COMPARISON OF SPECTRAL AND VOICE QUALITY PARAMETERS OF VOICE TRAINED ANGLOPHONES AND FRANCOPHONES IN CONTINUOUS SPEECH
William Weiss · Canadian acoustics · 1993
Continuous speech of 23 subjects was recorded. The group was composed of Voice Trained (n = l l) and Untrained (n=12) Male and Female Anglophone subjects. The objective of the investigation was to find out how are spectral levels and voice quality compared to an equivalent group of Francophones that were the object of a previous study. Long term spectral analysis was applied to all recorded samples and Spectral Levels were determined for regions of FO, FI, 1000Hz (Ajj^); and the fol lowing ratios were computed: 1. Q=Fl-F0, 2. a ^ g = >1000Hz- <1000Hz. Analyses of variance were carried out in order to ascertain differences between Trained and Untrained subjects and between the linguistic groups. Results show: 1. a ^ g greater for Trained Anglophones than for Untrained Anglophones. 2. a ^ g not greater for Trained Francophones than for Untrained Francophones. 3. a ^ g greater for Trained Anglophones than for Trained Francophones. 4. a ^ g the same for Untrained Anglo phones and Untrained Francophones. 5. Trained Anglophones have higher Spectral Levels than Untrained Anglophones. 6. Trained Francophones have smaller Spectral Levels than Untrained Francophones. 7. Trained Anglophones have significantly higher Spectral Levels than Trained Francophones in the higher frequencies. 8. Untrained Francophones have significantly higher Spectral Levels than Untrained