An acoustic study of plosive consonants produced by patients with and without reconstruction after partial or total glossectomy. Focusing on VOT

摘要

This paper deals with the analysis of temporal characteristics of speech for patients with or without reconstructive lingual surgery, after partial or total glossectomy using VOT (Voice Onset Time) as a spatiotemporal cue. We first examined the impact of glossectomy on articulatory timing in 6 patients compared to articulatory behaviour in 3 control speakers. For each patient, we recorded an acoustic speech signal one day before surgery and 3 recording phases after surgery. Secondly, particular attention was paid to understanding and explaining how some patients - who were subject to severe alteration - could adapt their speech production relative to a new configuration of their vocal tract. The data consisted of twelve VCV sequences. Our study is an articulatory-acoustic based approach which allowed us to observe, within the acoustic speech signal, different cues that may permit tracing articulatory timing patterns. It has also been noticed that statistically significant changes in the duration of VOT appear only for voiceless consonants, in the early post-surgery phases. Consequently, time and speech therapy have a positive impact on the patients' VOT values in the late recording phases which are close to those obtained for control speakers.