Listeners' perceptions of consonants produced by esophageal and tracheoesophageal talkers

The purpose of this study was to investigate the consonant intelligibility of 3 esophageal (E) and 3 tracheoesophageal (TE) talkers, and 1 dual-mode (DM) talker proficient in both E and TE speech modes. Audio recordings of 24 English consonants produced by each talker in a consonant-vowel-consonant-vowel-consonant (CVCVC) context were presented in the sound field to 15 normal-hearing, naïve, young adult listeners who phonetically transcribed their responses using an open-response paradigm. Listeners' pooled responses were converted to confusion matrices and analyzed for overall intelligibility, voicing and manner features, and consonant omissions. Ratings of speech proficiency were also obtained. Overall, the intelligibility of the TE talkers was significantly better than that of the E talkers. The DM talker was also more intelligible in the TE mode. Voiced consonants, plosives, fricatives, nasals, and liquid-glides were significantly more intelligible when produced by TE talkers. Affricates were also more intelligible for the DM talker in his TE mode. The different patterns of intelligibility observed between the E and TE talkers studied may be due to temporal speech distinctions evolving from the influence of dissimilar driving sources upon the vibratory characteristics of the pharyngoesophageal segment. Clinical implications are presented.