🎵 neuro-sync

Two brainwaves, one soundtrack — links two Emotiv EEG headsets over the network and fuses their emotional states into a single real-time AI-generated music stream.

Each participant's headset feeds into a local brain processor. The two emotion streams are merged on the host and sent to Google Lyria, which generates music that reflects the blended emotional state of both people in real time.

How it works

Person A — Emotiv headset ──▶ client_brain_processor.py ──(HTTP)──▶ ┐
                                                                       ├─▶ social_audio_service.py ──▶ Google Lyria ──▶ 🎵
Person B — Emotiv headset ──▶ host brain_processor.py   ──(HTTP)──▶ ┘

1. Each device runs a brain processor that converts raw EEG → emotion state
2. social_audio_service.py collects both states and runs an emotion fusion algorithm
3. The blended state is converted to a detailed music prompt (BPM, key, instruments, dynamics)
4. Google Lyria streams the generated audio at 48 kHz stereo, < 500 ms latency

Emotion → Music mapping (sample)

Emotion	BPM	Key	Instruments
Happy	120–140	Major	Piano arpeggios, light percussion
Sad	60–80	Minor	Strings, sparse piano
Angry	140–160	Diminished	Distorted guitar, aggressive drums
Relaxed	70–90	Major pentatonic	Ambient pads, soft guitar

16 emotion classes supported in total.

Stack

• Hardware: 2 × Emotiv EPOC X (Cortex SDK v3)
• EEG processing: Python · numpy · asyncio
• Services: FastAPI / Uvicorn
• Music: Google Lyria real-time (lyria-realtime-exp)
• Audio: sounddevice · pyaudio (48 kHz stereo)

Quick start

pip install -r EEG/requirements.txt

# Host machine (also runs the audio service)
python EEG/host_main.py

# Client machine (second headset)
python EEG/client_main.py --host <HOST_IP>

Requires two Emotiv headsets, Cortex App on each machine, and a Google API key with Lyria access.