A Speech Quality and Intelligibility Assessment Project Using Google Voice

Abstract

Speech quality and intelligibility assessment is an essential topic in the area of audio and speech signal processing. It has become increasingly more important due to the rising demand in areas such as digital protocol analysis, Dolby-dts compliance analysis, audio device testing, etc. “Audio and speech signal processing” or similar courses are widely offered as senior elective or graduate level courses in the electrical engineering curriculum. One of the biggest challenges in conducting a course project focused on speech quality and intelligibility assessment is the lack of available data from real telephony systems. In this paper, a speech quality and intelligibility assessment project using Google Voice is introduced. The voicemail feature of Google Voice allows users to download their own voicemails which are real Voice-over-Internet-Protocol (VoIP) speech. Google Voice also automatically transcribes the voicemails, providing great material for students to examine the effects of vocabulary and contextual information on the intelligibility of speech. The Students’ feedback regarding their Google Voice user experience and learning experience were collected through an exit-survey. The survey results suggested that in general the students found it convenient and easy to record their speech using Google Voice; they also agreed that using real-life data offered realistic tests of the theory, and that the automatic transcription system allowed them to investigate the performance of a real-life speech recognition system. In closing, the conclusions and future plans are presented.