Eight subjects, including one laryngectomee, were invited to participate in the assessment. First, the empirical parameters of phonation onset/offset estimation were compared with the optimal parameters obtained by minimizing simulation errors during video-EL. Second, a reaction time test was used to evaluate the ability of subjects to pronounce a single word with video-EL. Third, the fluency of subjects in producing long sentences with video-EL was calculated. Finally, the intelligibility of speech produced with video-EL was compared with that produced with button-EL.
The empirical parameters were not significantly different from the optimal parameter and resulted in fewer interruptions during voicing. Video-EL resulted in slower voice initiation and termination when compared with button-EL, which affected the intelligibility of an isolated word. However, video-EL provided a sufficiently fluent voice source so that the intelligibility of speech produced with video-EL was not significantly different from speech produced with button-EL when producing sentences.
The method proposed in this study is effective in the automatic on/off control of an EL. Subjects produced fluent speech with video-EL that was as intelligible as that produced with button-EL when Mandarin sentences were produced continuously.