I’m working on MikuMikuDance viewer.
GLSL toon shade
Inverse Kinematics (CCD)
btGeneric6DofSpringConstraint support(ported from Bullet Physics 2.76)
WOW, nice work!!! Rigid body rig is really cool. As as i understand, this program will be useful for motion capture, animation viewing, model viewing, glsl rendering. And it could be used in cartoon video production.?
I did not find a license of your program on your wiki, so i dont know if it is GPL or not. I suppose it’s not.
I just would like to ask if you can to share with your toon shader with our community…
I contributed some of my shaders to JME3, because i think that the community does not have good shader prefabs yet…
If you are able to share, so I can easily add your shader to my repository. http://code.google.com/p/jme-glsl-shaders/
Also, I can add you as a member to my repository for future shader improvements.
WOW! That’s really great. Thank you for sharing.
Can I add your shader to my repository? I can add any link or trademark you want. If yes, so where can I browse your shader code?
Or you can post pixel and vertex shaders here.
> Can I add your shader to my repository?
> I can add any link or trademark you want
Please link my blog.
> If yes, so where can I browse your shader code?
> Why did you “port” that? Its available in the jME3 bullet system. com.jme3.bullet.joints.SixDofJoint
JBullet supports btGeneric6DofConstraint , but JBullet doesn’t support btGeneric6DofSpringConstraint yet.
In other words, GLSL skinning can save fps if there are many animations in the scene, right? It’s interesting what kind of skinning preferred in cryengine, source engine… at present.
On the other hand, there will be in two times more shaders… with skinning and without it.
Or possibly, GLSL skinning could be better for the bullet physics calculations? I mean about bone rigged body, ragdoll…
Many modern setups have souped up graphics cards but slooooow buses. In this case, minimizing the mesh transfer is preferable.
For example, in Mythruna, some machines seem to get really high and consistent frame rates, but whenever they turn around it can take seconds (yes seconds) to send the new unculled data to the card. These machines would really like hardware skinning.
By my experience:
GeForce Go 6150
GLSL skinning works, however it’s very slow and instability.
GLSL skinning works.
GLSL skinning works fast.
GLSL skinning works very fast.
All vertex data transferred only once per running application.
But matrix data transferred per frame.
All vertex data transferred per frame.
Uniform variable capacity is very small.
uniform mat4 m_BoneMatrices;
This code works on GeForce9600GT, but it doesn’t work on GeForce Go 6150.
GLSL skinning doesn’t work with MikuMikuDance models because bone size of MikuMikuDance is over than 200.
I divided mesh per 20 bones and at last, it works.
I tried Vertex Texture Fetch like this:
However, it’s too hard for me.
You must be logged in to reply to this topic.