Page 1 of 2

4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 12:23 am
by Dragonlord
Dragons are not really what this forum is about ( although they are dinosaurs but that's another problem :P ) but since this work here is also usable for dinosaurs ( 4-legged ones at last for the time being ) it might be interesting nevertheless. Eventually got the ground-control animation code to work in a way it is usable although not fully mature yet. Ground-control animation for those not in the know is about how game actors ( actors are any kind of mesh which acts on their own usually player or non-player characters ) interact with the ground below. Most games simply play the animation or at best tilt the entire model which causes feet to dangle in the void or penetrate through the ground. More interesting to code though is a proper ground control where actors properly interact with the ground. This is especially useful for large actors where this can add to the look and feel of a scene. I don't want to waste too much text here. Just read the video description for some wall of text about what it's all about. Inverse kinematics still sucks as I somehow can't get the penalty mode to work which is why broken legs happen. Hopefully I can get rid of this over time.
Wakeboard testing | Dragon on bumpy terrain | Dragon on a real wake-board :P
The last video happened due to a joking comment of somebody from ModDB. Couldn't resist to whack it together for shits and giggles.

Last but not least a bit of physics playing around. Now trespasser dinosaurs with that kind of physics would be interesting :D
Dragon Rag-Doll Test
Hopefully I didn't annoy you with that folks.

Re: 4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 6:50 am
by Nick3069
Good job! That's quite impressive for an indie game engine. :D

Re: 4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 10:12 am
by TheGuy
Wow, haha, that's really cool! It seems quite life-like, weirdly.

It's amazing what a difference proper ground interaction makes to the believability of what you're seeing.

Re: 4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 6:03 pm
by Dragonlord
Yeah it gives an entire new dimension to immersion ( if I am allowed to use this over-used word ). The code itself works also with 2-legged actors although I'm still mocking around with that code. Imagine some raptors hunting around in the woods using this kind of ground interaction paired with good animation sequences. That would give quite some intense survival type game I could imagine :P

Re: 4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 7:53 pm
by Draconisaurus
Pretty awesome stuff, there. 8) Always enjoy a good dragon in action...

My only crit - in the ragdoll video, I noticed the rear legs falling into that mayanesque step-pyramid a lot :P

Re: 4-legged animation testing / rag-dolling

Posted: Tue Dec 22, 2009 8:30 pm
by Dragonlord
Draconisaurus wrote:Pretty awesome stuff, there. 8) Always enjoy a good dragon in action...

My only crit - in the ragdoll video, I noticed the rear legs falling into that mayanesque step-pyramid a lot :P
This is because they are not rigged with collision objects yet. If you look at the beginning I turn on the collision volumes ( orange shapes ) for a short time. Only bones with a collision object and constraints are affected by physics. The other bones are considered kinematic and move together with their parent bone or can be animated. I'll post another video once I fully rigged the guy. I need though to first implement some improvements to avoid self-collisions where they are not wanted. Most problematic will be the wings. I still have to figure out a way on how to do them best.

Re: 4-legged animation testing / rag-dolling

Posted: Wed Dec 23, 2009 1:33 am
by hppav
Draconisaurus wrote:Pretty awesome stuff, there. 8) Always enjoy a good dragon in action...

My only crit - in the ragdoll video, I noticed the rear legs falling into that mayanesque step-pyramid a lot :P
Draco fails at reading video descriptions :P

Nice job, I still remember when proper foot-ground interaction was an unheard of thing.

Re: 4-legged animation testing / rag-dolling

Posted: Wed Dec 23, 2009 3:32 pm
by Draconisaurus
Dragonlord wrote:Most problematic will be the wings. I still have to figure out a way on how to do them best.
Yeah I noticed they were not really working yet. I think first they need to be fully rigged - that is, with the complete joints that they'd use to do a flying motion (and so a flying animation might need to be finished first). They are just kinda sticking out atm, it seems.. and I think simply making them a two-part flopping mass or something would not look good.

Re: 4-legged animation testing / rag-dolling

Posted: Wed Dec 23, 2009 6:28 pm
by Dragonlord
Draconisaurus wrote:
Dragonlord wrote:Most problematic will be the wings. I still have to figure out a way on how to do them best.
Yeah I noticed they were not really working yet. I think first they need to be fully rigged - that is, with the complete joints that they'd use to do a flying motion (and so a flying animation might need to be finished first). They are just kinda sticking out atm, it seems.. and I think simply making them a two-part flopping mass or something would not look good.
The wings are fully rigged down to wing web bowing/shearing. So they are fully animate-able and have animations. That's actually not the problem. It's the wing web itself. The wing fingers I can do using collision shapes. The wing web though alters shape while the wing fingers move as it counts as a cloth material. The Bullet Physics library has support for cloth physics but I'm not sure yet on how to mix this all together. Kinda delicate problem.

Re: 4-legged animation testing / rag-dolling

Posted: Thu Dec 24, 2009 11:36 pm
by Draconisaurus
Hmmm, yeah.. As I think of it, I am picturing an imaginary string that connects two wing support bones, where nothing is allowed to pass the 3-point plane between them and the joint.

Re: 4-legged animation testing / rag-dolling

Posted: Fri Dec 25, 2009 1:43 am
by Dragonlord
Am currently experimenting with spring constraints for the wing web. It's not bad but collision shapes are rather small causing tunneling problems. But the wings get some kind of movement which looks okay. Just have to fully rig it.

Re: 4-legged animation testing / rag-dolling

Posted: Fri Dec 25, 2009 5:13 pm
by Dragonlord
This spring-wing stuff if actually working... sort of :D . A bit of playing around with 0-gravity. Yielded a couple of bizarre poses. Rag-dolls can be fun :P . Simple put gravity to 0 and whacked him a bit waiting until the auto-disabling kicks in and he gets' frozen in time.

Image 1 | Image 2 | Image 3 | Image 4

Need to make some video sometimes soon but first I need to fix some problems.

Re: 4-legged animation testing / rag-dolling

Posted: Fri Dec 25, 2009 6:56 pm
by Draconisaurus
Haha that dragon looks like he's having fun :lol: I must say, it's a great skin he's got.

Re: 4-legged animation testing / rag-dolling

Posted: Fri Dec 25, 2009 8:32 pm
by Dragonlord
Draconisaurus wrote:Haha that dragon looks like he's having fun :lol: I must say, it's a great skin he's got.
Thanks. The skin has meant to be realistic ( in contrary to what you'll usually find ). Try to find him: http://lordofdragons.deviantart.com/art ... n-86884952 ( full view helps ).

Re: 4-legged animation testing / rag-dolling

Posted: Fri Dec 25, 2009 11:23 pm
by Draconisaurus
Pretty nice o-0 If you'd give him more blur in that image, he might be almost impossible to catch.