This presentation will demonstrate how language technology professionals can re-purpose existing speech recognition APIs in order to automate the evaluation of student pronunciation of the target language. Through low-cost and no-cost web applications, students can receive immediate and meaningful feedback on the accuracy and intelligibility of their output. The implementation of this software serves to reduce or eliminate the lengthy and tedious task of correcting and grading student audio recordings on an individual basis. We will explore and compare implementation options from Nuance, Google, Apple, AT&T, and iSpeech. The presentation will conclude with a demo that incorporates Google’s multilingual speech-to-text API that has been embedded into an online assessment tool.
