No Spring Chicken....

A bit more progress on the digging character, this time he's rather worryingly doing the chicken.....

This is really just an exercise to test the bind on the character but it was peculiar enough to make me render it out and present it as part of the workflow of this project.

I downloaded the mocap data from the Carnegie Melon University database.... it's pretty extensive and most importantly it's free for anyone to play with!!! The data hasn't been cleaned up or adjusted and, as you can probably see from the rendered animation, I've not bothered doing any of that for this test as I am just checking the bind.

Initially I wanted to rig the character with IK, FK and a third skeleton that could be controlled by mocap data but I have decided to forgo the IK and FK parts and have opted to have only mocap driving the main body of the rig.I did still bind a version of the character with a manual rig and IK FK switching on the arms and legs.... I will probably only use it for posing still renders rather than animation but it was a good refresher exercise in rigging and binding.

The hands and face of the character do still rely on manual keyframes which, in this video I have not demonstrated but please believe me... they work. I have used a couple of simple lines of MEL script to drive the mouth and the eyes, it uses noise over time to drive the rotation of the joints... I felt that the lack of movement ni the face of the last video was a bit unsettling... I suppose that in it's own way this new video is equally, if not more unsettling but it's not because of the face being motionless.

The next stage for this build is the blend shapes for the facial expression and geometry correction in extreme poses.

Texturing & Re-targeting......

So, a progress report on my character is probably about due so here goes.....

Since my last post I've made quite a bit of progress on the old guy character that I'm working on. I spent a little more time on sculpting and then moved onto texturing the mesh.

As I'd been using Mudbox for the sculpting I thought I might as well investigate it's texturing tools as well, these allow you to paint directly onto the model in 3D space.... this workflow was a revelation for me. For years I have created 2D textures in Photoshop using the UV maps that I have made in Maya as a template, this workflow is fine but even the best UV maps can distort the finer detail.

Mudbox still uses the UV map as a template but being able to paint directly onto your model means that what you see is what you get, there's no need to jump between Photoshop and Maya to check that the texture is fitting properly or to tweak UV's after the texture has been applied. Obviously, 3D painting tools are not a new thing, it has actually been possible to paint textures in Maya for many generations of the software but the functionality has been more suited for roughing out rather than detail work. The painting tools in Mudbox, are fantastically responsive, working on layers allows the build up of texture and being able to use reference images as colour stencils is genius.....

The above image illustrates a visual evolution of the shader network that I have used in Maya for this character. Until this point I had never really used the Sub Specular Scattering (SSS) shader so I figured that with all of the other new knowledge that I have crammed into my brain through this project, a little more development wouldn't hurt.

The SSS shader works like skin, different layers allow light to penetrate and scatter as it interacts with the material. Rather than my getting bogged down trying to explain it all here's a link to an explanation from Autodesk.......

Mentalray Fast SSS Tutorial

Another area that I've been looking at since my last post is joint re-targetting through the Human IK (HIK) system in Maya. Since this character is going to be used as a motion capture puppet I figured it would be prudent to begin to develop my understanding of how this system works.

I figured it would be more entertaining to see the old guy move than a naked rig so I quickly bound a very rudimentary skeleton to the mesh and defined it as an HIK character. The fundamentals of the HIK system are pretty straight forward and the GUI (see below) pretty much guides you through the process of setting up the character.

It is as easy as selecting a joint on your rig and then assigning it to the relevant indicator on the GUI. So, after defining my simple rig as Character 1, I defined a mo-cap example rig as Character 2 and used it as a motion source for Character 1.... it's really that simple.

Below is a render of the driven motion..... Like I said the rig is very rudimentary and there was no attention paid to setting the joint influence but it kind of works. I find that because there is no facial or finger animation the resulting motion is rather unsettling, dead eyes and limp hands give it a peculiar reanimated appearance reminding me of Overtime, a dark tribute to Jim Henson from Supinfocom

The actual rig that I eventually apply will have facial controls as well as hand controls which will be animatable through keyframes... I don't think I want to start looking at facial mocap for this project but who knows I might end up doing just that!!

Getting Muddy....

In the little time I've had to play recently I've managed to make some progress with the character that I started in the previous post.

One of my aims in this collaborative project is to develop an understanding of the sculpting software Mudbox, it's toolset and the workflow between it and Maya. Up until this point I've never really had the desire to sculpting software but packages like Mudbox and Z-Brush are becoming a more and more integral part of the production pipeline and so I figured it was time to get my hands dirty.

I think that my reticence to use sculpting software stems from the heavy handed approach that so many artists have demonstrated through their practice. I have found that there is an abundance of work created through these packages which carries a certain look, a look that screams "sculpted" at the top of it's voice. Deep gouges and symmetrical lines are the most common offences and it is this style that I am hoping to avoid in my efforts.

Anyway, once I got over my fear of carving heavy marks into my model found the transition to sculpting pretty rewarding, the tools in Mudbox are very intuitive and after I had learnt to navigate 3D space with my Wacom it became a rather therapeutic exercise. Developing a workflow took a little time but I soon realised that, much like sculpting in real life, it's best to work down to the building finer detail from larger strokes.

Now that I have a project that benefits from the use of this software I am beginning to understand it's potential and will most certainly be using it again in the future. I still believe that it is important to understand that not everything needs to be sculpted and good old fashioned texture and bump maps are still absolutely valid practice in the professional workflow.

Can U Dig It?......

I've recently been working with artist Garry Barker to realise one of his characters as a 3D model and it's already been an interesting learning experience for both of us.

For me, this project began when Garry and another of my colleagues, Annabeth Robinson, had been discussing ways that Garry's digging character could be presented through alternative media, this led to talk of motion capture which ultimately requires a 3D model which in turn led to my own involvement.

Although there is plenty of reference through his work and the attributes that make up the character are pretty identifiable, it was difficult for me to consider it as a 3D model so I figured I might as well ask Garry to provide me with some orthographic drawings to work from. I thought  that this wouldn't be particularly hard for him considering it is his work after all and the character shares a very similar blueprint to the rest of the figures that he draws but, it turns out that asking an artist who deals in two dimensions to consider anything that they've created and then drawn countless times as a three dimensional object that can be animated might throw up all sorts of new thought processes for them. He got there in the end and produced some images that I could easily work from in Maya.

The character style is rather doughy, no insult intended Garry, which leads to interesting proportions and an asymmetric look which is very different from the style of character that I might normally build. Still, modelling is modelling and this figure didn't really cause me any problems in it's construction.

I'm pretty pleased with the base mesh, pictured above, the geometry is clean and even without textures and bump maps it has a pretty good level of detail... I could feasibly rig it right now and have a nice model to animate or motion capture but why have a nice model when a flippin' brilliant one could be the order of the day..?

So what's the plan?

I'd say a bit of learning in the form of Mudbox, sculpting software which I am a familiar with but have never tackled anything like this before, and muscle systems, dynamic deformation that has been available in Maya for some years now but something that I have never really had the desire to use. I also intend to develop my understanding of Mental Ray shaders, in particular the SSS (subsurface scattering) shaders that allow rendering of realistic skin.

So, I've got my work cut out for the next while..... I'm not sure how long this is going to take considering the limited free time I have but I'm pretty excited about developing it into a pretty amazing puppet that can be used to bring Garry's art to life.

I'll be posting progress here as and when it happens.