Perception, 2015, volume 44, pages 145 – 156 doi:10.1068/p7779 Simulation of driving in low-visibility conditions: Does stereopsis improve speed perception? Kevin R Brooks, Mohammed E Rafat Department of Psychology, Macquarie University, Sydney, NSW 2109, Australia; e-mail: kevin.brooks@mq.edu.au Received 24 April 2014, in revised form 27 November 2014 Abstract. Laboratory-based studies of perceived speed show that, under most circumstances, perceived speed is reduced as a function of contrast. However, a recent investigation of perceived vehicular speed while driving around a closed road circuit showed no such effect (Owens, Wood, & Carberry, 2010, Perception, 39, 1199–1215). We sought to probe the source of this discrepancy, asking whether the presence or absence of stereoscopic motion information might account for the difference in results. In a two-alternative forced-choice psychophysical speed-discrimination task, observers compared the speed of high- and low-contrast driving clips filmed with a 3-D camera and presented either stereoscopically (3-D) or monoscopically (2-D). Although perceived speed was reduced at low contrast, the size of this misperception was equivalent for 2-D and 3-D presentations. However, the inclusion of stereoscopic cues to vehicular speed caused significant improvements in the precision of speed judgments. It is concluded that although stereopsis can provide access to valuable information on perceived speed, contrast-independent speed estimation as demonstrated by Owens et al. (2010) is more likely to reflect the use of the full visual field in a real driving situation (compared with limited field of view simulations), or the additional contributions of nonvisual cues rather than stereopsis. Keywords: speed perception, contrast, driving, fog, motion in depth, stereomotion, disparity, stereopsis 1 Introduction Road users are presented with significant dangers when driving in low-visibility conditions— for example, when drivers suffer from cataracts or are confronted with fog, or dust, or sand storms (Brazel, 1991; Edwards, 1998; Musk, 1991; Owsley, Stalvey, Wells, & Sloane, 1999; Owsley, Stalvey, Wells, Sloane, & McGwin, 2001). Despite widespread awareness of the need to slow down, drivers fail to reduce their speed adequately (Al-Ghamdi, 2007; Edwards, 1998; Musk, 1991; Trick, Lochner, Toxopeus, & Wilson, 2009). Under these circumstances, drivers may rely on their perceptual estimation of velocity while keeping their attention on the road, rather than risk glancing down at their speedometer (Brown, 1970). The visual input in such challenging conditions is degraded in a way that is summarised principally by a reduction of contrast, presenting multiple challenges to the human operator’s perceptual system. One potential source of visual error under low-contrast conditions, and a possible explanation for insufficient speed reduction, involves the underestimation of vehicular speed. The observation that perceived speed is reduced with decreases of contrast was made first by Thompson (1976). Since then, many subsequent studies have replicated this robust effect using simple sine-wave gratings (Campbell & Maffei, 1981; Hawken, Gegenfurtner, & Tang, 1994; Johnston, Benton, & Morgan, 1999; Stone & Thompson, 1992; Thompson, 1982; Thompson, Brooks, & Hammett, 2006; Thompson & Stone, 1997). Although these stimuli allow isolation of the spatial frequency channels known to operate in early human motion processing (Harris, 1986; Thompson, 1984), their unnatural appearance has led some to question their ecological validity. However, the effect persists when more complex targets, such as compound gratings (Brooks, Morris, & Thompson, 2011), random dot patterns,