Publications by Karl Isaac

Keywords: irrelevant speech effect, reminders, spearcon, speech, working memory

When screening participants for speech perception experiments, formal audiometric screens are often not an option, especially when studies are conducted over the Internet. We investigated whether a brief standardized self-report questionnaire, the screening version of the Hearing Handicap Inventory for Adults (HHIA-S), could be used to approximate the results of audiometric screening. Our results suggest that while the HHIA-S is useful, it needs to be used with extremely strict cut-off values that could exclude around 25% of people with no hearing impairment who are interested in participating. Well constructed, standardized single questions might be a more feasible alternative, in particular for web experiments.

Microtask platforms such as Amazon Mechanical Turk (AMT) are increasingly used to create speech and language resources. AMT in particular allows researchers to quickly recruit a large number of fairly demographically diverse participants. In this study, we investigated whether AMT can be used for comparing the intelligibility of speech synthesis systems. We conducted two experiments in the lab and via AMT, one comparing US English diphone to US English speaker-adaptive HTS synthesis and one comparing UK English unit selection to UK English speaker-dependent HTS synthesis. While AMT word error rates were worse than lab error rates, AMT results were more sensitive to relative differences between systems. This is mainly due to the larger number of listeners. Boxplots and multilevel modelling allowed us to identify listeners who performed particularly badly, while thresholding was sufficient to eliminate rogue workers. We conclude that AMT is a viable platform for synthetic speech intelligibility comparisons.