Character Design Project

   The last month I've been deep in Maya and Zbrush with character crafting. My original inspiration was a character I had drawn up for Dungeons and Dragons with friends, but despite having a rather short life (dying twice in quick succession to a blast of light from a much more powerful magician), he made impression enough that I have fond memories of my experience with him.

   So, with such an impression I drafted him as the first for possibly a whole line of such characters, and a showcase for anyone who might like a similar result.for their own player characters. Posable, usable in animations and quite possibly a source for 3D printing if the setup is right.

   Magic users have had quite a compelling allure to them - masters of mystic arts, scholars and academics who shape reality. This talent, and a lack of armour for their class gives the opportunity to make them one of the more casually-dressed character types in games of D&D. Other classes certainly have a style, but one way I find magic users fun is how they can bring court style and artisan outfitting to an adventure.

   For this particular character, this connection to courts was key to his design as the background for him was of an aristocratic family in decline. This particular member of the ruling classes, who I called Varthen, took to travelling the countryside after becoming an adult. Funding his travels with money made performing tricks which come innately to him at various villages with the occasional performance in a town pub. As well perhaps as other services that come from magical ability coming naturally to him or having a formal education.

   These two factors: A noble background and a life on the road, created a character who was of privileged status but lived with few of the trappings. While I could have taken the angle of just a few trinkets to show his wealth, such as a signet ring or a necklace, I thought he'd go a little more flamboyant. Namely he would still be in touch with the fashions at court, and underneath a (relatively) simple coat would be a silver-threaded blue velvet doublet, linen shirts and silvered leather boots. Underneath a simple overcoat is a man of proud upbringing, who could weave his way through the royal circles if he so wished.

   He's not fully human, more a half-elf. While this does make him rather Eldar-esque, elves in the D&D settings are often far more fairy-like than Lord of the Rings. Half-elves have often been described as a miawy point between human and elven beauty, so a half-human-half-elf is more along the lines of a handsome or beautiful human with somewhat pointy ears. To an elf, they'd look ugly with rather puny ears.

   Most importantly it was really fun to design his outfit, particularly the designs of silver and padding on his luxurious buttoned doublet.

  The other fun thing was learning tricks to rig up large pieces of fabric in the absence of ncloth programs. Varthen has two arrangements to make this work: The lower half of his coat has a grid of controls that link to nine different chains of bones while his half-cape is controlled by a single chain of bones much like a tendril coming from his shoulder.

   There was a lot more work than I initially thought but that is part of the learning process. Some of the most fun things to work on were his hair, eyes (which are rigged with a blendshape to dilate), the pattern on his doublet and getting familiar with the rivet tool. Which I've found can be far more than a way to pin an object to a specific point on a mesh.

   With him done, I've had various ideas on my mind where to go next. Such as offering various components for sale like his coat with attached rig or a blank version of his head. But also considering less human characters like dwarves, tieflings or lizardkin.

Further reading

Artstation entry: https://www.artstation.com/artwork/N5bkON

Mosquito Redesign: Working Legs

  I didn't have much success implementing an IK nor getting it to compress, but they say the simplest solution is often the most effective. In the case of my piston the simple solution was to parent each of the pistons with each other, tie them to 'aim' themselves at the opposing hinge and create a few driven keys to mimic the pistons extending and compressing.

  After that it was a simple matter of constraining each of the hinges. and the pistons so they stop at a fixed instance into and out of their sockets. By adding a fifth segment to the leg I am able to extend its reach to a motion window of between 92 and 168 degrees (any more open and the piston will collide with the knee hinge). Which might be one of them ore complicated elements of the model out of the way.

From what has been assembled so far, I think I can easily fit six of the legs into the main body with plenty of room to spare.

A Modeling Sandbox

  I've been doing a little practice work on a model that I made long ago, I've had plans to revamp it for a while. With time in Mudbox I hope to make the skin considerably more convincing, which right now primarily serves as a base texture to build on. I mostly wanted ot go deeper with my modeling - better skin, better texture, better skinning, a test-bed for ways to create models that are a little more convincing

  The process was similar to the refinement of Four: A 3D model created in Maya transferred over to Mudbox then brought back to Maya as a low-poly model overlaid with a normal map.

It's not all geometry, underneath this structure is a skeleton so when the whole mini-project is finished this creature should be poseable, animatable and something that can be brought in to a more complex scene. At the very least I'm getting more familiar with Mudbox sculpting tools.

Four: Last Few Controls

  So far almost everything has been linked up. The last few elements required are the fins. As I found the IK control system quite handy, I decided that barring other suggestions I use that to make controlling the fins easier, as I had to get rid of a few controls for the old method when rebuilding the neck.

  There was some minor warping when I linked up the blendshapes to the model, but a loophole I found to reduce the severity was to reduce two of the shapes - one for the lips and around the eyes, one for the cheeks, are working at half-strength. While this does mean their controls need to be moved further for a better effect, it seems like a worthwhile compromise. 

The counter-roll for the spine's IK also works effectively. However since copying the rotation for the core joint overshoots the counter-rotation it might have to be controlled on a by-eye basis.

Four: Correction Sculpting Complete

  It's taking its time but the pose space corrections are about finished. The hard part was correcting the changes, now what remains are linking up the blendshapes above and creating splines for the tendrils on the head. Turning the head has proven tricky but I have a solution of tying the twist and/or roll parameters of the spine IKs into attributes on the body, allowing for easier correction when they twist the body out of shape.

  Some of the controls have already been tied in but not all. That is something I am currently working on. I may also need a new system for controlling the fins.now that I have fewer controls along the creature's spine.

