Enhancing movie experience by speech rate design of audio description

Speech rate conversion in screen readers is a crucial human-computer interface that allows persons with vision disabilities to access audiovisual media. However, its potential utility in audio descriptions (ADs) has not been researched extensively. This study investigated the effect of AD speech rates on information access and film aesthetics. The appropriate speech rate for human-narrated and synthesized ADs of short scenes from Japanese movies was evaluated by involving blind, partially blind, and sighted participants. An arranged staircase procedure and software developed to enable AD speech rate adjustments through keyboard operations were employed. The results obtained from blind and partially blind participants indicated lower mean appropriate speech rates of the single AD with human and synthesized voices for information access than those of screen readers in previous studies but higher than the standard spontaneous speech rate. The presence of the film sound did not significantly impact the average speech rate of either the human or synthesized AD. However, the variability of the speech rate decreased specifically for the synthesized AD. Additionally, a statistically significant positive correlation emerged between the appropriate speech rates determined for human and synthesized ADs, strengthened experimentally with short movie scenes with multiple ADs. The comments by blind and partially blind participants and sighted participants’ ratings demonstrated the positive effect of appropriate sequential changes in speech rate on the sense of presence and immersion in movies. The speech rate design of ADs presents an opportunity to improve the comfort of listening to ADs and augment the movie experience.