This dissertation investigates the perceptual limits of head‑mounted displays (HMDs) with respect to two interrelated phenomena: (1) the smallest amount of scene motion that users can detect when they turn their heads (scene‑motion thresholds) and (2) the smallest system latency that produces a perceptible lag between head movement and visual update (latency thresholds). The author built a simulated zero‑latency HMD using a projector and a high‑speed head tracker to precisely control and measure latency. Five psychophysical experiments were conducted. Experiments 1‑3 measured scene‑motion thresholds under different head‑yaw directions, speeds, and visual‑field conditions; Experiments 4‑5 measured latency thresholds directly and validated a mathematical model that predicts latency thresholds as an inverse function of peak head‑yaw acceleration. The model explains why latency perception is tightly coupled to the dynamics of head motion. The work also explores the use of visual distractors for reorientation, showing that distractors that trigger head turns in the same direction as world rotation can mask scene motion. Based on the empirical data, the author proposes a practical latency guideline of ≤3 ms for immersive virtual environments, outlines a hierarchy of hardware and software design choices to achieve minimal latency (display response time, frame‑buffer architecture, tracker latency, V‑sync handling, filtering, rendering speed, synchronization, and prediction), and discusses the trade‑offs of prediction‑based latency compensation. The dissertation concludes with recommendations for future HMD development, measurement protocols, and research on how attention, distractors, and user tasks influence perceptual thresholds.
Topics
Perceptual Limits and Design Guidelines for HMD Latency and Scene Motion
Establishes why quantifying the smallest detectable scene motion and system latency is essential for immersive VR/AR, linking these thresholds to user comfort, presence, and visual stability. Synthesizes experimental findings into a practical design target of ≤3 ms latency for imperceptible lag, highlighting the challenges revealed by the prototype system.
Simulated Zero‑Latency HMD Platform
Describes the custom hardware and software testbed built for the dissertation: a high‑speed head tracker, low‑latency video pipeline, and projector that together emulate a zero‑latency head‑mounted display, enabling precise manipulation and measurement of system delay.
Mathematical Model of Latency Thresholds
Presents the core equation T = k / a_peak, showing that latency perception thresholds are inversely proportional to peak head‑yaw acceleration. Details derivation, parameter fitting, and how the model accounts for the experimental data.
Distractor‑Based Reorientation for Masking Scene Motion
Explores how visual distractors that induce head turns in the same direction as world rotation can reduce perceived scene motion, drawing parallels to change‑blindness literature and suggesting a potential mitigation strategy for latency‑induced motion artifacts.
Low‑Latency Design Hierarchy and Prediction Trade‑offs
Ranks eight engineering actions—display response time, frame‑buffer architecture, tracker latency, V‑sync handling, filtering, rendering speed, synchronization, and prediction—by their impact on overall system latency. Analyzes the benefits and risks of head‑movement prediction, noting artifact potential when predictions err and emphasizing the need for consistent delay.