The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
Rank posts some new animation process work
Rankenphile
Passersby were amazedby the unusually large amounts of blood.Registered User, Moderator Mod Emeritus
Passersby were amazedby the unusually large amounts of blood.Registered User, Moderator Mod Emeritus
For my second semester Junior Project, we keep a weekly progress journal defining our progress and documenting our process. I was keeping it in a word doc, but decided to move it to my blog to share the progress of the project. To get feedback form the community here and to share the process with those that may be curious, I'll copy the updates to the forum as well. I'll be updating this at least once a week until the project is done.
For the class, we have to create an animated short from scratch, using Maya, Max and other software. We start by building out a story and storyboarding it out, analyzing the beats and making sure everything works while we build out character analysis and art, then model, rig, skin and texture the character or characters, build and light the scene and then animate the entire thing. Depending on our focus, our scene can be between 15 and 45 seconds.
Below is the first entry for the project, detailing my story, breaking it down to a Logline, and then the first pass at analyzing the beats of the story itself.
Shot opens on a moonlit roadside exterior, somewhere in a thin forest. An ancient sword sticks out of an anvil at the side of the road, an old twisted fence surrounding it. Grass has grown high after years of neglect. Clearly no man has been able to free the sword from the anvil in a very, very long time.
A small goblin coming down the road appears in the left of the shot and, just as he's passing the sword, spots it. He stops to examine the object and gets nearer, getting a good look at it. As he touches it, a beam of light illuminates him. Clearly he is destined by the Gods to be the next great leader of men. He decides to try his brawn and grabs it by the grip, jerking it upward. It doesn't budge. He tries again, both hands on the grip, and it refuses to budge. Grabbing it by the cross guard, he strains and bends and jerks the sword as hard as he possibly can. Finally it comes free and he falls to the ground, the sword in his lap.
Coming to his senses on the ground, the goblin shakes his head, then realizes what he's done. He leaps to his feet, elated, and tosses the sword away into the grass. The beam of light immediately fades as he runs over and grabs his new-found anvil and drags it off screen right.
A goblin, finding a sword in an anvil, pulls the sword out but discards it for the anvil itself.
Opening shot: sword stuck in an anvil beside a road under moonlight. Goblin enters screen left.
For the class, we have to create an animated short from scratch, using Maya, Max and other software. We start by building out a story and storyboarding it out, analyzing the beats and making sure everything works while we build out character analysis and art, then model, rig, skin and texture the character or characters, build and light the scene and then animate the entire thing. Depending on our focus, our scene can be between 15 and 45 seconds.
Below is the first entry for the project, detailing my story, breaking it down to a Logline, and then the first pass at analyzing the beats of the story itself.
Story – Pitch and Beat Analysis
Shot opens on a moonlit roadside exterior, somewhere in a thin forest. An ancient sword sticks out of an anvil at the side of the road, an old twisted fence surrounding it. Grass has grown high after years of neglect. Clearly no man has been able to free the sword from the anvil in a very, very long time.
A small goblin coming down the road appears in the left of the shot and, just as he's passing the sword, spots it. He stops to examine the object and gets nearer, getting a good look at it. As he touches it, a beam of light illuminates him. Clearly he is destined by the Gods to be the next great leader of men. He decides to try his brawn and grabs it by the grip, jerking it upward. It doesn't budge. He tries again, both hands on the grip, and it refuses to budge. Grabbing it by the cross guard, he strains and bends and jerks the sword as hard as he possibly can. Finally it comes free and he falls to the ground, the sword in his lap.
Coming to his senses on the ground, the goblin shakes his head, then realizes what he's done. He leaps to his feet, elated, and tosses the sword away into the grass. The beam of light immediately fades as he runs over and grabs his new-found anvil and drags it off screen right.
Logline (story pitch in 25 words or less)
A goblin, finding a sword in an anvil, pulls the sword out but discards it for the anvil itself.
Story Beats
Opening shot: sword stuck in an anvil beside a road under moonlight. Goblin enters screen left.
- Goblin notices sword.
- Goblin approaches cautiously, examining sword in anvil.
- Goblin touches it tentatively. Shaft of light appears above him.
- Goblin recoils from anvil, cautiously.
- Goblin touches again.
