Task #1: Character Turntables
For our presentation this week, I wanted to make sure our characters were presented in the best way possible, so I created turntables of both the close-up facial details and full body wireframes. When I started animating the characters, however, I soon realized that a lot of the dynamic properties I tweaked to sculpt the hair did not translate well into animation. After fighting with the hair and fur systems for a while (the characters’ eyebrows kept flying away and Dyer’s hair completely exploded), I realized I had to backtrack a LOT and redo most of the hair. Sigh.
This is what I started with, and when I began to animate it only got worse from there. 😦
The good news: It all works, and is completely animatable! This took muuuuch longer than I budgeted though, so I had to table a few of my smaller deliverables. I was originally intending to texture the clothing modeled by Dave Maguire, but I decided it would be better to have some REALLY polished head turntables instead of half-assed clothing. I tweaked Dyer’s model to give him more pronounced eyelids, modeled eyelashes, painted additional texture maps (for more detail), and rendered it all out.
- Positive Stuff: I got some really positive feedback! Everybody seemed to love my changes, and the turntables rendered without any SSS flickering. And, perhaps most importantly, even the extreme close-ups rendered at only 3 minutes per frame.
- Problems / issues:
- Danforth’s fluffy hat texture took too long to render, so it was left out of the renders. It’ll be added in when it’s optimized a tad better.
- I didn’t re-UV Dyer after modifying his model, so his wireframes have seams down the center axis. I’ve already fixed this problem, but I didn’t fix the renders.
- Mustache could be a little fuller.
- No clothes… yet!
Hours: 15 😦 😦 😦
Task #2: The Presentation
I created the Keynote for our end-of-fall progress report presentation. I used the same template as our pitch, so thankfully putting everything together didn’t take too long. The assets themselves took a while, though — because so many of our deliverables include motion, I edited together a few videos from everyone’s assets and embedded them in the presentation.
- Positive Stuff: I’m really, really happy with all the progress everyone has made! Looking forward to showing it all off tomorrow. 🙂
- Problems / issues: None so far, but I suppose we’ll find out tomorrow what the faculty thinks.
LIST OF UPCOMING TASKS:
Before the end of the term, I’d like to have the following completed:
- Textured clothing
- Tweaked Dyer model (his eyes are slightly too feminine for my tastes)
- All textures, shaders, and character details DONE and ready to be imported into the rig
Summary: This week I continued my work in learning Houdini. The main issue that I was able to over come was exporting FBX files to Houdini correctly. I also figured out how to export alembic files to Houdini.
Pros: I actually exported a full scene correctly from Maya to Houdini.
Cons: Alembic and FBX in Houdini both have their strength and weaknesses.
Summary: Nothing too new this week for rigging. I finished up dyer. I added all the finger controls/ sdks. I fixed the ikfk blending. I also added in eye controls.
Pros: At this point the rigs are finished! Expect fixing some weighting issues and any other issues.
Cons: None other that hand sdks take a long time to make.
For next week:
Prepare for the presentation.
Move onto the face rigging.
This week I fell a little behind because I went into hospital in the beginngig of the week for a back injury.
Summary: To this point I finished all that need to be put into danforth for the body rigging. Unless anything else comes up, that being rig issues, the rig breaking, and helping gabe with the weighting. This week I finished all the SDKs for the hands. I switched the IK FK switch so they can be blended between one and zero. I rigged the hat. The hat still needs a little work on the weighting. Finally I added controls for the eyes.
Pros: Danforth is very close for facial rigging.
Cons: Need to catch up and finish dyers extra rigging.
Summary. I continued my weekly study of learning Houdini. This is mainly for advanced animation but anything I learn in Houdini I believe will help in the future when I begin working on effects.
Pros: I am slowly starting to understand different parts of Houdini. Excited to try implement. Houdini in the project
Cons: Still a lot of Houdini to learn.
Being that I was not getting great weighting from last week and to save time gabe decided to help out this week with rigging. The way things worked out is that I worked on base weights for both danforth and dyer. I got all the weights to an alright state, then gabe went in and did all the fine tuneing. I am really happy with the result of the weighting that gabe did. The weights will still change as the animation progresses. Gabe also added in a extra joint to help with the squash and stretch in the body.
Along with weighting I worked on some rigging fixes. The animators found some issues with the rigs that I went in and changed. The setup for the neck control was not as good as I hoped it was going to be. Therefore I went and changed the neck back to a regular Fk neck setup. I also fixed the fk controls in the arms. I set up the controls up so the clavicle controls are on the same rotation axis as the fk controls. While gabe was weighting he saw some joint placement errors, so I fixed those issues as well.
Finally I worked on hands. It took me some time to find a good setup for the hands. I was able to set up the hands in a way that will let me go in and add in extra control in the future. As well I avoided using sdks and used the connection editor, which will also save time in the future. The rest of the hands will be finished tomorrow.
Pros – It was great working together with gabe, he helped a lot. I am excited to start the face rigging.
Cons – Some of the issues in the rigs could of been prevented.
This week I continued learning Houdini. I worked on learning the smoke solver in Houdini and attached smoke to broken up geometry. I was able to get some pretty good results with fast sim times compared to maya
Pros – Houdini is fun to learn.
Cons – Figuring out how to incorporate Houdini into everything with the project.
This week I started working on the weights for dyer. I was not able to work as much this week because I had some issues at home to deal with.
Summary: I spent a lot of time painting the weights for dyer and did not get too far. The tight edge loops around the legs is giving me trouble to get good deformations out of the legs. I was going to make corrective blendshapes, but I want some what good deformations for making the blendshapes. After awhile I got something working with dyer.
Pros: Nothing at the moment
Cons: The weight painting was not working well
Summary: I was able to make clouds in Houdini. They seem like they will be a good option for some of the shots in the animation.
Pros: The clouds look really good.
Cons: Long render times.
For this week I rigged more, and practiced n-cloth.
- Summary. Well rigging had its ups and downs this week. Danforths file crashed on me again which meant I had to re rig a lot of the upper body. But doing it for the third time was much faster than the second and first time. From there I started building the proxy rig for Danforth. Once again Danforths file crashed on me multiple times. Luckily I saved often. I gave the file to the animators to see if they could crash the file. None of them had a crash. I am not sure what exactly is going on, i might just re install maya to see if it will fix the issue. The proxy rig geo was finished for danforth. I currently am working on dyers proxy rig as I write this post. Aside from crashes I spent some time on getting the hierarchy to work with my Fk and Ik arms. As well I cleaned up the my file names/ structure for my own sanity.
- Pros. The animators can start animating. Unless they break the rig.
- Cons. The weird issue with dyers file constantly crashing. Also that fixing the issue most likely is a simple issue.
- I worked on cloth for awhile. I started with testing marvelous designer. It is a great program for layout and design of cloth. I could not get any good results when trying to bake a cloth animation out of marvelous. For sake of time it might not be worth looking into a solution for baking cloth from marvelous. I was going to attempt cloth animation for characters, but before I attempted it I talked to Dan Bodenstein. He explained to me that there is no point to do cloth for characters in Houdini and I should stick with ncloth for clothing simulations. At first ncloth was really hard to get a hold of but over time I understood it to point. It is a lot of trial and error. In the end I was able to get marginally alright results.
- Pro. I learned a lot about N-cloth in a short period of time. I found it to be a lot more powerful than I thought. Everyone this week practiced n-cloth.
- Cons. Ncloth is hard. Cloth exploded a lot. Cloth exploding then maya crashing. Maya crashing.
Various simulation testing.
Continue learning Houdini.