Wow, that's pretty much the first full project I've run from start to end through the blender 2.5 pipeline. Its not bugless yet, but its getting there. Automerging keyframes still seems to have a few problems on the animation front, and I encountered one or two armature bugs along the way - all thankfully now fixed in the SVN builds. Hair sim seems a little crazy at times - It was adding a humongous amount to my render times in the calculating vector pass (probably a bug?) and getting blender to render from the same particle cache which is on display in the viewport takes some persuading. Much fewer crashes than I was used to last time I tried to render a complex scene in 2.5 but I still had to do the late night babysitting to restart the render when I hit an occasional glitch. Hopefully I'll get some time to do some bug reporting over the weekend.
Thanks to some help from Jahka on #blendercoders I've now fathomed what the cache buttons do and a few other options relating to hair dynamics:
Cache Step: This affects how often results are saved and as a result the accuracy of the played back simulation cache/bake. When you run a simulation each frame (and any subframe steps if you have them enabled) gets calculated, but if your cache step is set at 10 then only every tenth frame is actually saved for playback.
Quick Cache: With this on the cache step is used for the simulation. If cache step is at 10 then only every 10th frame is simulated and only every tenth frame is saved.
Effectors and Hair Dynamics: If you have hair dynamics off then effectors (objects with force fields etc) get used to shape/twist/curve the hair in all sorts of interesting and exciting ways. You can use empties with forces on (on the same layer) to fake the effect of the wind/follow-though etc or curve guides to change the hair style. If you switch on hair dynamics suddenly all these forces have a different meaning. They are now used to effect the dynamic simulation of the hair rather than to effect its grown shape.
Showing posts with label mood change walk. Show all posts
Showing posts with label mood change walk. Show all posts
Friday, 10 December 2010
Saturday, 4 December 2010
Opening Shot Render Test
Set for the Mood Change Walk: Going for a 'Night on Earth' Feel.
Still got some AA problems - will have to play with texture settings.
Labels:
approximate AO,
blender,
camera move,
indirect light,
mood change walk
Thursday, 18 November 2010
Disco Guy - Character for the mood change walk exercise
Disco Guy: Rig and Wire
Some of the texture painting's not perfect but in the end I was pretty happy with how the rig turned out. I have IK/FK switching on arms and legs, free/fixed hip choice, stretchy IK body with volume preservation and a nice B-bone torso, as well as clavicle deformers to keep the chest shape. On the hands I have palm fan controls and finger spread/fist controls as well as individual finger curls. Not too bad for half a days rigging and weighting.
Shape keys yet to come, only the eye controls are done for the face so far. Not sure if I'll bother with custom bones as I know the rig inside out and spent a while sorting out layers. Sadly the 32-bit PC in the studio can't cope with deep shadow buffers and hair without running out of RAM very quickly.
Friday, 5 November 2010
Mood Change Walk
First attempt at a mood change walk with the Ludwig Rig (and some special FK/IK adaptations) in blender 2.5. Background flies are boids with a vortex field.
Talking of boids I've been working on some python code for boid like crowd simulations, the difference being they're slightly less flock like, constrained to a 'ground' geometry and they can't collide or pass through each other.
Talking of boids I've been working on some python code for boid like crowd simulations, the difference being they're slightly less flock like, constrained to a 'ground' geometry and they can't collide or pass through each other.
Subscribe to:
Posts (Atom)

