Hello!
I am interested in creating a model for DiffSinger for one of my characters, but do not know how I would go about it. And my computer is far too weak to train an AI model as well (it can barely run the program itself lmao). I would also pay, but I am a high school student with no way to work due to my current living situation. And so, I have to make this a request.
For my first request, I need help in the creation of the voicebank.
The voicebank would ideally be bilingual in both Japanese and English, maybe with the added phoneme of a rolled R sound. I would want it to have access to auto pitch and vocal modes as well.
There are two vocal modes (besides my normal singing style) I would want no matter what:
Fire: A powerful and bitter style with a punkish feel
Air: A soft and calming singing style with a wide open mouth
As well as some ones I would want to add a lot, but would be okay not adding (maybe in later updates?):
Water: Cool, strong, and uniquely accented vocal with a sweeter tone than normal or fire ideally used for music in the style of artists like Hiiragi Kirai
Light: A bright and thick vocal with an idol feel, meant to sound like it's smiling and energetic
Dark: A sad and mature vocal with a very breathy feel, maybe a little bit tenser than Air
Fairy: A very breathy and whispery falsetto style in the highest and cutest voice I can make with a childish accent, based after Miku's sweet append
The samples for these are not recorded yet, but I hope to get it through soon.
Now, for the second request I'd like to make: concept sketches
I don't need anything too detailed or anything, just some simple things to get the character's outfit to be used for his main vocal synth design. I will send you some ideas for some parts of the design that I want.
And for my last request: just some assistance in recording!
What are things to do with my singing? Not to do? Good songs for diffsinger training?
I am interested in creating a model for DiffSinger for one of my characters, but do not know how I would go about it. And my computer is far too weak to train an AI model as well (it can barely run the program itself lmao). I would also pay, but I am a high school student with no way to work due to my current living situation. And so, I have to make this a request.
For my first request, I need help in the creation of the voicebank.
The voicebank would ideally be bilingual in both Japanese and English, maybe with the added phoneme of a rolled R sound. I would want it to have access to auto pitch and vocal modes as well.
There are two vocal modes (besides my normal singing style) I would want no matter what:
Fire: A powerful and bitter style with a punkish feel
Air: A soft and calming singing style with a wide open mouth
As well as some ones I would want to add a lot, but would be okay not adding (maybe in later updates?):
Water: Cool, strong, and uniquely accented vocal with a sweeter tone than normal or fire ideally used for music in the style of artists like Hiiragi Kirai
Light: A bright and thick vocal with an idol feel, meant to sound like it's smiling and energetic
Dark: A sad and mature vocal with a very breathy feel, maybe a little bit tenser than Air
Fairy: A very breathy and whispery falsetto style in the highest and cutest voice I can make with a childish accent, based after Miku's sweet append
The samples for these are not recorded yet, but I hope to get it through soon.
Now, for the second request I'd like to make: concept sketches
I don't need anything too detailed or anything, just some simple things to get the character's outfit to be used for his main vocal synth design. I will send you some ideas for some parts of the design that I want.
And for my last request: just some assistance in recording!
What are things to do with my singing? Not to do? Good songs for diffsinger training?