- Goblin decides it is safe and grabs the handle, jerking upward swiftly. Sword doesn't budge.
- Goblin grabs with both hands and pulls hard but sword doesn't budge.
- Goblin lets go of sword, then musters his strength.
- Goblin grabs sword by the cross guard, steps onto anvil and pulls with all his might.
- After a long pull, goblin frees the sword from the anvil. Resulting forces goblin to reel back and fall onto his butt with sword across his lap.
- Goblin shakes off the resulting confusion, notices sword in lap, and jumps to feet.
- Goblin pitches sword off-camera into the weeds and approaches anvil, excited.
- Goblin grabs the anvil and, straining, lifts it, dragging it off screen right.
Rankenphile on
0
Posts
I initially designed the character after the Warhammer-styled goblins, going for sort of a mean, tough and realistic look, trying to develop a highly detailed sort of character the type of which you might see in a video game.
I developed the first iteration of this character after going through a few different initial sketches, then spent time painting it up in photoshop as a greyscale rendered sketch to figure out the dimensionality of the forms and the contrast needed to make each area readable easily. Unfortunately, I just wasn't drawn to the character as much as I'd like, and once I saw what some of the other students were working on, I went back to the drawing board with a new inspiration, wanting to build a character with more appeal and interest in the design.
After a number of sketches, I hit on the following design, which I inked down a turn-around and build a color comp layout, utilizing a number of color combinations to figure out what might look the best in the scene.
I really like the new character and it feels like a big break from work I've done in the past and appeals to my sensibility of animation and design. I think this will help hold my attention and interest a lot more.
I've shown the color comp to a number of fellow students, professors and friends that aren't trained artists to get a wide range of opinions. I have been getting a lot of interest and reaction to numbers 5, 6, 9, 11 and 12. I think, personally, I'm leaning toward 5 and 6, but I want to give it a few more days to gather up more opinions before making any final decision. Creating textures for the character is still a long ways away, anyway.
Using the cooler, fairly monochromatic colors of a forest under moonlight, I'll be able to use warmer and more saturated colors on the character and the anvil, as well as higher contrast on the character and “hero prop”, to help draw the eyes onto the story elements of the shot. Using the trees to segment two frames of sky, one on either side of the screen, I can build two distinct stages for the action – one for the goblin to enter to and establish his motivation and character, and a second to frame the hero prop and story moments.
I'll be drawing on my experience in last semester's CG300 class to model and light the scene. Examples of those renders are below.
01 – The goblin enters the screen from Stage Left. Seeing a sword stuck in an anvil, he gets naturally curious.
02 – As he reaches for the sword, a shaft of light appears from the heavens. He tentatively accepts this new fact and begins to struggle with pulling the sword free.
03 – The goblin, struggling with all his might, pulls and tugs at the sword, stuck fast in the anvil.
04 – Finally, despite his small stature, he manages to tug the blade free. The resulting momentum, however, propels him backward onto his butt on the roadside.
05 – The goblin shakes his head and comes to his senses. Realizing his situation, he beams at his luck, and then, to our astonishment, pitches the new-found blade into the bushes.
06 – The goblin's true intentions are revealed as he grabs the heavy anvil and struggles to carry it off-screen Stage Right.
The head stands at 586 quads (1,172 tris) right now, a very respectable low-poly count for a cinematic like this. I spent a lot of time studying edge loop topology to make sure I'd be able to have a lot of good deformation in the face when it comes time to animate. I bought a couple of books, one on facial animation and rigging and another book on how to set up a character rig in Maya.
My plan for my workflow right now is to build out the models in 3DSMax, then import the models into Maya to rig and skin it. I much prefer the rigging setup in Maya, as well as the animation and lighting tools.
(The following was created a day or so after the post above and was merged with this post as they cover the same topic)
I've finished much of the modeling for the character, having roughed in much of the costume and completed the limbs and extremities. I managed to complete the model without using only quads, which will help subdivision and sculpting for normal maps in Zbrush, if time allows.
I want to have an advisor take a look at my topology before advancing much further to ensure that the deformation will work as I hope it will. I still have yet to model the interior of the mouth or the tongue, but that should be rather simple. Once the design is finalized, I can begin making morph deformation targets for the face to help with the facial animation. I'll be sketching out the needed poses for the face before doing the target poses, and will be sure to work with an advisor to ensure that this is done correctly and wisely.
