Abstract
Speakers adapt their speech in response to communicative context and listeners' needs. For example, when talking to hard of hearing or non-native listeners, or in the presence of background noise speakers make their speech more intelligible by speaking clearly. However, it is not clear which adaptations speakers will make when the goal is not intelligibility, but voice recognition. In this paper, we describe a speech database collected using a novel Wizard of Oz technique to investigate speakers' vocal adaptations when they are prompted to sound either more intelligible (clear speech) or more recognizable (identity marked speech). We recorded 39 speakers interacting with a mock speech recognizer, which repeatedly misunderstands speech, and a mock speaker recognizer, which misrecognizes the speakers' voice. We also collected read speech which served as a baseline. All recordings have been orthographically transcribed, forced aligned and manually corrected.
Perepelytsia, Valeriia