I have a question concerning the Wavenet voices of google text-to-speech (gTTS). When I look at the spectra of the synthesized speech, I can see that all Wavenet voices (in