Not the most productive week so far, due to considerations for other classes – both CG215 and ANI350 hit very hard with their demands for work due this week, so much of my time was spent on those classes.
Prof. Johnson expressed his concern about the story, in that the technical requirements for the animation were substantial – showing weight and transference of weight is very, very difficult stuff, and he felt that ending the story at the point where the goblin gets the sword was a good “if all else fails” plan.
In a way, I agree that it would be a good failsafe, but I'm confident in my ability to take on this challenge, providing I don't run into major disastrous issues on the technical side as I did last semester. I am really trying to put this piece together to showcase my ability to portray both acting skill through animation as well as raw, technical animation skill, using the techniques learned in ANI300 and those I will be aquiring in ANI350. Additionally, having gone through the process of early vid ref, I think that the needed poses aren't horrendous, and the entire scene plays out in approximately 30 seconds, which is a manageable goal, in my opinion.
Having looked at the technical requirements for what I need to do, I have planned for the following:
I also received feedback from other students and faculty about my general topology. For the most part, I'm in good shape, but I need to revise the belly a little to provide the most flexibility, and I will need to reconfigure some of the topology of the shoulders by re-routing the edge loops, always a tricky process. So far the goblin is modeled completely in quads, resulting in great subdivision. I'd like to keep it that way.
So what happens to the beam of light when the goblin pulls out the sword and falls down with it in hand? Does it follow the goblin or the sword? Your thumbnail makes it seem like it follows the sword as he throws it away.
Maybe if it follows the goblin, indicating that he's the new king of men, it should stay on him as he drags the anvil away? I mean, just because he ditched the sword doesn't mean he isn't a king now
"You are a child of the universe, no less than the trees and the stars; you have a right to be here.
And whether or not it is clear to you, no doubt the universe is unfolding as it should."
I'm planning on having it follow the sword and begin dimming the second it leaves his hand. Keeping it on him infers that he would be the king of men, but that's a bigger story beat to try to explain in the limited amount of time then I think I can accomplish. It is suggested by the existence of the light, but it would need something additional to help tie that down, otherwise it is too ambiguous. Having it tied to the sword, and his contact with it, is already apparent in the actions laid down, and doesn't require further explanation.
In my previous entry, I mentioned my vid ref I did for basic acting beats and timing. I'm including it as a YouTube link below.
http://www.youtube.com/watch?v=hvJ00HYoLts
Began the week with a group critique with Pam Matheus and a number of fellow students. Notes from the critique are below.
Thoughts for resolving these issues:
I completed the final first draft of the base mesh model of the goblin today, unifying the entire model into a single mesh, with the exception of the shoulder pad itself, which will be parented to a special controller or a bone. Should work just fine. Thanks to Josh Jones, I resolved the issue of the topology around the shoulders, and added more edge loops around the belly area to allow better deformation when the character has to bend. Prof. Lu also suggested removing a couple unneeded edge loops to create better, more even quads, and that made good sense.
I deleted the body mesh that would be covered by his shirt/tunic and extruded the mesh out to match the geometry of the shirt. This went swimmingly well – I originally created the shirt by duplicating that part of his body, using the push modifier to increase its size and then pushing and pulling verts to get it into the right shape, which meant that it shared the same number of verts and edge loops as the mesh below, making it a trivial task to attach it later and patch up any potential holes. I used the same extrusion method to create the belt, and may add a few more edge loops to the edges of the belt if I choose to take the model into a subdivision sculpting program like Zbrush, but that's another issue for a later time.
Having it as a single mesh will make skinning much easier a task then I had last semester, as long as I don't run into the same technical problems, such as having the Paint Weights skinning tool crash every time I use it. I've included pictures of the progress below.
The overall model with the first draft of the final topology, before adding the eyelids. I may extrude the geometry on the right hand to create a fingerless glove, but for the time being I'm happy with the results.
Detail of the connection of the mesh to the boot at the ankle.
Detail of the connection of the mesh to the shirt at the neck.
Detail of the shirtsleeve connection to the mesh. These additions keep the entire model watertight, so that there won't be any holes to view. I also connected the underside of the shirt to the hips/legs area, but it is a tricky one to get a clear picture of.
