Tesla Dashcam Telemetry Overlay

Extracts embedded SEI telemetry from Tesla dashcam MP4 files and composites a real-time overlay displaying vehicle data.

Features

The overlay displays:

• Speed (MPH)
• Accelerator pedal position (vertical bar)
• Autopilot/FSD state (OFF, FSD, AUTO, TACC)
• Rolling acceleration chart (5-second history)
  • Front/Back (X-axis) - coral
  • Left/Right (Y-axis) - teal
  • Up/Down (Z-axis) - yellow
• Brake indicator
• Turn signal indicators (left/right)

Requirements

• Python 3.8+
• ffmpeg
• Tesla dashcam video with SEI metadata (firmware 2025.44.25+, HW3+)

Python dependencies

pip install numpy pillow matplotlib protobuf

SEI Extractor

This tool depends on Tesla's SEI extractor. Clone it to /tmp/tesla-dashcam:

git clone https://github.com/tesla/dashcam /tmp/tesla-dashcam

Or adjust the path in the script.

Usage

# Output defaults to <input>-overlay.mp4
python overlay_dashcam.py /path/to/dashcam-video.mp4

# Specify a custom output path
python overlay_dashcam.py /path/to/dashcam-video.mp4 -o /path/to/output.mp4

How it works

The script uses a dual-pipe architecture for efficient processing:

1. ffmpeg decoder streams raw frames to Python
2. Python composites the overlay using:
   • matplotlib with blitting for the acceleration chart (only redraws lines, not axes)
   • PIL for indicator icons (pre-cached by state)
   • numpy for alpha blending
3. ffmpeg encoder encodes the output to H.264

Single pass, one frame in memory at a time.

Output

• Codec: H.264
• Quality: CRF 18
• Frame rate: ~36 fps (matches input)

License

MIT