Electrolaryngeal speech of Japanese untrained users (Sato et al., 2021)
mediaposted on 21.09.2021, 19:16 by Koji Sato, Junji Genda, Ryoya Minabe, Takumi Taniguchi
Purpose: The aim of this study was to investigate the characteristics of electrolaryngeal (EL) speech among untrained speakers to aid in its effective introduction and to identify syllables and words that are easy or difficult to pronounce.
Method: A total of 21 healthy individuals who had never used an EL were included. The participants were briefed, and tests comprising 100 Japanese syllables and 50 single words were conducted to evaluate EL speech intelligibility. A trained speaker was defined as a certified speech-language pathologist who underwent EL training for 3 months. A 5-point electrolarynx effectivity score (EES) was used for the subjective assessment of EL.
Results: The median (interquartile range) intelligibility scores of the untrained and trained groups were 24.0% (20.0%–34.0%) and 40.0% (36.0%–45.0%) for syllables and 48.0% (38.0%–60.0%) and 88.0% (82.0%–90.0%) for words, respectively. The intelligibility scores for syllables and words were higher in the trained group than those in the untrained group. Only two syllable subgroups (/m/ and /w/) had > 80% correct answers among untrained speakers. A total of 14 syllable subgroups (/k, kʲ, s, ɕ, t, tɕ,͡ ts, ɲ, h, ç, ɸ, p, pʲ, and a/), a number of which contained voiceless consonants, had < 40% correct answers among both speaker groups. A greater number of morae were associated with higher intelligibility scores. An EES of 4, indicating that the EL was effective, was the most frequent score.
Conclusions: It was difficult for untrained speakers to produce intelligible speech using an EL. Syllables, including voiceless consonants, were difficult to pronounce using an EL. Longer words with a greater number of morae were more intelligible, even for untrained EL speakers.
Supplemental Material S1. The schema showing a session to evaluate speech intelligibility using an EL. Each session included three participants (A, B, and C) who never used an EL before. One of the participants (untrained speaker) pronounced the syllables and words using an EL, whereas the other participant transcribed. Thus, every participant became an EL speaker once. Next, a trained speaker pronounced using an EL, while every three participants transcribed. A trained speaker was a certified speech-language pathologist and underwent EL training for 3 months. The same trained speaker participated in every session. We repeated a total of seven sessions in this study.
Supplemental Material S2. Video of a session.
Sato, K., Genda, J., Minabe, R., & Taniguchi, T. (2021). Characteristics of Japanese electrolaryngeal speech produced by untrained speakers: An observational study involving healthy volunteers. Journal of Speech, Language, and Hearing Research. Advance online publication. https://doi.org/10.1044/2021_JSLHR-21-00069
The principal investigator (K. S.) received a Grant-in-Aid for Scientific Research (KAKENHI) from the Japan Society for the Promotion of Science (18K08912).
Read the peer-reviewed publication
electrolaryngealspeechcharacteristicsuntrainedspeakersJapanesehealthysyllableswordspronunciationidentifyintelligibilityspeech-language pathologistSLPelectrolarynxtrainedvoiceless consonantsmoraeproductionintelligibletotal laryngectomyLaboratory Phonetics and Speech ScienceLinguistic Processes (incl. Speech Production and Comprehension)Communication Technology and Digital Media StudiesMedical Devices