As I delve into the technical aspects of getting the character rigged, skinned and get the deformation targets built, I'll be adding my notes and plans into this section. My primary resources for this are the books Body Language: Advanced 3D Character Rigging by Eric Allen & Kelly L. Murdock, and Stop Staring: Facial Modeling and Animation Done Right by Jason Osipa. I'll also be drawing on the knowledge and experience of faculty and fellow students, and will try to attribute my findings as I write them.
Known Needs:
I've begun talking with Prof. Chun Lu, who has extensive experience with rigging in Maya, to help plan out the methods to best approach for the needs expressed above. I'll post progress as I get it.
Thanks. That was my idea for sharing this - I read a lot of blogs and get to see a little bit of the process, and every little scrap of it I find is super valuable, as a student. I figured since I have to document it anyway for class, why not share it with the forum. I'm keeping a copy of it all on my blog, as well, and basically just sharing a copy of that content here to get critique and share it with a community I've gotten a lot from in the past.
Here's the blocking stage of my first 3d animation set to dialog. Lots to tweak and fix, but I'm fairly happy with the results so far. Hopefully I can keep the quality up as I move into splines this week.
http://www.youtube.com/watch?v=yBBwm-ougic&feature=player_embedded
I agree with adding a few touches to subtly imply hes gonna go for that anvil, But I think you can do this pretty simply by having him look at the sword like one would look at sword if it was sticking in their living room couch. Just with an an expression that goes "Sweet, check out my new anvil. How..... and why is this sword in here thats really inconvenient and I need to get this out of here and...... what the hell light? Goddamn it come on get out of there. That was more trouble than it was worth! Sweet! Anvil!"
If you can get those facial expressions of confusion and annoyance, it'll work a bit better than the child like wide eyed discovery, in my opinion.
1. The acting/video reference - I highly suggest that you redo the video using a heavy prop for the sword. Convincing feeling of weight and motion is key and is a large part of the animation. I know a sword is kind of a weird prop to make, but ask around. I know Jarcho was able to make something up for me when I needed to make a huge lever.
2. Character - Right now, I don't understand why he wants the anvil. It's a great gag, but I think if you introduced some elements to the character that made him look like a blacksmith (example like an apron or some heavy gloves) could help sell it for me. It kind of seems random and I think it'd be funnier if there was some reason behind doing it.
This is also a cinematic model - I feel you can increase the polycount on the character. Be sure to add atleast three edge loops around the elbows and knees. If you need some awesome topology reference for models, go to www.hippydrome.com
3. 3D Programs - What are you using to animate/render to? I assume you're learning Maya more and came from more of 3ds Max background from sophomore year. Maya is great for animation and I highly recommend it. Learn how to properly light and make mock up models/prototype early. It'll save you a lot of hastle towards the end.
If you need any reference or suggestions, feel free to PM me and I can give you my contact info for any other advice. Great start dude.
I've got a lot more refinement to do on the model, and was already forwarded Hippydrome by another student. I'm going to do what I can to understand those concepts more and do some deformation tests with a basic rig before going to final.
And yeah, I'm doing all my animation, rigging, lighting and rendering through Maya. I like Max's modelling workflow a little better and its UV tools seem more capable, especially for organics, but animation, lighting and rendering tools seem more intuitive and better in Maya. If I can, I'll be using the school's MusterRender farm for the final render, but time will tell on that end.
Thanks for the feedback!
Maybe
It's not part of our workflow, but that doesn't mean it isn't a tool still in use in the industry somewhere
I spent last night re-working much of the model I began for my CG315 class, in which I was modeling a crust-punk mermaid to take into ZBrush to begin subdivision sculpting. I threw out the old head and redesigned it, and am far happier with the results. I'm still not thrilled with some aspects, such as the hat, and need to connect the hands to the rest of the body and finish out the forearms - the elbow joint could use some attention, as well, but I think much of this work will be completed in ZBrush itself rather then struggling to get it done in Max working against deadline when I have other work that needs attention.
Here you can see a comparison of the old head (left) to the new one (right). It is far less flat and much more feminine looking, and fits the character design much better. The ears still need a lot of attention, and the cranium shape isn't quite right, but it is a dramatic improvement overall.
