Characteristics of voicing source waveforms produced by esophageal and tracheoesophageal speakers

Voicing source waveforms produced by 10 laryngectomized esophageal speakers, 12 laryngectomized tracheoesophageal speakers, and 10 age-matched, normal male speakers were obtained by inversely filtering flow functions recorded with a circumferentially vented mask. The data from these speakers was used to evaluate differences in source signal properties on a group basis. In a second analysis source signals produced by a single laryngectomized individual, capable of producing esophageal and tracheoesophageal speech in a highly proficient manner, were analyzed and compared. The overall results of this project provide an initial and novel description of voicing source signals produced by esophageal and tracheoesophageal speakers. The characteristics of these signals appear to be highly variable, particularly in comparison with the homogeneous pattern of source waves produced by age-matched, normal speakers. This initial description of voicing source signals produced by alaryngeal speakers is generally supportive of the hypothesis of differences in source properties among normal, esophageal, and tracheoesophageal voices.