By Senior Artist Kim McCarthy
I am the resident CG Artist here at Cloth Cat and I first joined 8 years ago as the character artist on Shane the Chef. Since joining I have worked on a variety of projects in games and animation with varying demands and have learnt so much along the way!
I’ve always taken interest in anything creative growing up, exploring every area I could, including illustration, painting, dress making, product design, furniture, computer programming, game design and animation. These have all served me well going into the world of CG where I’ve created worlds, characters and spectacular particle effects. I eventually came to the decision to delve into CG after taking a Foundation course in Art and Design, focusing on illustration for character design, and then going on to a BA and MA in Computer Animation at the Atrium USW in Cardiff. Here I specialised in character creation in 3D, using Maya and ZBrush throughout. I became proficient in both and developed an understanding of the requirements needed in 3D characters.
After leaving university, I did some freelancing until I was lucky enough to get the role of 3D character artist on Shane the Chef. Here I used Maya, Zbrush, and Modo for modelling and Substance Painter / Designer for texturing, and finally Redshift was used for rendering. When Shane the Chef came to an end, I was asked to stay on here at Cloth Cat working with research and development projects, games projects and helping out on some 2D projects with my background skills in illustration.
Over the years I have become aware of Blender and its ever increased popularity in the industry, particularly using it for creating assets for games and animation at the same time. I was always reluctant to leave Maya having been told this is what the industry used, and other similar programmes I had tried before didn’t quite match Maya’s capabilities. My understanding was also in Maya so having to learn a new software was always off-putting.
Blender’s open-source nature played a crucial role in creating a strong community among 3D artists. Unlike proprietary software I have mentioned such as Maya, which can sometimes feel isolating due to its specialized user base and steep learning curve, Blender encourages collaboration and sharing. The vast array of tutorials, plugins, and resources available through the community has been invaluable in my development as an artist. As a result, I’ve been able to connect with other artists, share insights, and exchange ideas, creating a vibrant and supportive environment where creativity thrives. I found that I could quickly iterate on ideas and visualize concepts more seamlessly than ever before. This newfound freedom led to a more organic creative flow, enabling me to bring my artistic vision to life with greater clarity and enthusiasm.
The expense of large sets and render times has been a challenge throughout every project. In 2019, Unreal Engine released real-time ray tracing as a Beta feature in Unreal Engine 4.22. Real time rendering allows users to view how what they are working on will be rendered as they are working, rather than making changes and waiting for them to be sent down the pipeline and rendered before having to make further changes. This could save a lot of time going back and forth, instead providing instant visual feedback and therefore allowing artists creativity to flow by being able to get instant results.
I found a vast and friendly community of Unreal users willing to share their knowledge with others. As soon as Unreal 4.3 was released we set out to investigate the plausibility of using it as our rendering software. Previous projects had seen render times soaring, often needing to render things multiple times when developing a shot. This meant we were losing a lot of time here that could be used more effectively during development.
Unreal at the time wasn’t able to handle rigging and animation within the engine and this would need to be created in another 3D software. Our goal was to use Blender for the creation of assets, rigs and animation. Then Unreal for building sets and rendering. This was slightly daunting to me as I was still learning the software alongside working on the project. I soon learnt that having an understanding of any 3D software gave me the tools / insight to adapt and learn in any other modelling programme.
To achieve the relationship / bridge between Blender and Unreal, we needed to understand the data that needed to be sent between the two programmes. Working with a programmer and our technical artist at Cloth Cat we first set out to understand this by rebuilding Shane the Chef assets in Blender / Unreal. This was different to how the assets had originally been created in Maya and rendered with Redshift. Instead of working with sub division and smoothing methods, I now needed to think like I was creating assets for a game; thinking modularly when modelling, modelling with hard / smooth edges and the use of different levels of detailed versions of one asset for optimisation; LOD’s.