Four: Presenting Play and A New Adult

Written into the script, one idea was for Four to knock over the phone as Amelia is talking. This scene should be more interesting than Four simply wandering into David's office and giving the camera puppy-eyes.

Anyone with a pet cat knows they love playing with inanimate objects by giving them pokes, jabs and pushes to see what they do or how they fall over. As a creature of predatory heritage, Four might have similar instincts so this might be entirely in-character for her.

Preparing for the third act I went back to the adult model. While installing the spine to simplify control over Four, a slight problem came out in that when installing a curve to a spine, the joints will contract to fit the CV curve, which would make reattaching the blendshapes problematic. The solution so far (which has given freedom in correcting the skeleton) has been to essentially fix the skeleton then rebind the skin and paint the weights back on. 

As I had done this before a couple of times now I might be able to re-paint the skin weights fairly promptly. That is the plan, as the deadline of mid-May is coming and I have to reserve a week or two for export to a render farm and post-processing.

Four: Signing Off The Grub

  The grub is almost ready to sign off. Controls are in place and tied to their bones while a basic skin has been added, and a texture map is to be in the works. With time in mind, it is fair to say that I am ready to start animating with this model.

  What is slightly jarring that might be coming from the glossy reflection is the skin appears to be reflecting the body around it, something that I've noticed skin to not really do. Turning off the model's ability to be seen in reflections does appear to remove this mirror effect however this is likely to make the creature's  invisible in the reflections any of the chrome surfaces in the lab (such as the floor of its incubation unit).

  The grub, compared to the others, is quite a simple design. Only one key control is needed for opening and closing the mouth. I might need to devise some teeth to provide something for inside the mouth, although it might not be as easy to see as other 

Four: Something Grubby

  I started work on a grub form the Sunday after I returned from my break, today I made considerable progress on it, crafting the rest of the body, adding definition and then rigging and skinning. As it is quite a simple shape compared to the other forms, the skinning process did not take long.

Aside from the legs, already the grub is about ready for controls. Like with the juvenile, it will require a spline IK in the spine, as it is quite a flexible creature and attempts at bones serving two segments did not work out particularly well.

It might need pose space corrections to give solidity to the segments. So when the segments compress they push upwards for that fatty "squidgyness" the grub will have.

I have considered that it might be best if there would only be one grub due to time constraints, There's no drastic change for the two considered forms apart from maybe the grub getting fatter, which could be achieved either with a blendshape or controlling the scale of certain bones. Or the talk of scale mentioned in one scene could simply be represented by the camera looking at what is simply a bigger grub in the tank. David's shock could easily come from simply how gigantic a little bug like this gets - growing from 5-10cm to almost a metre across the months it is studied.

Four: Finer connections.

With some planning, Four has a slightly nicer eye, something that is a little less human

  I may go a little larger with the iris but at present the iris' coverage is quite solid, leaving only a small amount of white. The eye could be signed off.

  Four's boy also needed some simplification. Going over I removed controls that could be covered by a spline IK, which bend the joints along a curve tied to control joints.

  The eyes are rigged and controls are set, the eyelids are connected and the membranes are set as well. So the priority for the weekend is to link up the last nodes, the digits on the and and feet and the fins, tying the controls for the tendrils on the head, then limiting everything so the pose corrections don't break.

In hindsight, this was a more important model, as it will be present for much of the second half of the animation.

Four: Juvenile Almost Complete

  Four's juvenile form is nearing its end. The last deformers have been modeled, the controls are in place, most of them are rigged up to their respective joints and the deformers have all been connected.

  What remains are the finer controls - the channels for the digits, same for the fins and jaw, and the replacement of the eyes with a better model and texture.

  I feel optimistic that I can sculpt out a base grub form and set up UVs and bones by Sunday.

Four: Improving Her Eyes

  Speaking to Alan today I took a minor detour on rigging Four to work on her eye. I want ot try something less human-looking so to start with I experimented with a few eye forms. Currently I like 2 and 6, Speaking with Phil and asking Alan for help in how to install one, I also implemented the idea of a nictitating membrane. A third eyelid that some animals possess to protect their eyes. In Four's case, it would be used to allow her to see while swimming in water.

  A texture still needs to be applied but the mechanics for the membrane currently work.

I plan to rework the geometry of the eye at some point but at the moment that is a secondary priority to finishing off the skinning process, which is almost halfway done.

Four: Weighting the Skin and Facial Deformers

  I set myself a deadline for about the end of this week where my primary assets would be completed. So much of today was centred around binding the skin to the body and hopefulyl signing off and starting on the grub before the month is out.

  I tried a technique for blending the weights where the skin is flexible, such as on the webbing, to make it flow more organically between the toes. For the most part it appears to work, the skin blends and moves more smoothly than on the adult model (which may need revisiting at some point).

  Also in need of creation were some of the key blends for the face. In addition to the setups I created for the first model, I added a set of deformers for the cheeks, to allow Four to visible puff and huff. These deformers are set slightly inside the mouth so that when she motions, the interior of her mouth will hopefully move with the exterior.

  One problem I had previously rigging the eyelids was that when rotating, the control joints would move slightly in a direction that I didn't need them to. To alleviate this I considered an alternative method, where an EP curve would be constructed to find the midpoint.

  Stretching between the corners of the eyes, the EP curve was rebuilt with two segments, a cluster deformer was created so that a locator could be snapped to the middle, which was then moved slightly up and down depending on whether the lid was for the top or bottom. Then the joints were attached.