Creating Mechanical Parts Animation Inspired by Metal Skin Panic MADOX-01 Anime

Introduction 

My name is Jake Woodruff, a principal environment artist with over 13 years working on AAA games. Primarily, I specialise in organic modelling, but I always love to push hard-surface design in my free time.

Metal Skin Panic

This project started with me discovering the anime Metal Skin Panic MADOX-01. I've always been interested in 80s/90s anime and traditional animation; it was a period in time that showed off lots of intricate and over-the-top mechanical designs and ideas, perfect for interpretation into 3D forms and development with modern tools and techniques.

The opening sequence is a montage of different parts of a mech being tested and shown, still camera shots with animation, and as soon as I saw this, I knew this was what I wanted to do. It was really digestible and easy to break down into manageable chunks (important when working on large projects) but also contained a wealth of inspiration and details.

I always want to develop new skills and try new things with each project, and Metal Skin Panic was perfectly aligned with developing a kitbash library to translate it into 3D. I had done some kitbashing before, but this would be a lot more detailed and larger to really expand my workflows. It was also a perfect opportunity to continue to develop my knowledge of movement and animation.

While Metal Skin Panic provided incredible designs to work from in a hand-drawn style, real-world references were still hugely important. This would be key in creating realistic materials and grounding everything in reality. This reference consisted of different mechanical parts, both close-up and contextual, to always refer back to when in the modeling or texture phase. Essentially, no little part or greeble is ‘made up’ but derived from real-life objects, which is key to making everything believable.

I also gather references of different types of lighting visuals I want to develop and tone-setting images that inspire the final renders as I progress. I think this is really important to initially guide decisions, even if the final results deviate from them.

Modeling

I started this project with a very thorough blockout phase. This was the key to working out all the different pieces that needed to be made and what could be shared. It helped plan how all the mechanisms and movements would work and also provided a very strong starting point for lighting and rendering the scene. Everyone world different, and while it may look chaotic, iterating quickly on a design like this is important to refine the visual design.

While the anime was a very good reference for the designs, to translate it all into 3D, I had to work out how all the mechanisms worked, how all the pistons moved, and what potentially existed at angles that couldn't be seen normally. This is where the blockout phase was key.

Once I had blocked in all 4 of the vignettes, seeing what pieces could be shared, I broke down the meshes into a library to go through a detailed high-poly stage. This was actually one of the easier stages, but it took a long time to complete. I wanted to be able to use all of the pieces at any distance from the camera while making sure each piece was highly detailed.

While it was a lot of work modeling all the parts individually, I could easily swap the pieces out to finalize in the blockout as I went and see the models come to life without much additional work, even with just flat materials.

UVs

The models themselves are fairly straightforward high/mid poly modeling using Blender. Blender is such a joy to do hard surface modeling quickly using a combination of booleans, bevel, and subdivision tools. The only Blender addon I use currently is ZenUV for unwrapping; everything else I run in default. I found that this makes it easier to pick up tutorials or change versions.

All the meshes could also have been baked down further to fit lower, real-time budgets, but for now, that wasn't the goal of my project. Keeping geometry clean is always a good idea, though, especially if you want to reuse pieces or even come back to the models further down the line from your library.

Texturing

I wanted texturing to be fast and effective, but I didn't want to compromise on details. I chose to break up the models into similar groups of type and size and have them share their own unique texture space, ending up with around 6 materials. While having a completely procedural shader/texture solution would be quicker to iterate on, I felt it wouldn’t give me the detail and visual I wanted for this project without a bigger time investment.

So from there, I created smart materials for each of the main types of material I had. This included metal, painted metal, rubber, plastic, etc., and once I had these made, I could drag and drop them on anything I needed and export the materials out from Substance 3D Painter.

I actually found getting the look of the base metal materials tricky, as I wasn't relying on heavy wear and paint surfaces to cover things up. I wanted to go for a more bare-bones approach to match the reference, which was hard to get the right layering of wear and variation. The look I ended up with was mostly concentrating on surface colour variation and subtle highlighting of edges and recesses. Lighting and shadow in the renders were how I could add more macro information. 

I also needed to consider that the kitbash parts would be used in all different orientations and scales, so I couldn't bake in any directional information into dirt or damage layers. While getting the look I wanted was tricky, having a few robust smart materials meant the base texturing process happened quite quickly.

Lighting & Rendering

While lighting is sometimes the last thing that people consider, for me, it was one of the first things I set up. I don't usually leave it until the end, which might not always be the best approach, but I think it really helps to learn what will be seen and what to focus on.

I took a more realistic approach while taking inspiration from the anime: harsh studio lighting in an abstract space. Studio lighting involves different cinematic techniques like rim, key, and fill lighting to really make the subject pop out from the background. I also kept the lights very grounded and neutral so as not to influence the colour in the scene and materials overall.

I used a lot of rim and backlighting to separate parts from the background and add depth. Wherever parts were silhouetted directly against the background, I made sure there was a light. Where I wanted more depth in the model itself, I added extra lights to make things pop.

To cut through the fairly monochromatic lighting, I placed smaller coloured lights to add pops of colour and interest. I deliberately restricted my use of colour to see what I could achieve and make the moments of colour more impactful.

Due to this project having a lot of moving elements, I also incorporated DaVinci Resolve into my workflow. While Blender can handle post-production comfortably, I wanted to push the realism and cinematic feel, which I knew DaVinci is really good at – adding subtle film grain and other effects that can be applied to static and moving shots easily. It was also the first time I rendered and exported in EXR, so I could properly colour grade my images.

An overlooked technique I also use is to take regular renders and track your progress through images. This guides the render and allows you to see where things are improving or where they may have regressed. This is especially useful for lighting, as you can view the images in thumbnail format, which gives a good impression of the balance and readability of the image. For this project, I experimented with lots of different lighting and colour scenarios, but ultimately landed back at my original vision.

Animation

Now, as I had a really strong blockout, I had mostly figured out how everything was going to move – or at the very least, where all the pivot points would be. This made things a lot easier when it came to learning how to create a basic rig using Blender.

For the pistons, I used bones and constraints so they behaved as expected when moving the asset as a whole. This meant all I needed to do was keyframe any movement to a root bone or controller, and everything would follow along nicely.

For the assortment of pipes, instead of using complex cable physics, I simply parented the spline verts to the rig using hooks in Blender so they inherited the same movement. I could also add additional noise to the spline to imitate a wiggle. The visual mesh is simply fitted to the curve to give the appearance of movement and physics without costing anything in render time.

The biggest reward for me was taking the time to learn basic rigging and animation in Blender to bring everything to life, even if it's not something I normally do.

All of the smoke and steam was actually one of the easiest elements from a technical standpoint, but it added a lot visually. It's all alpha cards and looping smoke animations. These could be created from scratch, but there are also plenty of places to download pre-made ones, such as Action VFX.

Conclusion

I used this project to continue developing my hard surface skills and push the detail and complexity of the assets further than I had before. Breaking down the complex machines into digestible parts that work together as well as on their own was a real challenge. I also learned much more about rigging and movement, which can elevate an asset to another level.

For anyone wanting to tackle assets like this, my advice is to extensively block out and plan exactly what parts are needed and how each mechanism fits together. Once that's achieved, replacing simpler meshes with final ones becomes a much simpler process and has a great payoff.

Also, don't skimp on the lighting and rendering of your assets. Lighting can elevate an already complex asset but can also destroy details and composition if not thought through carefully.

I'm always happy to share my techniques and answer any questions people have. Feel free to message me over on ArtStation!

Jake Woodruff, Senior Environment Artist

Interview conducted by Gloria Levine