Acoustic analysis of the effects of alcohol on the human voice

An experiment has been designed and implemented to determine the physical effects of alcohol on the human voice Sustained phonation and prescribed text voice recordings were taken from twelve volunteers at different stages of alcohol consumption. A computer based analysis has been earned out on these voice recordings to determine whether a significant change in voice frequencies occurs when alcohol has been consumed. The effect of alcohol consumption on sentence duration has also been investigated. The results obtained from the analysis show that there is no association between the fundamental frequency of a sustained phonation and the level of alcohol consumption (regression coefficient = 0 146, F = 0 078, df= 71, p= 0 781). Similar results are recorded for the formant frequencies. For the utterance of vowel sounds within spoken sentences it was found that the fundamental frequency rose on average by 1 7Hz (std err = 0 57, F = 8 9, df = 237, p = 0 003) per ml alcohol per kilogram bodymass Prescribed-text, sentence duration is observed to increase by 6 4% (std err 0 5%, p < 0 0001) per ml of alcohol consumed per kilogram bodymass. There is no evidence that the sex of the speaker is related to the effect of alcohol on sentence duration. We conclude that there is no correlation observed between the level of alcohol consumed and variations m the characteristics of the human voice. A system has been developed whereby speech samples are filtered as a means of simulating noise induced hearing loss. A Chebychev low pass filter has been designed using the Matlab programming package and this filter has been applied to voice samples that are spoken in environments of different noise levels. The resulting voice samples give an indication as to the level of hearing damage incurred by individuals suffereing from noise induced hearing loss