Skip to content

Speech recognition: overview and accuracy

JALTCALL 2022

Gary Ross & Stephen Henneberry

Abstract: Speech recognition and synthesis are perhaps the last stage in interface technology. For the language learner, the ability to speak to a device that can simultaneously play different roles and accents (female, male, old, young, Irish, Scottish, New Zealand, Fijian, Canadian, US etc.) is opening up speaking opportunities for L2 learners. However, the machine learning engines that allow such interactions are somewhat of a black box. In this study supported by a JSPS Kakenhi Grant as part of a Japanese MEXT Grants-in-Aid for Scientific Research, we look at the accuracy and use of such systems, and this presentation will give an outline of how the technology works, its future, and the preliminary results of this ongoing study.