As said in title. The ultra realistic voices of "Wilbert" and "Donovan" change with long generation of text. Wilbert changes, after few sentences, into female british voice. ( Even in sample its heard?) And Donovan changes into normal instead of deep sound after few sentences. It would be good if someone could cjeck this out.