Testing of voice interface devices across multiple languages and locales is difficult due to factors such as the lack of availability of native speakers, inconsistency of human speech samples across languages, difficulty in scaling the number of human-provided query samples, etc. This disclosure describes the use of automated translation and text-to-speech generation technologies to obtain machine-generated audio in various languages. A set of query strings is translated into multiple languages and a text-to-speech synthesizer generates a consistent set of audio samples. The generated audio samples can be used to test voice interface devices.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Singhal, Amit; Chang, Yao-Jen; Subramanian, Srikrish; and Wang, Xinchen, "Automated Audio Generation for Testing Voice Interface Devices", Technical Disclosure Commons, (August 30, 2021)