Back to list

Perceived Breathiness, Pressedness, and Vocal Fry in Synthetic Voice Stimuli

Authors Windtner, S.
Year 2021
Thesis Type Bachelor's thesis
Topic Psychoacoustics
Keywords auditory perception, psychoacoustics, speech processing
Abstract In the present thesis, different voice types, i.e., breathy and pressed voice as well as vocal fry are generated by means of an existing model based synthesizer. The determination of the voice type is important for clinical health care, because it affects the indication, selection, evaluation, and optimisation of clinical treatment techniques. The voice samples are generated by using different parameters and values. In two listening experiments, the samples are assessed by experts and students. The listening experiments provide information on whether the chosen parameters can be considered representative and whether they were perceived as natural for the respective voice type. The thesis shows which parameter can be representative for pressed voice, breathy voice as well as vocal fry. Furthermore, parameters have been achieved in order to generate the aforementioned voice types. Results from the listening experiments showed that the opening quotient is highly significant for the perception of the respective voice type. Furthermore, an increase of the fundamental frequency leads to a weak perceived strength of the vocal fry.
Supervisors Höldrich, R.