Another shot of the new head. Some of the overall issues are easier to see here, and hopefully they'll be addressed in ZBrush. Since this shot was taken, I've completed the head shape and addressed some of the issues with the hat and cranium, as well as extending the neck down below the shirt and jacket.
A look at the rest of the body mesh. Not perfect, by any stretch of the imagination, but it'll do for now. I wish I had more time, and a lot more experience, but that's how things go with a project of this scope and limited time frame.
The hand, always a difficult model to make. This is severely revised from the previous model shape, and should work well for the purposes I need it for.
I'll be taking the model into ZBrush this week and beginning my first foray into SubD sculpting. I hope I can do the character justice.
At the beginning of this week, I had started with the UV unwrap process when Mark Peasley suggested I revise some of the mesh to optimize it for deformation before going too far into the unwrap. As much as I wanted to just get the model unwrapped and start painting the texture, I knew he was right, so I made sure to put attention into the trouble spots on the mesh, which is where the majority of my attention went this week for this project.
I made sure to add additional loops around the facial features that were going to see the most deformation, as well as optimizing the joints.
I added additional loops around the ball of the foot area, as there was going to be substantial bending here and it needed to work naturally. I also added a "patch" around the knee to help keep the form from flattening oddly when it was bent.
The elbows received the same treatment. I also made sure to optimize the loops around the knuckles and wrists.
I still need to go in and add the geometry for his single glove, but that's a trivial matter of just extruding and pushing the mesh and then pushing a couple verts around to clean it up.
I've also begun testing out my first foray at rigging in Maya, and already really like the setup far more then its Max counterpart. I've got a ways to go, especially getting the IK-FK blends set up properly and the facial deformations in place, but I've begun the research and sketching for that process, as well, and should be good to go in two weeks time.
With that schedule in mind, getting the model unwrapped and a base texture down in the next week or two is a goal I can handle. Getting a block model built of the environment, the props built and the initial shader research all in the next four to five weeks is do-able, and I'll be starting my animation planning in that time frame, leaving me plenty of time to complete the project and hit all of my goals, providing I don't hit any major technical or personal snags along the way.
http://www.youtube.com/watch?v=Ru8mRXQnG48
Less progress then I'd like, but lots of technical background work, prepping the model for the next phase of development.
I spent a little bit of time re-familiarizing myself with projection mapping tools in Max, and was able to create a reasonable tiling texture of chainmail, to be used on my little goblin friend's loincloth, which will add a little fun secondary animation to the model. To do so, I created a torus, rotated it into place and then duplicated it and moved it, creating a single pair of linked rings, then duplicated that set and moved it into place. By doing this step over and over, I created a good base mesh for the chainmail.
Once I got it set up over a plane, I scaled it to the point where I was reasonably sure it would tile well, then used the Render to Texture panel to project the chainmail mesh onto the plane below, creating an alpha, normal and ambient occlusion map set.
This resulted in three 1024x1024 texture maps, one for each of the passes. The transparency map will be hidden in the alpha channel of the diffuse map, once I paint over it. The ambient occlusion pass will be used as the base color channel for a paintover to create the diffuse map, and will help inform the specular map, as well. I'll be laying over rust colors and textures to help that process, when the time comes.
To test the new setup, I imported the file into CrazyBump to ensure it tiled well, and it does, at least reasonably well. It isn't perfect, but for my purposes it'll do well.
I'll be using this method to create better normal maps for a number of the props and pieces of my environment, and using CrazyBump to help create texture maps for much of the process, as well, as it seriously helps speed up the process - making normals for pitted iron for the shoulderpad and anvil, for instance, or the bark of the trees, is a snap in CrazyBump.
I've begun unwrapping the model, and am looking at a number of tutorials and plugins to help expedite the process. There are good ones available at the classily-titled site Poopinmymouth.com. I've gone back and forth on this process a bunch of times, getting mixed results, but I'm determined to have this portion done by the end of the weekend.
Finally, I've also begun rigging the goblin in Maya, and after speaking with a number of fellow students I'm moving forward with the process.
I've got the skeleton itself in place and have begun basic tests for the skinning showing a lot of progress - the joints seem to deform quite well already without having gone into the weighting refinement process, which means that completing the skinning should go pretty well once the rig is complete.
