OPTION 1) Delete the Human IK rig (leaving me with a skeleton with no controls or Ik systems, but skinned clothing), and reconstruct the IK solutions and manipulators. This would be re rigging the feet, legs, hips, clavicles, arms, forearms, spine, neck and head. This would be a lengthy job and would need me to essentially reproduce the work I did for my last piece of work which took a number of weeks.
OPTION 2) Get the original Maya rig / mesh (which is already rigged and has controls etc) and import the clothing meshes and re-skinning them. This would also be a lengthy option, but the rig skeleton and controls would already be in place.
I decided as I didn't initially have a reply from the CG Society, to go for option 2. This would require me to 'pull an all nighter' and took from 6pm Tuesday the 12th, until 7am Wednesday the 13th. This may seem of of been a little crazy, but as the Human IK rig is different to that of the original rig, I couldn't import the skinning data of the clothing and so had to re-skin them. This did however give me a good opportunity to improve the original rig (constructed for an NSD module). Improvements other than the skinning of the clothing included:
- Spine joints 2,3 and 4 were re-skinned as influence covered too much of the torso.
- Hips upper and lower legs also re-skinned as too much pull was occurring in the hips area, effecting the skinning quality of the trousers / shorts.
- Advanced eye facial control nodes have had re targeted with new set driven keys (so now the eye lids open and close properly when the mesh is in smooth mode)
- All clothing meshes and the rig mesh have been given the custom 'Mesh Smooth' attribute for quicker animating and test rendering.
- The clavicle rig control's joint influence has been relocated closer to the neck, giving the manipulation a more natural look.
The rig now looks the same as the Human IK rig as al the same clothing models and rig meshes have been used. I feel that this was a smart move as the next day, I had the thread conversation with Shawn about both the Graph Editor problem, as well as the Human IK's uses. He said that although the Human IK can be keyframed, it is primarily used for Motion Capture and re-targeting. He said that any useful keyed animation would have to be on additional animated layers, increasing the workload 3 to 4 times (which I simply haven't the time to do).
Therefore, I have made the executive decision to abandon the Human IK rig and use the appropriately dubbed 'Contingency rig'. I am a lot more comfortable with this rig and feel that I will be able to get good results out of this rig in a shorter space of time. In addition, the stepped tangent function in the graph editor in much more accurate as there isn't a full body IK solver constantly trying to correct itself.
It is a shame that I will not be able to use the Human IK rig in the way that I wanted to, however the way I see it, I now have a rig that I can plot motion capture data on and possibly use this for a future project. In addition, I will still be using other tools such as the Trax Editor, as well various 'Character sets and sub sets, the Graph Editor and hopefully the Sequencer.
So the project has gone as follows:
- Dee character rig to be used but not original enough
- My custom Maya rig to be used, but want to apply Human IK (FBIK)
- Custom rig has IKs attached so cannot be used, so made new rig
- Skeleton generator used and rebuilt rig, then applied Human IK
- Initial tests were successful, but cumbersome to use and Graph Editor / usage problem
- Fell back on original rig, but reskinned clothing and tweaked skinning / controls.
So the outcome is that I have now got a rig for motion capture data re-targeting, as well as a new and improved custom rig.
At this stage, I do not plan on using the Human IK rig as this project is more about key-framing as extensive use of the Graph / Trax Editor, but it would be good to at least experiment with motion capture data, so we shall see. The project is now currently back on track.
No comments:
Post a Comment