FA+
Shop
Serious progress on jiggle engine!
6 years ago
I've been hard at it since Becumming #8, and my soft body physics engine is really starting to take shape:
https://twitter.com/bendzzTransform.....43124686540801
https://twitter.com/bendzzTransform.....74960184778754
https://twitter.com/bendzzTransform.....85214764904448
For those just joining us, I think physically simulating the body down to the bone is the missing key to making sexy 3D characters, especially naked ones. Hollywood's been doing it for a decade, but it's never been in games before- I'm not exactly sure why, but soft body sims can take hours to process per frame, so that's my bet.
Well. My simulation runs fine not just at game speeds, not just at VR speeds, but even on cheap mobile VR headsets like the Quest! The sim only eats about 20-25% of that little mobile GPU atm, hitting an easy 72 fps and leaving plenty of room for graphical TF goodness. I'm super hyped!
There's still a lot to do yet obviously, this isn't the final look. Aside from rebuilding the models and lighting to match my pay comics, we need:
-Collision detection and handling (for VR groping and sex)
-Proper point weighting (so the nipples don't drag like lead weights)
-Volume preservation (for squishy fun)
-Stretch limiting (really important for breasts)
-A zero G base model (so the body reacts to gravity accurately)
Plus lots of little things like preventing element inversions and more serious physics weighting tools. But good news! Based on past research papers all that stuff ought to require way less GPU power than what's already done.
Whew. It's been a long 3 years of R&D but it's really starting to look like this might happen! Thanks again to Bio and NRN for all the help. The porn implications here are obvious, but I'm perhaps most excited to see what it does for face animation =)
Of course, it's not a sure thing yet. I've been at it about 6 months full time and I bet it'll take another 6 months until sexy TF is happening. I can keep self funding, but dealing with the setbacks and coding grind is going to be the tough part. Cross your fingers.
https://twitter.com/bendzzTransform.....43124686540801
https://twitter.com/bendzzTransform.....74960184778754
https://twitter.com/bendzzTransform.....85214764904448
For those just joining us, I think physically simulating the body down to the bone is the missing key to making sexy 3D characters, especially naked ones. Hollywood's been doing it for a decade, but it's never been in games before- I'm not exactly sure why, but soft body sims can take hours to process per frame, so that's my bet.
Well. My simulation runs fine not just at game speeds, not just at VR speeds, but even on cheap mobile VR headsets like the Quest! The sim only eats about 20-25% of that little mobile GPU atm, hitting an easy 72 fps and leaving plenty of room for graphical TF goodness. I'm super hyped!
There's still a lot to do yet obviously, this isn't the final look. Aside from rebuilding the models and lighting to match my pay comics, we need:
-Collision detection and handling (for VR groping and sex)
-Proper point weighting (so the nipples don't drag like lead weights)
-Volume preservation (for squishy fun)
-Stretch limiting (really important for breasts)
-A zero G base model (so the body reacts to gravity accurately)
Plus lots of little things like preventing element inversions and more serious physics weighting tools. But good news! Based on past research papers all that stuff ought to require way less GPU power than what's already done.
Whew. It's been a long 3 years of R&D but it's really starting to look like this might happen! Thanks again to Bio and NRN for all the help. The porn implications here are obvious, but I'm perhaps most excited to see what it does for face animation =)
Of course, it's not a sure thing yet. I've been at it about 6 months full time and I bet it'll take another 6 months until sexy TF is happening. I can keep self funding, but dealing with the setbacks and coding grind is going to be the tough part. Cross your fingers.
I figured I'll write a paper once the game's better established and popular, so I can use this research as an early advantage. But tbh I've posted most of the core concepts, papers and innovations all over. If someone wanted they could duplicate the work in less time than it took me.
The second link one looks quite good really. The others might be a bit too jiggly.
And I'm super happy your considering Quest users and such, I can't afford a good PC, but a Quest is more within my price range, so I've been saving to get one for my birthday this year ;p So I'm glad someone is looking out for me!
-Since this has taken 3 years of effort I'm going to hang onto the algorithm for a while for my TF, instead of say, giving it to other community games right off the bat.
-I'm not PROMISING it's coming to the Quest, since that's actually a really big challenge. But so far it's looking good! And I really do like the Quest, it'd be awesome if it's feasible.
Also yeah, Totally understandable~
Unless you're suggesting something like not simulating parts of the model facing away from the camera. That'd be tough to do with how shaders handle IF statements, and probably a bad idea (how are her thighs supposed to squish out against a chair if her underside isn't simulating collision and volume?)
Turns out building people is hard
(that seemed better in my head)