Users of speech recognition technology often hyperarticulate (i.e., exaggerate) their speech in response to recognition failures and subsequent requests to repeat (e.g., \I'm sorry, I didn't... Show moreUsers of speech recognition technology often hyperarticulate (i.e., exaggerate) their speech in response to recognition failures and subsequent requests to repeat (e.g., \I'm sorry, I didn't understand, please repeat the input."). Hyperarticula- tion usually leads to further recognition failure. The goal of the current project is to develop a protocol for testing di erent talker characteristics of voice prompts in speech recognizers with an aim towards minimizing hyperarticulated speech from users. This IPRO is equally suited to students interested in the more tech- nical aspects of acoustic phonetics and voice recognition as well as the cognitive aspects of predicting user behavior in technology-mediated environments. Show less
