A Comprehensive Evaluation of the TRACKPixx3 Eye-Tracker

High-quality data is essential for interpreting findings from eye-tracking experiments. However, manufacturer specifications are often limited to spatial accuracy and precision values obtained under ideal conditions. These metrics are insufficient for evaluating performance across a wider range of eye movements (e.g., smooth pursuit, microsaccades, blinks) that are critical in many research paradigms. While SR Research’s EyeLink 1000 Plus is widely considered a reference instrument, a comprehensive, independent evaluation of its direct competitor, the VPixx TRACKPixx3 (introduced in 2020), is so far missing. This study addresses this gap by directly comparing the two devices using an extended version of the benchmark proposed by Ehinger et al. (2019), which we supplemented with a new text reading task. Through simultaneous recordings across eight tasks, we evaluated key performance measures including accuracy decay, fixation durations, pupil dilation, smooth pursuit tracking, and the classification of microsaccades and blinks. Our findings indicate that the overall performance of the two eye-trackers is comparable. A key difference emerged at the sample level: The TRACKPixx3 produces substantially smoother sam-ple data due to a non-deactivable and undocumented internal filter, which has the potential to distort saccade kinematics and fixation onsets. Conversely, the EyeLink 1000 Plus exhibited slightly weaker performance on some metrics, an outcome we attribute to our specific recording setup which, due to practical constraints, was suboptimal for this device. Although the EyeLink 1000 Plus remains a de-facto benchmark, our evaluation establishes the TRACKPixx3 as a viable and competitive alternative for high-precision eye-tracking research.