Fig. 3From: A unit selection text-to-speech-and-singing synthesis framework from neutral speech: proof of conceptSinging pitch curve generation. Preparation (upper left), overshoot (upper right), the applied masks (middle), and the resulting mix (bottom)Back to article page