Comparing synthesized versus pre-recorded tutor speech in an intelligent tutoring spoken dialogue system

Abstract

We evaluate the impact of tutor voice quality in the context of our intelligent tutoring spoken dialogue system. We first describe two versions of our system which yielded two corpora of human-computer tutoring dialogues: one using a tutor voice pre-recorded by a human, and the other using a lowcost text-to-speech tutor voice. We then discuss the results of two-tailed t-tests comparing student learning gains, system usability, and dialogue efficiency across the two corpora and across corpora subsets. Overall, our results suggest that tutor voice quality may have only a minor impact on these metrics in the context of our tutoring system. We find that tutor voice quality does not impact learning gains, but it may impact usability and efficiency for some corpora subsets.