I've started the process of setting up three arm bone sets so that I can have a switch between IK and FK-controlled arms, which will be crucial to the animation I have planned. I've found a good tutorial here that should help the process.
I've also learned how to set up additional attributes and link attributes in Maya to controllers or controller sets, so I can do things like have a single controller for all of the fingers on one hand, for example. This will greatly ease the animation process, and is good experience to get under my belt.
I'm going to keep pushing forward on this stuff all weekend, and I have an appointment for later this coming week with Micah Zahm to help fix any issues I can't resolve myself. All in all, I should be through with this technical hurdle soon, which is exciting - this is what nearly sank my project last semester, and I've learned that it is best to go into something this technical that I'm not particularly strong at by planning carefully and seeking help from the folks that know it well, rather then keep bashing my head against the wall and hoping for the best.
So I spent a bunch more time this weekend getting the model unwrapped and prepped for texturing. I found a great rainbow and numbered checker pattern so I can be sure I'm distributing the UVs where I need them, and that I'm laying them out in the orientation I want. The unwrap went pretty smoothly, with only a few parts that needed to be overlapped - the legs, arms, ears and eyeballs all share the same UV space, since they can be textured identically. All other features, including hands, the head, the feet and the costume, are textured individually.
The only parts I'll have to be cautious about in terms of seams are the cheeks and the wrists, for the most part. I had to make cuts there to get the mesh to flatten out without too much distortion, but painting over the seams carefully in photoshop by directly importing the model should help create a fairly seamless paint.
Once I had the UV layouts arranged, I went in and created a flat-color flood, to test out the model and be sure it looked okay.
I then applied the texture to the model. Success. So far, things were coming together, and it was exciting to see him with some color on him - it meant I'm that much closer toward my goal.
I then applied an Ambient Occlusion diffuse map to the model and set it up with a number of directional lights to eliminate hard shadows, so that I could render out a burned AO texture pass.
Once rendered, it created a basic shadowed and flat version of the model texture, which I could overlay on the original flood fill to help inform the shape and give the illusion of light hitting the model. This will greatly speed up the texturing process, and give a much better overall look.
The new texture was then applied to the model, and the components were re-separated from the mesh. I'm now having to step away from the project for a bit, as I've been neglecting my other work to get this far on it, but I'm glad to see so much progress and I'm looking forward to getting to paint him up.
Anyway, huge update today.
Huge amounts of progress this week. I spent a lot of time learning how to use Subsurface Scattering Shaders in Max and Maya, as well as getting my rig prepped and ready for animation and skinning.
The detail and effect of SSS was astonishing, as it was my first chance to play with it and see the result. Even better, it really didn't add to the render time to any major degree. I burned off another closeup render to use as a target for the Maya shaders and got to work in the other program.
The Maya shader system, being node-based, offered a lot more control but was considerably trickier to get fully functioning. I found a wonderful set of free video tutorials over on Gnomon's site and went through them, taking a couple hours to really make sure I understood the concepts - longer then I hoped to spend, but I think it will be far worth it in terms of quality of the final product gained for the investment.
First I got the basic shader set up and began with the normal map, making sure it was still going to function properly, as the added simulated contour would help inform how the shapes were going to look in the renders, and it would tune the basic three-point lighting system I had set up.
I then began tweaking the different shader settings, beginning with the back scattering.
Once I got this blocked in, I recorded the settings, turned it off and blocked in the Epidermal Scattering layer.
Again, once this was blocked in, I turned it off and began the Subdermal layer.
I then began turning them back on, first adding a color map to the backscatter layer.
Then I brought the color map to the other layers and spent time fine-tuning the weights between them.
As you can see, it really knocked back the effectiveness of the shader, especially the normal map, but much of it is due to the fact that the different shader layers need to be painted more to get the most out of this. Also, at this point I didn't have the specular layer applied. I brought in the specular map I had created before and set it up in a rudimentary fashion, and then had to call it a day for this part of the process, as I had an appointment to work on the rigging.
