Who Am I?
Hello, I’m Paul Scott and I’m a 2D/3D generalist but I generally lean towards creating characters. During my education I slowly transitioned from focusing solely on 2D images to mainly 3D. The major shift towards 3D happened during my second year at university (during a concept artist course) when I was introduced to Zbrush and told to make a few head sculpts in less than 2 hours each week.
The whole idea behind that was to stop me spending hours cleaning up a crappy attempt and instead move on to a new one and use the knowledge gained on that one instead, gradually improving through iteration. It was difficult at first, however. After 20+ head sculpts, things started coming together and I would recommend it as a great learning exercise to anyone starting out with sculpting.
After university I learnt the basics of programming, design and animation to ideally create my own projects as a side goal after working professionally in the games industry. I currently work for Pixel Toys in the UK and focus on creating characters and VFX for our various projects. Every new character or effect is a blast to work on and I learn new things constantly to speed up my processes.
Currently a friend and myself are trying to create a game with a food based theme and every beast based enemy will drop some specific materials. This idea was the basis for all of the creatures I’ve been creating recently, so on to the Maple Leaf Spider. We needed some way of introducing syrup drops for cooking and went through various animals thinking of what makes the most sense. We thought it would make the most sense if it were some type of creature that people already associated with shooting things that slow the player, as the player should be able to assume some type of slowing mechanic when seeing this enemy for the first time.
Eventually we thought about it being dropped from a spider and began designing how we would want it to look.
The game is designed with a top down camera that is at a far distance from the player, so all creatures require a strong silhouette to make them easy to identify at a distance. Due to this, having a large single form as the basis of a creature really helps make them identifiable.
Since I am the one creating both the final concepts and 3D models, I don’t spend too much time in the concept stage as it would be unnecessary extra work. As long as the concept is identifiable and we like what it generally looks like we move on to the 3D stage.
Building Blocks and Sculpting
For every character I make I start with a series of spheres in Zbrush that are hammered into the vague shapes and overall forms of the models. This process helps me when working out what I can get away with duplicating, to save time on sculpting. Also, it quickly gets the overall silhouette solidified to see if it still stands up in a 3D space.
With the main body completed in a rough state it was time to move onto creating the focal point for the creature, the maple leaf. I could have modelled that out of a sphere like everything else but it would have taken a lot of time to get the correct shapes and consistent thickness desired for the leaf. Whenever I’m in that sort of situation I like to go into Maya, create a mirrored plane and use the cut tools to work out the key shapes. Once I have that, I clean up the topology and use deformers to get the nice curved shapes I desire. When I’m happy with the shape, I smooth it to make it ready for importing into Zbrush (if not smoothed I find Zbrush has a tendency to shrink the edges of objects).
When I am happy with all the major forms of the model I begin creating the more intricate shapes and forms. This part of the process generally takes the most time and sometimes I can spend way too long on unneeded details (something I need to improve on). For super complex forms that cannot be achieved with a plane (Such as the furry legs) I create a single detailed ‘fur’ clump and duplicate/skew it into chunky forms to work out the flow of the details.
When I am happy with these forms I merge all of them together and blend them together into flowing clumps. I could sculpt them all directly onto the main mesh but it causes so many headaches when a brush interacts with a finished clump; doing them separately at first gives me full control over each individual clump without the worry of moving the others.
Making the Sculpt Usable
Now to my least favorite part of the process, retopology. I decimate the separate pieces of the model in Zbrush and import them all into Maya, then begin the process of using quad draw to build the forms in a lower poly resolution.
The spider has been quite a special case for retopologizing. The furred sections could not be created at a super low resolution then smoothed to add detail without the flow of the polygons being incorrect. Instead the major flow of polygons were created around the fur clumps then the gaps were filled in following these guidelines. It took a lot of extra time to draw the major lines manually, but automatic methods were not producing acceptable results for the higher polygon count they would give and their topology flow would have made UV mapping difficult.
When I’m creating a low poly I try to keep in mind where the model will be deforming the most and I give those areas extra polygons to make deformations better.
When creating UVs I always try to make as much of the model mirrored as possible so that I get much more texture room for the model on a single sheet. With UVs I prioritize the size of islands based on which part of the model is going to be visible most of the time. An example of this is having the front of the leaf use twice the space on the UV map as compared to the back of the leaf, which is rarely seen.
To speed up my workflow with texturing I have a load of premade smart materials that I have created. A lot of creatures can get away with using these materials with minor tweaks and they make colour palette swaps relatively simple. I tend to set up some basic smart materials at the beginning of a project and use them as a base for all assets in that project. It helps keep everything looking consistent in the world.
Rigging and Animation
Finally, on to bringing the creature to life! For most creatures I create a simple rig, starting with a root bone at the 0,0 then creating the rest of the rig based on that root. This makes it so that you can control the location of a creature without affecting its bones, which is perfect for making run/walk cycles.
This spider has a lot of its bones in its legs and leaf so that they have a nice amount of deformation points. I am a fan of flowing animations and that requires a lot of bones to get looking right. Also using a set of basic nurb controls and an IK for each leg made keeping the legs in the right location a breeze.
When animating I create a series of layers each with either a main animation or a tertiary animation on it. The tertiary animations are things like a subtle sway or breathing. These layers are used in conjunction with a single main layer to add extra subtle animations without having to recreate them every time a new animation is needed.
When creating the idle animation below I used 3 tertiary layers, swaying, eye twitch, breathing and 1 main layer, the main idle animation. This enabled me to focus on only the big movements for the idle whilst knowing that a tonne of subtle animation was being blended on top.
The walking animation was definitely the most difficult with this creature, just because of how many legs have to move. This animation is where the root movement layer comes into it; this layer simply moves the root bone linearly from one side of the grid to the other. Using this guide I snap each leg to a point on the grid and keep the inactive ones in that spot as the creature moves forward. When all the legs are set up correctly I turn off the root movement and I now have a functional walking animation with the creature in a static place. Without this root movement layer it would have been difficult to guess how fast the creature is moving and where its legs should be placed.
Why I Use Sketchfab for Showing Work
I first started using Sketchfab to get some inspiration and look at the techniques people have thought of to create different styles and effects. Once the quality of my projects were at a point I was happy with, I started using it myself and haven’t looked back since!
It is great to have full control over so many options in my models and easily/quickly be able to get a nice looking lighting setup. With the new model inspector it makes me work a little harder to make all my textures as efficient as possible. I generally create my lighting setup for still renders in Maya/Marmoset and I find that I can easily replicate the set ups on sketchfab, which is fantastic! Also with new features constantly being added I can keep improving the rendering of my models.
For my lighting set up I always try to have one overall light source, a strong side light and subtle light coming from underneath to fake bounce lighting. I find that having this setup and a floor plane really help make a character feel solid due to the clear shadows cast onto the floor.
I also use quite a few of the post processing filters to help pretty up a render (Ambient Occlusion is great!); most are on a low strength so they don’t overpower the scene.
I’ve tried many places to display my work but none of them have the capabilities of Sketchfab for showing off my work. As a bonus I get some feedback from a great 3D community.
Thanks to the Sketchfab team for giving me an opportunity to show some of my workflow and for creating a great website for 3D artists to get inspired and gain useful feedback for improving their work. Hopefully you guys found this breakdown interesting, thanks for reading!