Texturing assets using a trim sheet method rather than one texture sheet per asset meant higher detailed textures with lower memory costs. For example; allowing a 4k wood grain being shared across many assets which would have been previously textured completely on one 4k texture map meaning much higher details at a lower memory cost. This was a whole new way of thinking to me as I had primarily focused my studies to a more animation based workflow. Though I soon realised this way of thinking was adaptable to both workflows and was all about optimisation.
Some time after our research using the assets from Shane the Chef we were given the project named “The Secret Spring”, which would be 12 episodes (3 minutes each) created in two language versions. These would be used in classrooms in Austria alongside workbooks that would contain questions based on what they will have just watched. The possibility of creating an interactive game alongside the animation as well as what we had just learnt drew us to the use of Unreal and Blender for this project. An early version of Ynput’s Ayon formed the pipeline and the studio was able to secure the Epic Megagrant which funded the development of the bridge created between Blender and Unreal alongside the creation of our project “The Secret Spring”.
Working on Secret Spring taught me a lot. For the first time I was directing another artist alongside Jon in creating assets and building the environments whilst I could focus primarily on characters. Using vertex painting to paint landscapes again allows high detailed textures to be painted across the landscape. We used Quixel megascans at the time to add high detailed scans to be added with low memory costs as they were already optimised for use in Unreal. The major challenge here was the style of the project was to be rendered with a toon style shader. The high detailed textures and scans that were available were too high detailed for this style of shading. To overcome this we created a pre-set filter in photoshop that would add some stylisation to the textures but retain some of the detail, we could then paint on top of these if we felt the need to.
One of the really cool features I wasn’t expecting to work so well was the ability to have multiple users working in one Unreal project. This took some time to work out how to set up the project which needed to be on a shared drive, but would allow different artists from different departments to work at the same time seeing instant updates. The main thing here that concerned me was changes being missed and not reviewed properly. Another issue we encountered was the memory requirements for such large project files being edited by multiple users. This required each user’s machine to be upgraded with a large cache drive to handle such data. Little did we know that we were about to go into Lockdown, this could be a valuable tool in such a situation.
Animation was to take place outside of Unreal and the sets that will eventually end up in. These sets being huge meant it wasn’t feasible to have all that information in an animation Blender file. To overcome this we exported the lowest level of detail asset for each set and removed any unnecessary details that didn’t concern the animators. Occasionally we would need to adjust foliage in our Unreal sets as these are so heavy in terms of poly count for say each leaf; that we often left that out in the animation version of the set.
Once an animator had finished working on a shot they were then able to send across to Unreal and view their animation in viewport with a fully rendered view, allowing them to review their work how it will eventually be seen and act on any further development instantly. A more natural workflow compared to working in a cheaply shaded viewport before waiting for a render.
Modelling the characters was very similar to my previous workflow. I worked from 2D turnarounds to sculpt the characters in Zbrush. I then re-topologised and UV’d in Blender before sending to Substance Painter for texturing. Though I now needed to think about modelling without the option to subdivide and the possibility of creating lower level details of assets that would be triggered depending on camera distance. These needed to be set up differently than in a game however as we would never want to see the transition of an asset between two LOD’s in a shot, therefore these were selected on a shot by shot basis.
Since the project I have worked solely from Blender as my main modelling software, I am however still using ZBrush for some sculpting. I’ve collected a lot of Zbrush tools over the years, while leaving them behind is off putting, I can see other artists have made the transition so I have started to also. I am now starting to dabble more and more with Blender’s ever growing sculpting tools.
In conclusion, my journey into 3D animation and character production has been profoundly impacted by my time at Cloth Cat and delving into the use of Unreal and Blender. Working in Unreal / a games environment made me approach assets differently with the use of modular assets, trim sheets and vertex painting. Trying to be as efficient as possible.
Blender and Unreal are community-driven platforms which have made sharing and gaining feedback easier than ever. The availability of tutorials, forums, and collaborative projects means that I can continually improve my skills and adapt to new techniques. The software has not only enhanced my efficiency but also elevated the quality of my characters. As I continue to create and innovate in the realm of animation and games, I’m excited to see where my journey with Blender will take me and what Blender has in store for us in the future!