Adapting Speech Synthesis Systems to Users with Age-Related Hearing Loss

This paper summarises the main results of a pilot study into the effect of auditory ageing on the intelligibility of synthetic speech. 32 older and 12 younger users had to answer simple questions about a series of meeting reminders and medication reminders. They also underwent an extensive battery of audiological and cognitive assessments. Older users only had more difficulty understanding the synthetic voice than younger people if they had elevated pure-tone thresholds and if they were asked to unfamiliar medication names. We suggest that these problems can be remedied by better prompt design. User interviews show that the synthetic voice used was quite natural. Problems mentioned by users fit the results of a previous error analysis.