By CHARLOTTE KIESEL, PhD CANDIDATE

Grammar and vocabulary acquisition are typically emphasized in foreign language
learning. Pronunciation is interconnected with their acquisition, and good pronunciation
with a sense ofof belonging. Due to its role in learning, researchers have investigated both
audio, visual, and audiovisual feedback methods to support improvement. Among these,
audiovisual feedback is generally the most effective at improving pronunciation. It is
unclear, however, what makes it effective. Since there is a myriad of visual feedback
possible, and even more languages in the world, it is important to uncover why. In this talk,
I introduce my research in computer-assisted visual feedback for second language
pronunciation training. I discuss design implications based on participant feedback through
two studies of audio-only and audiovisual feedback. An exploratory study compares the
two methods through qualitative analysis of participant thought processes and reflection.
The following study builds on the results of the exploration through deployment of the
updated tool during a six week period. I present the initial results from interviews of
students during the period of practice. Finally, I discuss possible future directions for
second language pronunciation research, such as voice normalization to aid learner
comparisons, integrating speech to text, and annotating a database of audio through

crowdwork.