Voice Transformations for Extreme Vocal Effects

Extreme Vocal Effects in music are so recent that few studies have been carried out about how they are physiologically produced and whether they are harmful or not for the human voice.

Voice Transformations in real-time are possible nowadays thanks to new technologies and voice processing algorithms. This Master's Thesis pretends to define and classify these new singing techniques and to create a mapping between the physiological aspect of each EVE to its relative spectrum variations.

Voice Transformation Models based on these mappings are proposed and discussed for each one of these EVEs. We also discuss different transformation methods and strategies in order to obtain better results.

A subjective evaluation of the results of the transformations is also presented and discussed along with further work, improvements, and working lines on this field.