RealEye Webcam v1 vs v3

Published: Updated: 4 min read

RealEye offers two webcam eye-trackers: RealEye Webcam v3 and RealEye Webcam v1. Both use the participant's webcam to measure where people look on the screen — the difference is in the technology behind them, how calibration works, and which features they support.

RealEye Webcam v3 is the default eye-tracker for new studies. RealEye Webcam v1 is the older, legacy engine — it is still available and selectable in the study settings, mainly because it is the only option that supports facial coding and mobile fielding devices.

#Which RealEye Webcam should you choose?

In most cases, you should use RealEye Webcam v3. It is the modern, default eye-tracker and gives you accurate, reliable eye-tracking results on desktop computers with a quick and easy calibration.

Choose RealEye Webcam v1 when your study needs something that v3 cannot provide:

  • you want facial coding (emotion/expression analysis) — facial coding works only with v1,
  • you want to field your study on smartphones or tablets — v3 is optimised for desktop, so selecting a mobile fielding device automatically switches the study to v1.

If you don't need facial coding and your participants use desktop computers, stick with the default RealEye Webcam v3.

#RealEye Webcam v3 (the default)

RealEye Webcam v3 is the default eye-tracker for new studies.

  • Technology: v3 runs a machine-learning (neural network) model in the participant's browser inside a web worker. The model automatically falls back between execution backends as needed (WebGPU → WebGL → WASM) and uses MediaPipe for face detection. This keeps the experience smooth and does not require any hardware beyond a standard webcam.
  • Calibration: v3 uses a quick 4-point calibration (four corners of the screen), so participants can get started fast.
  • Fielding devices: v3 is optimised for desktop computers (PCs and laptops). Selecting a smartphone or tablet as a fielding device automatically switches the study to RealEye Webcam v1 — you can select v3 again manually afterwards if you prefer.
  • Face/head recovery: when head/eye position control is enabled, if a participant's face is no longer visible, v3 shows a face-visibility popup. The study pauses and automatically resumes once the participant's face has been stably detected for about 3 seconds.
  • Facial coding: not supported. When v3 is selected, facial coding is disabled — use v1 if your study needs emotion analysis.

#RealEye Webcam v1 (legacy)

RealEye Webcam v1 is the legacy webcam eye-tracker. It is still fully selectable for new studies.

  • Technology: v1 is based on the WebGazer engine. It tracks the participant's face mesh and trains a per-participant regression model during calibration, which is then used to predict gaze.

    You can learn more about the technology and accuracy behind RealEye Webcam v1 in our RealEye Technology Whitepaper.

  • Calibration: v1 uses a device-specific multi-point calibration pattern. Participants can also reuse a stored calibration from a previous session (within a 15-minute window and only when the screen size matches, within tolerance), which replaces full recalibration with a quick validation step.

  • Facial coding: facial coding (emotion/expression analysis) works only with v1. This is the reason v1 remains available.

  • Fielding devices: v1 supports desktop computers, tablets, and smartphones.

  • Face/head recovery: when head/eye position control is enabled and the participant's eyes are no longer tracked, v1 shows the eye-fix (green dots) recovery flow to help the participant get back into position.

#Comparison table

| | RealEye Webcam v3 | RealEye Webcam v1 | |---|---|---| | Default status | Default for new studies | Legacy, but still selectable | | Technology | Machine-learning (neural network) model in a browser web worker; automatic fallback WebGPU → WebGL → WASM; MediaPipe face detection | WebGazer engine: face mesh + per-participant regression model trained during calibration | | Calibration | Quick 4-point calibration (four corners) | Device-specific multi-point calibration; stored-calibration reuse (15-minute window, screen-size gated) | | Facial coding | Not available (disabled when v3 is selected) | Works only with v1 | | Fielding devices | Desktop (smartphone/tablet selection auto-switches to v1) | Desktop, tablet, and smartphone | | Face/head recovery | Face-visibility popup when head/eye position control is enabled (~3 s of stable detection before auto-resume) | Eye-fix (green dots) recovery flow when head/eye position control is enabled |

#Related articles