Edit: Fixed permissions for the clips, sorry they were restricted for a minute or two!
I'm dumb/weird/have too much spare time and decided to attempt to recreate Rin and Len's talking with VocaListener for Vocaloid Editor 4. I don't own any other Cryptonloid, so I couldn't conduct tests with them. (I guess if people demand I finish this "research", then I guess I could ask someone to render Miku/Kaito/Meiko for me.)
Room for error: I can't talk as high pitched as they can (especially Rin) and I did not edit VocaListener's analysis so the mechanical errors remain. Also, I am not including a clip of me speaking even though it would help by providing more "scientific evidence", because I really hate my voice and it's too embarrassing. Sorry. Just know I am female and using a cardioid mic + audio box. Also, Rin's test is shorter than Len's because I was dying inside while recording and just can't maintain such levels of genki-ness, it puts my inner goth at terminal risk.
Rin:
Original Project Sekai audio > Power > Sweet > Warm
Len:
Original Project Sekai audio > Power > Serious > Cold
I think Power sounds closest to the original dialogue.
VocaListener and NT's Cherry Pie are by the same company (AIST), except Cherry Pie uses deep learning based on the gender of the input voice.
Edit 2: Yeah, I am aware this is like... mega cursed content.