A collection of audio projects for the CH32V003J4M6 RISC-V microcontroller, demonstrating various audio synthesis and playback techniques on an 8-pin, $0.10 microcontroller. From simple 1-bit music to compressed audio sample playback and speech synthesis. Click on the image to watch a video and hear the results.
This repository contains several firmware projects and development tools:
- SimpleSoundFirmware - Single-voice 1-bit music playback (monophonic)
- PolyphonicSoundFirmware - Advanced 8-voice polyphonic music with PWM synthesis
- SamplePlayback - Compressed audio sample playback (IMA ADPCM, QOA, 2-bit ADPCM)
- Talkie - Speech synthesis using the Talkie library
- Buzzer Studio - Interactive web app for designing 1-bit sound effects and exporting MIDI tracks
The Buzzer Studio project is available here: https://buzzer-studio.atomic14.com/
A word of warning - the Buzzer Studio codebase was creating with a lot of help from AI tools. It all runs client side, there is no backend for anyone to hack. But the code may not always be easy for a human to understand.
All firmware projects build with PlatformIO. To build locally:
Simple Sound Firmware (1-bit monophonic):
cd SimpleSoundFirmware
pio runPolyphonic Sound Firmware (8-voice PWM):
cd PolyphonicSoundFirmware
pio runSample Playback (compressed audio):
cd SamplePlayback
pio runTalkie (speech synthesis):
cd Talkie
pio runFor detailed firmware information, see each project's README:
- SimpleSoundFirmware/README.md - Original single-voice 1-bit music
- PolyphonicSoundFirmware/README.md - 8-voice polyphonic synthesis
- SamplePlayback/README.md - Compressed audio playback with multiple codecs
- Talkie/README.md - Speech synthesis implementation
Buzzer Studio (Web App):
cd buzzer-studio
npm install
npm run devThe app allows you to:
- Export tracks from MIDI files for 1-bit playback
- Design sound effects with real-time preview
- Export to C arrays for microcontrollers
- Export to Python for MicroPython/CircuitPython
- Choose from preset effects (jump, coin, laser, etc.)
- Play LPC encoded audio
- Encode WAV files to LPC
- Microcontroller: CH32V003J4M6 (RISC-V, 8-pin SOIC)
- Power: 3.3V supply
- Programming: WCH-Link or compatible programmer
SimpleSoundFirmware (1-bit music):
- Output: Piezo buzzer or small speaker (connected to PD6)
- Input: Trigger button or signal (connected to PC1)
PolyphonicSoundFirmware (8-voice music):
- Output: Piezo buzzer or small speaker (connected to PA1)
- No trigger input: Plays automatically on startup
SamplePlayback (compressed audio):
- Output: PWM audio output on PA1 (8-bit via PWM)
- No trigger input: Plays automatically on startup
Talkie (speech synthesis):
- Output: PWM audio output on PA1 (8-bit via PWM)
- No trigger input: Plays speech on startup
brain-transplant/
├── SimpleSoundFirmware/ # 1-bit monophonic music playback
│ ├── src/ # Source code (main.cpp, music data)
│ ├── platformio.ini # Build configuration
│ └── README.md # Documentation
│
├── PolyphonicSoundFirmware/ # 8-voice polyphonic music
│ ├── src/ # Source code (polyphonic player)
│ ├── platformio.ini # Build configuration
│ └── README.md # Documentation
│
├── SamplePlayback/ # Compressed audio sample playback
│ ├── src/ # Source code (ADPCM decoders, PWM player)
│ │ ├── AudioStream.h # Base interface for codecs
│ │ ├── IMAAdpcmStream.* # IMA ADPCM decoder
│ │ ├── ADPCM2BitStream.* # 2-bit ADPCM decoder
│ │ ├── player.* # PWM playback engine
│ │ └── audio/ # Audio data files
│ ├── platformio.ini # Build configuration
│ └── README.md # Documentation
│
├── Talkie/ # Speech synthesis
│ ├── src/ # Source code (Talkie library port)
│ │ ├── Talkie.* # Speech synthesis engine
│ │ ├── phrases.* # Pre-defined speech phrases
│ │ └── player.* # Audio playback
│ ├── platformio.ini # Build configuration
│ └── README.md # Documentation
│
├── scripts/ # Python audio conversion tools
│ ├── wav_to_adpcm_c.py # WAV → ADPCM C array converter
│ ├── adpcm_2bit_decoder.py # 2-bit ADPCM decoder
│ ├── compare_audio.py # Audio comparison tools
│ ├── pyproject.toml # Python dependencies (uv)
│ └── README.md # Scripts documentation
│
├── buzzer-studio/ # Interactive web app
│ ├── src/ # TypeScript source code
│ ├── public/ # Static assets
│ ├── index.html # App entry point
│ ├── package.json # Dependencies and scripts
│ └── README.md # App documentation
│
└── .github/
└── workflows/
├── build-firmware.yml # Firmware CI/CD
└── buzzer-studio-ci.yml # Buzzer Studio CI/CD
- SimpleSoundFirmware - 1-bit monophonic music playback
- PolyphonicSoundFirmware - 8-voice polyphonic synthesis
- SamplePlayback - Compressed audio playback with multiple codec support
- Talkie - Speech synthesis implementation
- Buzzer Studio - Web app for designing 1-bit sound effects and exporting MIDI
- Scripts - Python tools for audio conversion and ADPCM encoding
For Firmware Development:
- PlatformIO - Embedded build system
- CH32V platform support (installed automatically by PlatformIO)
- WCH-Link programmer (or compatible)
For Buzzer Studio App:
- Node.js 18+ and npm
- Modern web browser with Web Audio API support
For Python Scripts:
- Python 3.8+
- uv package manager
SimpleSoundFirmware (1-bit music):
- Waits for trigger signal on pin PC1
- Toggles output pin (PD6) at specified frequencies
- Uses microsecond-precision delays for accurate timing
- Simple 1-bit audio (square wave)
PolyphonicSoundFirmware (8-voice music):
- Automatically starts on power-up
- Manages up to 8 simultaneous voices using PWM synthesis
- Outputs mixed audio through PA1 at 8 kHz sample rate
- Software mixing with real-time voice generation
SamplePlayback (compressed audio):
- Decodes compressed audio using codec decoders (IMA ADPCM, 2-bit ADPCM)
- Outputs 8-bit audio via PWM on PA1 at 8 kHz
- Supports multiple codec formats with extensible architecture
- Compression ratios of 2:1 and 4:1
Talkie (speech synthesis):
- Uses Linear Predictive Coding (LPC) for speech synthesis
- Pre-defined vocabulary of phonemes and words
- Outputs synthesized speech via PWM on PA1
- Real-time speech generation from phrase data
