Evaluation

I created my character from the ground-up, but forewent creating storyboards or line test as this is not the usual means by which I go about creating digital art. I did, however, as advised, research examples of existing pieces of motion graphics.

It was tempting to go down the typical path of deconstructing and analysing Pixar movies, but I thought I would take a different, rawer approach and look at often overlooked areas of motion graphics such as opening title sequences and commercials.

As demonstrated throughout this blog, I created certain aspects of my character in conjunction with new software (Mudbox) that I was trying to learn. While I am content now that I can use Mudbox effectively, I felt that this approach was somewhat restricting to the final look of my character. I was using certain tools just for the sake of using them -- such as the scales or stamps on the back of the toadstool. This would be a feature more attributed to a video game villain. Perhaps lighting and a properly textured and shaded scene would have aided the look of the character, but overall, it didn't fit in with the theme of the cartoon and he audience it was intended for.

I'm not against the idea of having stamps or scales. It was the way they looked. A bit too real and possibly too scary for children.

The recent Fanta adverts demonstrate the point I'm making:

http://www.youtube.com/watch?v=ExDZMbfN1vM

http://www.youtube.com/watch?v=bteq9_CFFlo

I like the CG direction very much. It has a colourful summer theme which all can appreciate, whether children or students, or middle-aged, and takes into account what makes a character interesting. For example, the characters in this commercial demonstrate the "exaggeration" principle of animation, with heads almost as big as bodies. And another reason why I like this, and it's a personal one, is because it looks as though it was predominantly made using packages similar to Maya or Max, which is how I like to work. It is quite irritating to have to migrate between software routinely.

Another point I wanted to make was the use of shapes. I tried to use some of the attributes that were so common in Disney villains, such as sharp corners and disfigured bodies, while utilising some of the theories that apply the protagonists, such as round shapes and curves. Hence why I opted for a perfectly round toadstool cap instead of the deformed one show in an earlier blog entry, and for sharp, menacing eyes.

This Fanta advert does that well. Round heads and rounded corners on the desks. It is a very friendly advert.

When going about my creations in 3d, I often have the goal of learning something new, as well as improving my skills. The next step to hone, after this, is to consider who I'm making my work for. Everything that I have learnt up to this point will enable me to do this.

Animation 01 - Walking

The animation for this blog entry is on the CD I handed in. It is called Walking.avi

The only problem I have with this is that, even though he isn't walking, I failed to make it look as if he is. People who I showed this to have said that they didn't get the impression that he was walking, but could recognise that I was trying to make it seem like he was. After I explained what happened to the legs to them, they said it was a good try, but could be better. The arms were okay though. If the character was walking, then the arms would still move like that.

One thing that stands out to me can be seen between 6 and 7 seconds of the animation, at the arm pit of the left arm. As the character raises his arm and points out, there is an apparent defect in the skin.

It is more apparent in the animation than it is in the stills, but it was to be expected. I said earlier that I the arm area may be compromised because of some changes I made in the topology in order to get the shoulder to look the way that it does.

In the future I will be more mindful of this.

The ending, where he points at the camera, was made in order to supplement the camera movement, and to give it a nicer ending that is vindictive of the character's personality.

I prefer the second half of the animation sequence to the first. Not because I have created smoother animations, but because the actions are more interesting. I'm pleased with the apparent seamlessness of the head movements working in conjunction with the arms moving around. The blendshapes worked charmingly, and I think I will strive to use them instead of skin clusters from now.

I am not so much an animator as I am a 3d artist, so I have surprised myself with some of the seamless transitions of the character, save for a few mistakes that I have outlined here.

Combating the lack of leg movement

I already touched on how I could elude the legs through camera trickery, so the next step was to decide where the camera would be placed. It would need to be close up, covering a span of the character's waist, to the top of his cap.

Something like the above. It hides the legs, an focuses on the areas that I have done well, with the face in plain view, meaning I can show off some of the blendshapes that I created.

Currently, I have a "mock" walk cycle planned for him. He will be strutting along a path with shoulders rocking back and forth in an imposing manner, with the camera bobbing up and down whilst he looks around and points.

Attaching the sword

The more disappointed I was that I couldn't use the character's legs, the more I thought about using the sword as a substitute for some kind of action. I spoke about the sword earlier, and how it may be problematic to attach it to the character, and it turns out, it is a task in itself.

If I wanted to convincingly attach a sword to a character's hands, I would have to create a different, and separate, kind of rigging system. There are also MEL scripts which apparently do this for you, but I have no knowledge of that, and wouldn't even know where the begin.

The closest I came to successfully attaching the sword was by using an open-source tool called "ZV Parent Master". Its use is demonstrated in the video below:

http://www.youtube.com/watch?v=dUrYuU3hCYo

It is a very useful tool that actually over-simplifies things. It's great if you're in a rush, but it lulls the user into a comfort zone and diverts him from learning new things or alternate methods that could be useful in other circumstances where the ZVPM is not available.

