The differentiation of Thai voiced, voiceless unaspirated, and voiceless aspirated stops in tracheoesophageal speech : acoustic analysis and perception
Abstract:
To study the acoustic characteristics of Thai initial and intervocalic stops in tracheoesophageal and normal speakers. It examines the hypothesis that the acoustic characteristics voice timing, closure duration, intensity and fundamental frequency discriminate voiced, voiceless unaspirated and voiceless aspirated stops for two groups of speakers. The research also has experiments for normal listeners to perceive initial and intervocalic stops spoke by tracheoesophageal speakers. It examines the hypothesis that listeners can detect differences between voiced and voiceless stops in initial and intervocalic position, but can not distinguish voiceless unaspirated and voiceless aspirated stops. The data were collected from words read by three tracheoesophageal and three normal speakers. The speakers are all male of the same age, education, domicile and size. The data includes initial stops of isolated words and intervocalic stops of bisyllabic words. For analyzing the data, the multi-speech program was used for voice timing, closure duration and intensity and the Praat program for fundamental frequency. For the perception test, tracheoesophageal words were recorded on compact disks. All 30 subjects listened through headphones in laboratory and chose answers from sets of responses printed on answer sheets. The research reveals that voice timing discriminate the three types of voicing in both initial and intervocalic stops in normal speakers. For tracheoesophageal speakers, voice timing distinguishes three types of stops, especially in intervocalic position. It discriminates initial voiced and voiceless stops, but can not discriminate initial voiceless unaspirated and voiceless aspirated stops. Closure duration in intervocalic position in tracheoesophageal speech and normal speech distinguishes voiceless unaspirated and voiceless aspirated stops, but does not distinguish voiced and voiceless stops. Both intensity and fundamental frequency of the syllable in these two groups of speakers do not discriminate three types of initial and intervocalic stops. This study concludes that the important acoustic characteristics for discriminating stops are voice timing and closure duration. The perception tests clearly reveal that the listeners detect differences between intervocalic stops for tracheoesophageal speech better than initial stops clearly. While the listeners can perceive three types of intervocalic stops, they can distinguish only initial voiced and voiceless stops only. They can not discriminate initial voiceless unaspirated and voiceless aspirated stops. The study of the ability of alarygeal speakers to produce Thai initial and intervocalic stops of tracheoesophageal speakers goes beyond the research of Gandour et al. (1987), which studied only Thai initial stops in esophageal speakers before tracheoesophageal speakers became available. The research reveals that intervocalic stops are better than initial stops in tracheoesophageal speakers. They have to control P-E segment instead of the vocal folds, so they are not good at the onset of speech. Later in utterance, they can control the P-E segment, so the rest of utterance is better.