Show HN: Open-source digital stylus with six degrees of freedom

Original Article ↗ Hacker News Discussion ↗

An open-source digital stylus that combines webcam-based marker tracking with inertial and pressure sensors is impressing people as a low-cost way to get six degrees of freedom and near-tablet-level accuracy. Commenters probe its technical trade-offs—rolling shutter artifacts, depth precision, drift during occlusion—and note room for improvement and integration with standards like OpenXR. The project also prompts broader reflections on 3D input hardware, from VR controllers to 6DOF CAD mice and space-sim joysticks, and on how open drivers and commodity cameras could unlock richer spatial interaction around a normal computer.

Overall impressions

  • Strong enthusiasm for an open‑source, 6DOF stylus that uses a standard webcam plus IMU and pressure sensor.
  • Many see it as realizing long‑standing “daydreams” about new 3D input methods.
  • Some concern that integration with existing ecosystems (e.g., OpenXR) is missing but fixable.

Tracking approach & sensor fusion

  • Uses outside‑in camera tracking with ArUco markers combined with inertial sensors and pressure sensing.
  • Commenters note similar approaches in VR (e.g., IR + IMU), but praise this for being open and using a commodity webcam.
  • Rolling‑shutter compensation on consumer webcams is highlighted as clever; the author notes it’s a crude but surprisingly effective approximation.
  • Depth is estimated via PnP from marker geometry; depth is the least precise axis but reportedly good enough once combined with IMU smoothing.

Accuracy, limitations, and ergonomics

  • Reported average tip error ≈0.9 mm over an A4 area; after alignment, stroke vs. ground truth differed by ≈0.16 mm.
  • Errors tend to be systematic (bias) rather than pure noise; edge distortions and corner‑drawing behavior could be improved via calibration and more tests.
  • Not yet as accurate as good graphics tablets, but commenters think software refinement could close the gap.
  • Practical drawing ergonomics (weight, balance, battery‑free operation) are flagged as critical if this is to replace tablets; some suggest current form is better suited as a VR/3D controller.

Occlusion, frame rate, and IMU limits

  • IMU can bridge low frame rates well (e.g., ~10 FPS still works), but cannot handle long occlusions; double‑integration drift dominates beyond ~150–200 ms.
  • Fully robust tracking through marker loss is seen as unlikely with this sensor setup.

Hardware, cost, and ecosystem

  • Parts list includes a specialized pressure sensor (~US$5 in some markets), small custom PCB, and a microcontroller/IMU module; total cost is modest but components are somewhat niche.
  • Some see a “missed opportunity” in not shipping an OpenXR layer; others note it’s open source and extensible.
  • Several users compare it to long‑standing 6DOF controllers (SpaceMouse, SpaceOrb, Wii‑based hacks), lamenting proprietary drivers and lack of open interfaces.

Potential applications

  • Suggested uses: VR and 3D manipulation, robotics control, precise 3D measuring (with added ranging), and remote whiteboarding without a tablet.
  • Broader speculation about future wide‑FOV or stereo webcams enabling richer 3D input around the screen, though some are skeptical given past failures (e.g., laptop touch bars, depth cameras).