I'll be going through the tutorial more to learn to fine-tune the specular layer, and will be following another tutorial on setting up a shader for the eyes separately from the rest of the mesh. I may also set up a new shader for the shoulder pad and the metal of the loin cloth and belt buckle, and possibly the cloth of the old sack he's wearing. The overall paint still has a lot of work to do, adding details and defining the shapes and contrast, but for a lay-in test this is working very well, in my opinion.
Megan Noble was quite helpful earlier this week, showing me how to create new attributes for controllers, how to orient joints and a number of tricks when it came to setting up the skeleton, as well as just answering questions when I'd get stuck or confused.
Earlier today, I spent a couple hours with Micah Zahm, who was gracious enough to give a couple hours tutoring me on the process of setting up a rig the correct way in Maya, teaching me about simple things I was doing wrong that can break a rig, and the best ways to do it right. It was incredibly helpful, as rigging has always been my downfall before, and he showed me a number of things I found very enlightening about the process.
Already, I'm finding Maya a much better system for all of this stuff then Max, at least in my limited knowledge. I was able to set up a full skeleton pretty quickly, and actually did it a few times before arriving at the final design.
I have a whole suite of features built into the design of the rig that I needed in order to have the most success in the limited amount of action I'm planning for the character. Ideally I'd like to use the character for other things later, but for the time being and my current needs, he's going to work out great, I think.
Micah was very helpful in showing me how to set up a reverse foot rig setup, and helped explain exactly how it works, which is something I've never been able to comprehend precisely before.
Essentially, the bone at the bottom of the foot doesn't get skinned to the mesh, it sits along the ground to allow the rig to pivot on its heel, but still allow the model to rise up on its toes or rock on the ball of the foot without lifting the foot controller but still moving the ankle IK, making the knee follow, as well.
You may also notice he's got a "wiener bone", as it was jokingly named. That is a series of bones created to be skinned to the loin cloth mesh only, allowing for secondary motion as it flaps around when he walks.
I've also set up a pair of bones for the eyes, which have Aim Constraints for the controllers which float out front.
The eyeball mesh is skinned entirely to those joints, and the joints themselves are parented to the head joint to make them follow in space. This lets the eyes rotate around inside the head without moving, but they follow the head as it moves. Just below the head joint is a joint for the jaw. I'm going to build two sets of teeth - one for the top row, which will be parented to the head joint, and one for the bottom row, which will be parented to the jaw joint. This will allow him to open his mouth and have the teeth follow. Finally, I have a pair of joints set up for the ears, allowing a lot of expression and secondary action in them.
I also spent a lot of time getting the hands set up to a controller. The hands themselves are IK controlled, and are moved by the wrist controller. The wrist joint itself has an orientation constraint to the controller, which allows the arms to be moved by grabbing and pulling them around, but keeping the hands oriented to the world, as they should be. This will be a little trickier in the animation, as it can be hard to maintain proper arcs with IK rigs, but they're necessary to lock down the hands in the world space when putting any weight or grabbing onto objects.
I also added new attributes to the hand controller and mapped them to the fingers, so that instead of grabbing controllers on every little joint I can manipulate them from a single space - much, much easier to work with and animate this way.
I still have a bit more to do to finish the controllers, and have a LOT of work to get the deformer targets built for all the expressions he needs (blinks, scowl, wonderment, confusion, grimace, grin, elation, surprise, etc), but I'm getting much closer, and am moving through a step that was previously daunting, thanks to the help of some very kind folks.
http://www.youtube.com/watch?v=2qHlpbiPPIo
I spent the first part completing the rig, learning the ins and outs of how to set up a biped rig in Maya. This is a very tricky, very technical process, and definitely not my specialty, but it is done now, and very functional.
I've got a lot of features built into the rig shown above - both arms and legs are IK driven, and I have a reverse-foot setup, allowing me to keep my heels locked on the ground, roll onto the balls of the feet and do all sorts of other great functions. I have a fully articulated spine, a jaw bone for controlling limited facial animation, duo-jointed ears, bones for the eyes, and a series of bones set up to allow secondary animation on the loincloth.
The eyeballs and eyelids themselves are set up on the eye controllers, and I created a locator object above the character's head to allow him to tip his head to the side without causing odd pivoting or rotation action in the eyes or eyelids.
