add natural timing variation, and groove extraction from performance data.
Set theory implements pitch class set operations (prime form, normal order, Forte number, interval vector, Z-relation), set-class catalog, and transformational theory (transposition, inversion, multiplication).
Orchestration provides instrument range and transposition data, scoring guidelines, instrument combination suggestions, orchestral balance analysis, and arrangement voicing distribution.
Genre classifies musical characteristics by genre with stylistic feature extraction and genre-appropriate compositional suggestions.
Ear training generates interval, chord, scale, and progression recognition exercises with difficulty progression and performance tracking.
Notation handles music notation rendering data, clef assignment, beam grouping, tie and slur logic, and notation export preparation.
Peer dependencies: @euterpe/core
@euterpe/genesis - AI Music Generation Engine#
State-of-the-art generative AI for music creation with ten sub-modules.
Text-to-music generates audio from natural language descriptions using diffusion-based or transformer-based models with prompt engineering for musical style, instrumentation, mood, and structure control.
Stems provides audio stem separation (vocals, drums, bass, other) using deep neural networks, enabling remix workflows and selective regeneration.
Voice generates vocal content including singing voice synthesis from lyrics and melody, vocal style transfer, and voice-conditioned generation.
Melody generation creates melodic lines constrained by harmonic context, style parameters, and structural requirements using both rule-based and neural approaches.
Arrangement automates arrangement creation from lead sheets or simple inputs, distributing musical ideas across instruments with appropriate voicings, rhythms, and dynamics.
Inpainting fills in missing or selected regions of existing audio with contextually appropriate generated content, enabling seamless editing and extension.
Style transfer transforms the style of existing musical content while preserving structural and melodic identity, enabling cross-genre adaptation.
Conditional generates music conditioned on multiple simultaneous constraints (chord progression, melody, rhythm pattern, instrumentation, energy curve).
Quality assesses generated audio quality using perceptual metrics, musical coherence scoring, and artifact detection.
Infrastructure manages model loading, inference optimization, GPU/CPU dispatch, batching, caching, and model versioning for the generation pipeline.
Peer dependencies: @euterpe/core, @euterpe/theory
@euterpe/studio - DAW Engine#
Complete Digital Audio Workstation engine with ten sub-modules.
DAW engine provides the core audio processing graph, real-time audio rendering, buffer management, latency compensation, and transport controls (play, pause, stop, record, loop).
Timeline manages the arrangement timeline with track lanes, regions, markers, tempo maps, time signature changes, and timeline zoom/scroll navigation.
Clip editing handles audio clip manipulation including cut, copy, paste, trim, fade in/out, crossfade, time stretch, and pitch shift with non- destructive editing.
MIDI editing provides piano roll editing, step sequencing, MIDI quantization, velocity editing, controller lane editing, and MIDI transformation tools.
Mixer implements a channel strip architecture with gain staging, pan, mute/solo, send/return routing, bus grouping, and master output with metering.
Effects provides audio effect processors including EQ (parametric, graphic, shelving), dynamics (compressor, limiter, gate, expander), time-based effects (reverb, delay, chorus, flanger, phaser), and distortion/saturation.
Instruments implements virtual instruments including synthesizers (subtractive, FM, wavetable, granular), samplers, drum machines, and instrument preset management.
Automation enables parameter automation with breakpoint editing, automation curves (linear, bezier, step), automation lanes per track, and automation recording from controller input.
Project management handles session save/load, project file format, audio file reference management, undo/redo history, and project export.
AI features integrates AI-assisted tools including auto-mixing, intelligent EQ suggestions, mastering assistance, and AI-powered audio repair.
Peer dependencies: @euterpe/core, @euterpe/theory, @euterpe/genesis
@euterpe/master - Audio Mastering#
Professional audio mastering pipeline with five sub-modules.
Mastering chain implements the mastering signal processing chain including multiband compression, stereo enhancement, harmonic excitation, limiting, and loudness normalization (LUFS targeting per streaming platform specifications).
Mix analysis provides spectral analysis, dynamic range measurement, stereo field visualization, loudness metering (integrated, short-term, momentary LUFS), true peak detection, and crest factor analysis.
Stem mastering enables mastering from individual stems (drums, bass, vocals, instruments) with per-stem processing before final limiting, providing more control than stereo mastering.
Format masters handles final format conversion including sample rate conversion, bit depth dithering, codec encoding (WAV, FLAC, AAC, MP3, Opus), and format-specific metadata embedding.
