Skip to content

Sample data

Real recordings from Trinet cameras, exactly as the camera wrote them to its memory card — so you can try your pipeline before you have a camera. Every sample includes the video for each eye, the motion data (.imu), the per-frame timestamps (.vts) and the camera's calibration.

Wrist Kit sample: a Trinet Stereo GS head camera (top: rectified left and right eyes) and two Trinet Mono wrist cameras (bottom left and centre), all on one timeline. Depth and the point cloud were computed afterwards on a computer — the camera itself does not output depth.

Download

All samples are in one public Google Drive folder: Trinet sample data ↗. Each sample folder has a README with a quick start.

Trinet Stereo GS (global shutter)

Sample What's in it Length (min:s)
Sample 1 — with wrists ↗ Wrist Kit: Stereo GS head camera plus left and right Mono wrist cameras, one folder each 2:23
Sample 2 — with wrists ↗ Wrist Kit: Stereo GS head camera plus left and right Mono wrist cameras 1:57
Sample 3 — stereo only ↗ Stereo GS head camera, H.264, with extracted calibration.json 1:54
Sample 4 — stereo only, H.265 ↗ Stereo GS head camera, H.265, plus an MCAP file for Foxglove / ROS 2 0:47

Trinet Stereo (rolling shutter)

Sample What's in it Length (min:s)
Sample 1 ↗ Stereo head camera, H.265 1:45
Sample 2 ↗ Stereo head camera 3:27
Sample 3 ↗ Stereo head camera 1:32
Sample 4 ↗ Stereo head camera, plus a rendered stereo-depth and motion video 2:01

Most samples also include a rendered visualization video — the same kind of overview as above — so you can see what was recorded before you download the raw files.

What's raw and what isn't

The camera writes only the per-eye .mp4 files, the .vts timestamps, the .imu motion data and a small .tel telemetry file. Everything else in a sample folder — visualization videos, depth, calibration.json (the calibration is also embedded in each .mp4) and MCAP files — was produced afterwards on a computer with the Python toolkit.

Open a sample

  1. Install the Python toolkit.
  2. Inspect a recording to see its camera, timing and calibration.
  3. Visualize it and check sync, or export it to MCAP / ROS 2.

For the file layouts, see File formats; for how the clocks line up, see Timing and sync.

Sample names

Folder READMEs may call the Stereo GS camera "Trinet Pro Stereo GS" — that's the same camera. Samples were recorded on firmware from July to September 2026 and read with current tools.