I've just find out how to do singing synthesizing using free software only.
0. Obtain audio data.
1. Separate to pronunciation tokens.
2. Joint.
3. Adjust length and pitch.
That's dope! Are you an aspiring musician? What program(s) are you using?
LMMS is probably your best bet.
What I would like to do is to imitate subeteanatanoseidesu and similar style.
I almost forgot that I would have liked to implement same thing as tio.run but whose does not require CGI; which completes on just client environment. Also what I would like is just for an esolang Dis. It's to build on GitHub pages.
I have recently found somewhere comfortable.