Acoustic cues to the voicing feature in tracheoesophageal speech

Tracheoesophageal (TE) speakers often have difficulty producing the voiced/voiceless distinction. This limitation has been attributed to use of the pharyngoesophageal segment as the phonatory source. The nature of this tissue may preclude precise control of voicing onset, a contributing cue to a phoneme's voicing feature, at least in laryngeal speech. The purpose of this study was to determine whether voiced and voiceless consonants produced by TE speakers could be differentiated from those produced by laryngeal speakers using four acoustic measures associated with the voicing characteristic of consonants in laryngeal speech. Sixteen TE and ten laryngeal speakers produced five stop and fricative cognate pairs embedded in a carrier phrase. Three of the four acoustic measures contributed significantly to the discriminant models that differentiated accurately perceived TE and laryngeal samples. The three variables were consonant sound pressure level, consonant duration, and preceding vowel duration. In general, values for each measure were higher/longer for the TE group. The discriminant functions were interpreted as a reflection of TE speaker attempts at overarticulation.