Built
- USB MIDI capture from Roland
- Raspberry Pi Python service
- Real-time WebSocket communication
- Precise 88-key LED mapping
- MIDI search and playback
- Chromium kiosk interface
- Automatic startup with systemd
Functional system · Active development2026
PLUTO Piano Brain
Interactive MIDI learning and visualization system
Role
Product Design · Hardware · Firmware · Backend · UI
01
Project overview
I transformed a Roland digital piano into a networked learning and visualization instrument using a Raspberry Pi, ESP32, touch display and a calibrated LED system mapped to all 88 keys.
Live USB MIDI is processed by a Python service and distributed over WebSockets. Played notes illuminate immediately, while MIDI playback visualizes upcoming notes across the keyboard for a piano-roll-style learning experience.
The system runs independently without a laptop. A dedicated touch interface provides the MIDI library, search, playback controls and system status, while the wider RediON²⁶ platform can monitor and control it.
02
Real project media
＋
Real photographs and screenshots for this project will be added here.
03
System architecture
Roland Digital Piano
USB MIDI source
Raspberry Pi
Python MIDI · Library · WebSocket API
ESP32 · 196 LEDs
Low-latency calibrated lighting
7-inch Touch UI · RediON²⁶
Control and visualization
04
Project status
In active development
- Interactive lesson scenes
- Improved sheet-music interface
- Home Assistant entities
- Expanded piano-library controls
Planned
- Piano learning games
- Adaptive AI lessons
- Wall-mounted EQ visualizer
- Family and multiplayer profiles
05
Engineering decisions
Why Raspberry Pi instead of only ESP32?
The Roland needs a real USB-MIDI host, ALSA support, persistent services, local storage and HDMI output—responsibilities better suited to Linux.
Why a separate LED controller?
The ESP32 provides predictable low-latency LED updates while the Pi remains focused on MIDI processing, library management and orchestration.
Why WebSockets?
One real-time event stream can drive the LED controller, touch interface, RediON²⁶ dashboard and future visualizers without duplicating protocols.
06
Challenges solved
- 01Keeping MIDI playback alive when the UI refreshes
- 02Physically calibrating LEDs to all 88 keys
- 03Powering the LED system separately and safely
- 04Starting all services reliably with systemd
- 05Keeping frontend builds away from the limited Raspberry Pi
- 06Using one real-time protocol across multiple consumers
07
Technology stack
Hardware
- Roland Piano
- Raspberry Pi
- ESP32
- 196 WS2812 LEDs
- 7-inch HDMI Touch Display
Firmware
- C++
- PlatformIO
- FastLED
- Arduino-ESP32
Backend & services
- Python
- ALSA MIDI
- python-rtmidi
- asyncio
- WebSockets
- systemd
Interface
- Next.js
- TypeScript
- Chromium Kiosk
08
What it demonstrates
PLUTO Piano Brain demonstrates my ability to engineer a complete embedded product across hardware, firmware, Linux services, networking, real-time protocols and user-interface development.