All eye activity is controlled through a series of controllers, using the Set Driven Keys commands in Maya. Through this technique, I can control blinking, eye direction and eyelid rotation, all through a single controller. This functionality was used on the hand controller, as well, allowing me to grab a single object to control the rotation of all finger joints, the spread of the fingers and limited rotation of the palm for exaggerating the form.
When I was rigging one of the hands, I came across a strange problem - all of the translate options for one of the finger joints was just... missing. It appears that somehow the transform node was, at some point, removed, leaving the only available keyframe-able option as Visibility, not a feature I need a lot in animation. The solution for the problem ended up being rather clever - I created an empty group object, basically a null object in Maya that doesn't have any actual data or anything attached to it beyond transform data. I then unparented the finger joint from its hierarchy, parented the null node onto the joint, and then moved the set back into the hierarchy. A little convoluted, but it solved the problem and allows me full functionality of the hand again.
I've also been working on my skinning, which is proving a little more problematic then I had originally anticipated, but it is coming along. I'm looking up further tutorials and other information from experts to assist, but I've got a good range of poses set up, it's just a matter of figuring out the strategy and learning what the model and rig can and cannot do. In the process, I'm learning a lot about how to better design topology for deformation, as I realize precisely what the theory really means, now that I can see it in action on a model that was this carefully planned.
Another success is that I finally got the mouth complete - I was having a huge issue getting teeth to parent correctly to the jaw bone and rotate right. No matter what I did, the teeth seemed to move twice as far as the lip would, meaning his teeth would receed or shoot right out of his mouth. It turns out that the lip was inheriting some of the skin weight from somewhere else, but that has since been remedied and we now have a great looking set of chompers.
Once I got them in, I ran the character through a number of facial poses, using the blend shapes I had set up and learning how he looks as his face distorts. Turns out? Adorable.
I even discovered, by moving and scaling the jaw bone, some additional functionality that the model possesses that I didn't realize - a happy little accident. Behold:
I've also begun work, and am about halfway done with the initial drawings for, the animatic, a useful stage in which staging and timing issues are resolved before moving toward animation. In doing so, I also created a more resolved version of the background image.
It is still a very loose and unresolved piece, but it solves a lot of the composition problems I had before, and getting the more final model of the environment complete this week, getting grass into the shot and getting hte first lighting pass is going to be very exciting.
Right now, though, I really, really just want to get this skinning done with so I can move on to starting the animation work. Time is running short, and there is a huge amount of work left to do.
(if you want to spoiler your post, you can reduce some of the thread clutter)
I think the little guy looks good so far. What would be awesome if you documented some of the problems so we could see them, also link to any good tuts you find.
Unfortunately I have no experience in modeling/rigging so I'm waiting to see some animation. Its pretty cool so far.
I've tried to document some of the problems, but often these posts are written against deadline of getting things submitted, among four other hectic projects. I try to screen-cap anomolies and such, but I do catch myself trying to cover up issues and making things look like they're going smoother then they actually are, since these entries are also being read by one of the faculty advisors that tends to get his fingers into everything if he hears there is a problem, and has a tendency to make things worse, rather then better. That's no excuse, though, if I'm trying to document this as an honest progress journal that would be useful for someone else going through the process, so I'll be more mindful of it as I go. Some of the issues in skinning I'll be dealing with this week, in particular, I think would be useful do document.
I'll go back and edit in spoiler tags on the previous posts on this page, at least. Thanks for the feedback.
edit: added spoilers. Any chance someone would quote my week 6 entry?
Added week 6 to my post.
What are the splines defining his head and chest doing in the rig? Are they just there for reference?
All spline shapes shown there are controller objects,. They don't actually effect the geometry, they're basically just handles.
Blocking Plus stage:
http://www.youtube.com/watch?v=Uu9fClGPewY
In this stage, I have begun working on smoothing the animation together in splines, but haven't yet had time to really finesse the motion. I'm mainly trying to get the first set of motion in, to ensure the timing works and the body moves between poses well.
First polish pass:
http://www.youtube.com/watch?v=1uldveCKNEU
n this pass, I begin working in the graph editor to really focus on getting the motion as smooth and natural as I can. I mainly focused on body mechanics in the feet, hips and torso, as well as tracking the arcs in the punching arm and the head. I'll be making one more pass at this to refine the animation before moving on to my final scene for the class.