Abstract
Change blindness has been demonstrated for objects and the visual form of whole scenes but one of the most striking and common examples of change blindness is our difficulty to detect cuts in an edited moving image. Edit Blindness (Smith & Henderson, JEMR, 2008) can be created by adhering to the continuity editing conventions of Hollywood such as coinciding a cut with a sudden onset of motion (Match-On-Action). In this study we isolated the roles motion and audio play in limiting awareness of match-on-action cuts by re-editing existing feature film clips. Motion before and/or after the cut was removed and presented within twenty second film clips either with their original soundtrack or in silence. Participants were instructed to detect all cuts as quickly as possible. Cut detection times for intact cuts were significantly longer than all other conditions and removing either motion before or after the cut resulted in longer RTs than removing both. Percentage of missed cuts was higher for intact cuts than any other condition, although this effect lessened when audio was removed. In a follow-on eyetracking study, the pattern of cut detection times and missed cuts was replicated and analysis of saccade frequencies revealed a delay in orienting to the new shot following intact cuts compared to the rapid orienting (~300ms) observed when pre and post-cut motion was removed. The presence of motion either before or after the cut also decreased saccade frequency but only when accompanied by audio. These results suggest that film editors intuited the power of motion to attract viewer attention and obscure the timing of a cut. The results also emphasise the role of audio in maximising visual continuity. These ideas are consolidated in the Attentional Theory of Cinematic Continuity (Smith, Projections, 2012).
Meeting abstract presented at VSS 2013