Web-based ear training for mixing engineers and audio professionals.
Train your ears to recognize frequencies, dynamics, reverb character, stereo width, transients, and panning — with reproducible A/B exercises built around real studio workflows.
Most ear training tools are either too gamified or too theoretical. ProAudioTrainer focuses on real studio decisions: clear references, measurable precision, and trackable progress — the same workflow you use when mixing professionally.
Upload your own reference tracks, train on real impulse responses from real spaces, and see your improvement over time with per-module score history.
| Login | EQ Trainer | EQ Match |
|---|---|---|
 |  |  |
| Dynamics | Reverb | Stereo Width |
|---|---|---|
 |  |  |
| Panning | Transient | Sound Library |
|---|---|---|
 |  |  |
| Module | What you train | Difficulty levels |
|---|---|---|
| EQ Trainer | Identify which frequency band is boosted/cut | 3 (±1 oct → ±0.25 oct) |
| EQ Match Trainer | Replicate a target EQ curve by ear (A/B) | 3 (visual → blind) |
| Dynamics Trainer | Distinguish compressor, limiter, gate, expander | 3 (obvious → subtle) |
| Reverb Trainer | Classify room character: cathedral, hall, room, tunnel, outdoor | Single (80+ real IRs) |
| Stereo Trainer | Perceive stereo width differences via M/S processing | 3 intensity levels |
| Panning Trainer | Identify L/C/R placement in the stereo field | 3 levels |
| Transient Trainer | Hear attack/decay characteristic changes | 3 levels |
| Sound Library | Upload, organize, and reuse your own reference audio | — |
| Progress Dashboard | View score history, training trends, and module stats | — |
| Layer | Technology |
|---|---|
| Frontend | Vanilla JavaScript, HTML5, CSS3 |
| Audio Engine | Web Audio API — BiquadFilter, AnalyserNode, AudioWorklets |
| Audio Library | Tone.js (pre-bundled) |
| Impulse Responses | EchoThief — 148 real-world IRs |
| Backend | Node.js 20 + Express |
| Auth | JSON Web Tokens (JWT) + bcrypt |
| Storage | JSON files + filesystem uploads (Docker volume) |
| Deployment | Docker Compose + Nginx reverse proxy |
Requirements: Docker + Docker Compose
# 1. Clone the repository git clone https://github.com/janwiebe05/ProAudioTrainer.git cd ProAudioTrainer # 2. Create environment file cp .env.example .env # 3. Start the stack docker compose up -d --build # 4. Open in your browser open http://localhost:8090Default credentials (change immediately after first login):
- Username:
admin - Password:
admin1!
Backend:
cd backend npm install npm run dev # nodemon, port 3001Frontend:
The frontend is static HTML/CSS/JS. Serve it with any static server or use the Docker setup. The Nginx config in nginx/nginx.conf proxies /api/ to the backend automatically.
backend/ src/ data/ # JSON persistence (gitignored — auto-created by server) middleware/ # JWT auth, error handler, ownership guard repositories/ # Data access layer (users, scores, library, exercises) routes/ # Express API routes (one file per trainer module) services/ # Business logic (auth, scoring, progress, admin) utils/ # JSON store helper, input validators uploads/ # User audio files (gitignored — persisted via Docker volume) server.js # Express app entry point frontend/ modules/ # One JS file per trainer module (self-registering) worklets/ # Custom AudioWorklet processors (gate, expander, MS-width) EchoThief/ # 148 impulse response WAV files (real-world locations) app.js # App controller: auth flow, module loader, VU meters index.html # Single-page app shell (vintage rack unit aesthetic) style.css # Core styles style-modules.css # Module-specific component styles nginx/ nginx.conf # Reverse proxy config (API + static file serving) Dockerfile docs/ screenshots/ # Interface screenshots and GIFs | Method | Endpoint | Description |
|---|---|---|
| POST | /api/auth/login | Authenticate user, receive JWT |
| POST | /api/auth/verify | Verify token validity |
| POST | /api/auth/register | Create new account |
| POST | /api/auth/change-password | Update password |
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/library | List uploaded audio files |
| POST | /api/library/upload | Upload audio file(s) (multipart, max 200MB) |
| GET | /api/library/:id/audio | Stream audio file |
| DELETE | /api/library/:id | Delete audio file |
| PATCH | /api/library/:id | Update file metadata |
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/eq-match/random?level=1|2|3 | Random EQ exercise |
| POST | /api/eq-match/evaluate | Evaluate EQ match answer |
| GET | /api/dynamics/random?level=1|2|3 | Random dynamics exercise |
| GET | /api/reverb/random | Random reverb exercise (IR-based) |
| GET | /api/panning/random?level=1|2|3 | Random panning exercise |
| GET | /api/stereo/random?level=1|2|3 | Random stereo width exercise |
| GET | /api/transient/random?level=1|2|3 | Random transient exercise |
| Method | Endpoint | Description |
|---|---|---|
| POST | /api/scores | Submit exercise score |
| GET | /api/scores/highscores | Top 10 global leaderboard |
| GET | /api/progress/overview | Per-user training statistics |
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/admin/users | List all users |
| POST | /api/admin/users | Create user |
| DELETE | /api/admin/users/:id | Delete user |
| GET | /api/health | Service health check |
Each trainer uses a combined time + precision score:
points = maxPoints × timeFactor × precisionFactor - timeFactor: Linear decay from 1.0 (instant answer) to 0.1 (timeout)
- precisionFactor: Based on distance from correct value (e.g., semitones off for EQ)
- Scores are persisted server-side per user and contribute to module-level leaderboards
- Spotify integration for reference-track A/B workflows (see issue template)
- Extended analytics with learning curves per module
- Genre-specific training presets (mastering, broadcast, film)
- Optional database backends (PostgreSQL / S3)
- Automated test coverage for scoring engine and audio processing
See CONTRIBUTING.md for setup instructions, coding guidelines, and the PR checklist.
For security issues, see SECURITY.md.
Impulse response audio files are from the EchoThief Impulse Response Library by Greg Hopkins, licensed under Creative Commons Attribution 4.0.
MIT — see LICENSE.