I have just finished another experiment with copying real voices. This song is maybe better suited for Rikka’s voice. I wanted to try esp. the speech part.
Method of copying is the same (Praat segmentation & analysis, pitch transferred as in the song, minor manual tweaks, phonemes manualy adjusted to copy original).
Used also Maki & Saki AI light for unison backing vocals - not tuned, flat pitch.
The shouts are not good - 16 times multiplied by VST plugin from one Rikka’s voice.