Voice fonts for individuality representation and transformation

Speaker individuality transformation is used to modify the speech signal's characteristics so that it sounds as if it is spoken by another speaker. Previous methods for individuality transformation use mapping functions which depend upon a pair of speakers. We introduce the paradigm of voice fonts to represent the individuality of a speaker, independent of other speakers. Several objective and subjective tests are conducted to evaluate the performance of the approaches proposed for the voice fonts paradigm. The results show that the voice fonts paradigm enables independent representation of a speaker's individuality and produces equally good quality of transformed speech compared to previous approaches. This independent representation will be useful in important applications which were not possible with previous methods.