Anyhow, I decided against having sword related animation. The rig I spoke of earlier -- the special "sword" rig -- that is something that I will look into further later on. It would be useful for many instances, not just swords, as characters are often expected to hold things.

Creating a context

In order to help envision my character's actions and motivations, I created a small scene. It is typical of the type of place where he would clash with the protagonist. It is supposed to look very lush and green, because it is the place that the protagonist is fighting to save. If it looked awful, who would want to save it?

There are intentionally lots of wild toadstools around to convey what both characters are fighting for. They are supposed to be the villain's "own kind", yet he is willing to sell them all out for the sake of money and power.

I also took a few shots of the complete character inside the scene, to see how he would fit in here. I imagine it is an area that he has invaded with his minions, and the protagonist must fight to get it back and preserve it. Bare in mind, no lighting has been applied here. It merely serves as a visual aid.

While I was creating this scene, I thought of the idea of the villain sitting, in a closer shot, amongst some of the toadstools cross-legged with his head bowed, almost camouflaging himself amongst them and appearing to be just a regular toadstool. But then he would raise his head and look into the camera, and suddenly burst into life. Of course, with the legs turning out the way they did, this wasn't a possibility.

Beginning animation

Now that my character is ready for animating, I need to decide on what exactly he will be doing. As I have said, the choice to create long-winded martial arts moves is now out of the window. I can still do that, but only wih moves that focus around the arms moving.

For now, I will concentrate on containing his villainous quality. I want him to look mean, as if he is a contentious and vain character who stops at nothing to get his way. Someone who treats his minions simply as subjects who are there to carry out his will. He already has the facial expressions for this, so it is a lot easier to envision it. So, lots of finger pointing and head gestures.

Skinning

Just like rigging, skinning was also a long task. I came across a few kinks too, such as incorrectly using the mirror skin weights tool when painting the chest are. This adjusted the arms that I had already done, and required me to re-do them, which was frustrating, considering how long painting weights properly takes.

Slowly but surely, I got through the whole of the upper body and created influences as best as I could, and was able to move my character in most positions.

I came into massive problems when dealing with the legs however. The way I had modelled the character made it impossible to properly distribute weights smoothly like I had done with upper body. I created only one edge loop between his two legs, which ran through his groin area. At least two or three are needed here.

This was the result of that:

Starting again at this point in time was not an option. Everything would have to be redone for the sake of two more edge loops. With this, I knew I wouldn't be animating the legs like I had planned. I wouldn't my character to perform a few action sequences in order to demonstrate that he is an effectively tough villain, which would involve kicks and jumps. I even rigged the legs in a way so as that animating jumps would be extremely easy.

Along side his action sequences, I wanted to create an animation of him looking like he is giving orders to minions. Naturally, in light of this, I will have to omit the legs and have the camera focus only on his upper body. Perhaps movement of the waist and an rocking camera can perplex a viewer into believing that he is walking.

Nevertheless, this is something to learn from, and if I'm going to make a mistake like this, I would rather that it happened now while I am student than during the course of my career, as it can be very burdening. 

Rigging

It was a lengthy process, but the rig is complete.

I wanted to utilise an IKFK (inverse kinematic, forward kinematic) switch handle, as shown above. It is a technique that I learnt during a previous module, but never used it then because I changed from character creation/animation to visualisation. Despite this, I was coming across problems in creating the rig. I think it was an orientation problem, and the root joint decided to ignore world space. Typically, I would persevere and continue until I found the problem, but it wasn't really necessary to use the IKFK switch here. I wasn't planning on having wild, over the top animations.

So instead I settled for a standard orient constrained arm set-up. It's boring but does the job.

It was a shame, because I have used an IKFK rig before with rigs that I have downloaded, and it's completely user-friendly and gives the user more control over the arm, as well as allows for fluid animations. I wanted to see how my one would fair in comparison.

Preperation for rigging

The sculpting is done, as are the textures, and facial animations. I am happy that I am now comfortable working between Maya, Mudbox, and Photoshop, and am familiar with the integration workflow. Now comes the rigging process, which is one of my favourite things to do, despite my reluctance to work in character creation. I enjoy the problem-solving element and overcoming challenges. I have yet to come across a rig where I didn't meet some sort of anomaly which required the use of my brain.

I am interested to see how this character turns out. Naturally, the limbs are expected to move, but what about the neck? Even though his neck is quite stocky, I intend for it to move and operate like any other neck.

Furthermore, In order to get the arms to look the way they do, I had to compromise the topology around the shoulder area. I suspect that will cause a little bit of a hiccup during animation, but don't expect a huge problem.

I have never created a character that was to hold something, so I'm interested to see how I will tackle the issue of the character holding the sword. I suspect a series of constraints and parenting will do the trick, but am not sure at the moment.

Other aspects are accounted for, however. The sword fits in the hand nicely, and the fingers match the hand nicely -- the fingers aren't too long and the palm isn't too small.

