Intelligibility of mandarin speech produced using different vocal rehabilitation technique in quiet and in noise

A growing number of patients are unable to use their larynges because of laryngectomy, tracheotomy or other health reasons. Most of them produced voice using alaryngeal speech. In this study, the intelligibility of Mandarin tracheoesophageal (TEP), esophageal (ESP) and electrolarynx (EL) speech in different level of noise was investigated. Two ESP speakers, two TEP speakers, two EL speakers and two normal speakers were invited to produce speech stimuli. Twenty naive adult listeners were asked to assign numbers to audio-recorded speech samples in each of eight signal-to-noise ratio (SNR) conditions in four separate magnitude estimation scaling tasks. The data from listener's responses showed that the score of intelligibility decreased with the increase of background noise level for all types of mandarin speech, and revealed that there are significant differences in the intelligibility of mandarin among TEP, ESP, EL and normal (NL) speech as a function of the background noise level. These results may be reveal some information related to the effect of background noise on the intelligibility of tone alaryngeal speech and provide clinical insights for the selection of speech rehabilitation technique for laryngectomees.