Second International Conference on Spoken Language Processing (ICSLP'92)
Banff, Alberta, Canada
The present paper reports a series of experiments designed to investigate the nature of perceptual compensation and the memory representations for spoken words produced at different speaking rates. The aim was to determine if variability in speaking rate has consequences for the encoding and processing of spoken words and if these consequences are comparable to those found for talker variability. A serial recall task was used to study the effects of changes in speaking rate and talker variability on the initial encoding, rehearsal, and recall of lists of spoken words. Presentation rate was manipulated to determine the time course and nature of processing. The results indicate that at fast presentation rates, variations in both speaking rate and talker characteristics incur a processing cost which influences the initial encoding and subsequent rehearsal of spoken words. At slower presentation rates, however, variation in talker results in improved recall in initial list positions while variation in speaking rate has no effect on recall performance. These results suggest that the processing of variability due to changes in speaking rate and talker differences may be the result of distinct operations. Talker information appears to be integrated into long-term representations of spoken words while rate information may be discarded or lost after initial stages of processing.
Bibliographic reference. Nygaard, Lynne C. / Sommers, Mitchell S. / Pisoni, David B. (1992): "Effects of speaking rate and talker variability on the representation of spoken words in memory", In ICSLP-1992, 209-212.