Blendshapes

The animations for this blog entry is on the CD I handed in. They are found in the folder called Blendshape Tests.

At first, I was convinced that I would create facial expressions by using skin clusters and painting weights, as this is my comfort zone, but I have already made a point here about wanting to push myself and learn things that are foreign to me. Mudbox again had more to offer in this regard. It is possible to use Mudbox sculpt layers to create blendshapes that can be used for seamless animation in Maya.

I had heard of blendshapes before, and had even used them, but never created any of my own as I thought that modelling polygons in maya at a sufficient level for them to be effective would take too long, because Maya is only good for creating base meshes as far as modelling goes. I never knew that people were using sculpting programs like Mudbox and Zbrush all this time.

Anyhow, once I knew what facial expressions I wanted, I set about sculpting them in Mudbox, and then brought them back to Maya for animating. 

Here are some printscreens of the expressions, starting with the default one I made with the base mesh:

During the creation of this character, I decided he wasn't going to be a dastardly, scheming villain like Jafar or Maleficent, but rather a fighter like Shan Yu. That is why the facial expressions are limited to angry grunts and hard blinks, instead of sinister smiles, or raised eyebrows of opportunity.

I tested them out and they all work. The final shape looks a bit odd, but I think that with some context it will be more suitable, like when he is pulling this face and performing an action that fits.

Here are some playblasts:

Final textured version

A few printscreens of the model fully textured:

The next step is the face.

Comparison between skinny and bulky

With the texturing complete, I imported my Mudbox file from Maya and decided to put my two creations side by side to see which one would be most convincing as a villain.

First, the new, tweaked villain alone:

And the comparisons:

I like the adorned Mudbox version better. The skinnier design is okay too, but I think that it is more suited to the role of foot soldier or minion. Someone who the bulkier one can command and boss around.

Texturing with Mudbox

I opted to use the painting tools within Mudbox rather than export to Photoshop as a means of fortifying myself with unfamiliar software and becoming adept in it.

Admittedly, it would have been so much easier for me to do this via Maya and Photoshop. But I wanted to learn how to do this and have the software under my belt. Luckily, all of the concepts and principles of 3d space, materials, and spec/bump maps that apply to Maya and Photoshop also apply to Mudbox. After a day or so of constant use, I felt comfortable using this software for this purpose, and am happy that I persevered.

Here is a shot of the paint layers that I used to get the shader settings correct:

Sculpting a bulkier villain

Having exported my low-resolution base mesh into Mudbox, I quickly added more sub-divisions and began sculpting.

Already, the character is taking up a much more shapely and charismatic form. One which so far looks much more befitting of a villain.

My first time using Mudbox, I really enjoy the attention to detail that I'm able to achieve almost effortlessly, such as crevices in certain muscle areas such as the lower leg.

Considering what I touched upon regarding attributes of villains, such as rough, gravelly voices, imposing physiques, and spiky/sharp corners, I thought I would make a few touches of my own. Of course, I followed suit and applied the theory to the eyes, as shown below. But I wanted to exasperate the feeling that this is character is a piece of foliage from the wild, so I thought to make use of the stamp tools within Mudbox and add some nice scale-like detail, which is true to the villainous form.

Inspiration for this idea came from the enemies from the video game Gears of War. I always admired the modelling techniques used to create them. I was aware that Epic Games used Zbrush exclusively when modelling for that game, and always thought it was a difficult task to do, but it's actually surprisingly easy.

Here are some shots of the said characters:

UV Mapping and texturing, and Mudbox

With the base mesh as complete as I could get it in Maya, it was time to start UV mapping in preparation for texturing him.

Once this was completed, I was free to take a UV snapshot and export it into Photoshop to begin painting. However, at this stage, I decided, what if I want to take the design further? I did mention that I wanted to say both a Jafar-like version and a Shan Yu-like version of my character. So far I have the Jafar-like version, and he looks a little bit plane, and furthermore, not very threatening as a villain.

With this mind, I considered the option to export my base mesh AND my UV map into Mudbox in order to get my Shan Yu-like villain, and to paint my textures. I was quite optimistic about this idea because I had never used Mudbox before, and the aim of this project, for me personally is to improve my prospects. So now, I am delving into the realms of character creation, and I'm using a brand new software.

Tweaking the cap

I modelled a few variations of a cap with imperfections, and then textured them to see what they'd look like.

And then applied it to the mesh:

I prefer the linear cap in truth. Not just because the imperfect one looks bad and uneven, but because it fits with the personality of a villain. Villains are vain and narcissistic, and he wants to look perfect and completely even. He makes it so that everything revolves around him.

Creating the character

The first stage in creating my character is to create an image plane which I will use as a built-in reference when modelling my character.

And as you can see, this plane helped in the creation of the foundation of my character:

Some hours later, the above became:

With the cap:

Finally, with the slippers and "sword":

So far, it looks like a good, bipedal character. However, I am unsure about how linear the cap looks, so I may do a few tests and experiment with an imperfect one.