Supra-threshold control of peripheral LOD

TL;DR

This paper investigates how level of detail (LOD) control should be adapted for supra-threshold perception in peripheral vision, revealing that it should differ significantly from threshold-based methods and depend on task and perceptibility levels.

Contribution

It demonstrates that supra-threshold LOD control in peripheral vision requires different strategies than threshold-based control, emphasizing task-dependence and contrast over size.

Findings

LOD should support task-dependent perceptibility levels

Perceptibility is better predicted by contrast than size

Above threshold, perceptibility should be minimized; below, maximized

Abstract

Level of detail (LOD) is widely used to control visual feedback in interactive applications. LOD control is typically based on perception at threshold - the conditions in which a stimulus first becomes perceivable. Yet most LOD manipulations are quite perceivable and occur well above threshold. Moreover, research shows that supra-threshold perception differs drastically from perception at threshold. In that case, should supra-threshold LOD control also differ from LOD control at threshold? In two experiments, we examine supra-threshold LOD control in the visual periphery and find that indeed, it should differ drastically from LOD control at threshold. Specifically, we find that LOD must support a task-dependent level of reliable perceptibility. Above that level, perceptibility of LOD control manipulations should be minimized, and detail contrast is a better predictor of perceptibility than detail size. Below that level, perceptibility must be maximized, and LOD should be improved as eccentricity rises or contrast drops. This directly contradicts prevailing threshold-based LOD control schemes, and strongly suggests a reexamination of LOD control for foveal display.