The blur paradox: Better recognition at a distance.

Abstract

When faces are blurred, presenting them at smaller sizes improves recognition. We term this unexpected advantage the blur paradox, which has been replicated in studies where face images are digitally blurred and scaled. To examine whether the blur paradox persists in physically realistic viewing conditions, we conducted two experiments using physical blur filters and varied viewing distances for size manipulation. First, we tested blurry celebrity face recognition at two viewing distances and found that recognition accuracy was significantly greater in the far condition than in the close condition. Second, we examined whether the blur paradox reflects gradual improvement across viewing distances or a sharp change in recognition performance at a particular distance. Across four viewing conditions, we found a significant main effect of viewing distance, with the highest recognition accuracy at the farthest viewing condition and lowest at the closest. Accuracy improved gradually, but nonlinearly, rather than showing an abrupt shift at a boundary. Exploration of participant demographics suggested a stronger effect among older participants (>50 years) and a weaker effect among left-handed participants. No significant sex differences were observed. These findings confirm the small-size advantage for recognition under blur and its persistence in physically realistic conditions, with accuracy improving gradually across a wide range of distances.