Mix assistant provides AI-powered mix improvement suggestions based on spectral balance analysis, dynamic range assessment, and reference track comparison.
Peer dependencies: @euterpe/core, @euterpe/theory
@euterpe/spatial - Immersive Audio and Spatial Sound#
Spatial audio processing platform with five sub-modules.
Atmos implements Dolby Atmos authoring tools including object-based audio positioning, bed channel management, metadata generation, renderer integration, and Atmos ADM (Audio Definition Model) file creation for immersive music production.
Binaural provides binaural audio rendering using HRTF (Head-Related Transfer Function) processing for headphone-based 3D audio. Includes HRTF database management, personalized HRTF estimation, and head tracking integration.
Ambisonics implements higher-order ambisonics (HOA) encoding, decoding, and transformation. Supports B-format and higher-order spherical harmonics with rotation, beamforming, and speaker-array decoding.
VR audio provides spatial audio for virtual reality and augmented reality applications including room acoustics simulation, distance attenuation modeling, occlusion, and real-time source positioning tied to 3D scene state.
Spatial upmix uses AI models to convert stereo or multichannel audio to immersive spatial formats, analyzing the source material to intelligently distribute elements across the spatial field while preserving artistic intent.
Peer dependencies: @euterpe/core
@euterpe/voice - Voice Technology#
Voice and singing technology with seven sub-modules.
Voice cloning creates synthetic voice models from reference audio samples, enabling voice reproduction for narration, singing, and content creation with speaker verification and consent management.
Voice conversion transforms one speaker's voice to sound like another while preserving linguistic content, supporting real-time and offline processing with multiple conversion approaches (encoder-decoder, GAN-based, diffusion-based).
TTS (Text-to-Speech) generates natural speech from text with support for multiple languages, speaking styles, emotion control, SSML markup, and prosody adjustment.
Text-to-singing synthesizes singing vocals from lyrics and melody input, supporting multiple vocal styles, vibrato control, breath simulation, and expression parameters.
Vocal processing provides vocal-specific audio processing including pitch correction (auto-tune style and transparent), formant shifting, vocal doubling, harmonizer effects, de-essing, and breath removal.
Choir simulates choir and ensemble vocals from single voice inputs, generating multiple virtual singers with natural variation in timing, pitch, and timbre to create realistic choral textures.
Voice analysis extracts vocal characteristics including pitch tracking, formant analysis, voice quality assessment, singing technique evaluation, and vocal range detection.
Peer dependencies: @euterpe/core, @euterpe/theory
Technology Stack#
| Component | Technology |
|---|---|
| Language | TypeScript (ESM) |
| Runtime | Node.js |
| Build | Nx with @nx/js:tsc |
| Testing | Vitest |
| Audio | Web Audio API compatible, custom DSP |
| Spatial | Dolby Atmos, Ambisonics, HRTF binaural |
| MIDI | Full MIDI 1.0 protocol |
| AI/ML | Diffusion, transformer, GAN-based generation |
| Notation | Scientific, Helmholtz, solfege, sargam, neume |
| Tuning | A440, A432, Baroque, Just, Pythagorean |
Development#
# Test a specific library
pnpm nx test @euterpe/core
# Build all Euterpe libraries
pnpm nx run-many --target=build --projects=tag:scope:euterpe
# Lint a library
pnpm nx lint @euterpe/theory
# Run all domain tests
pnpm nx run-many --target=test --projects=tag:scope:euterpe
Configuration#
- Project tags:
["scope:euterpe", "layer:domain", "type:lib"] - Build executor:
@nx/js:tsc - All libraries use ESM (
"type": "module") - Core has zero runtime dependencies
- Libraries use workspace peer dependencies for inter-library references
Dependency Graph#
@euterpe/core (standalone, zero deps)
|
+-- @euterpe/theory (peer: core)
| |
| +-- @euterpe/genesis (peer: core, theory)
| | |
| | +-- @euterpe/studio (peer: core, theory, genesis)
| |
| +-- @euterpe/master (peer: core, theory)
| |
| +-- @euterpe/voice (peer: core, theory)
|
+-- @euterpe/spatial (peer: core)
The dependency chain reflects the layered architecture: core provides music
primitives, theory adds analytical intelligence, and the production libraries
(genesis, studio, master, voice) build on both. spatial depends only
on core since spatial audio processing operates on audio buffers independently
of music theory. studio has the deepest dependency chain, requiring genesis
for its AI-assisted production features.