Ashli: Making a Mielgo-Style Character in 3ds Max & V-Ray

Introduction

Hey, I'm Math! I'm currently a lead 3D artist at VOLTA, a studio based in Québec City, Canada that focuses on concept, illustration, and benchmark assets. I've been in charge of most rendering work there for the last few years, and I've been able to grow immensely as an artist by tackling the widest possible range of visual styles in the most varied and challenging situations. Despite being a generalist on paper, the incredible skillset of the somewhat small 3D team made it possible for me to focus on the things I like most: surfacing, lookdev, lighting, and rendering. I've been lucky enough to do so for many amazing clients, from juggernauts like Microsoft, Ubisoft, and Hi-Rez to local partners in crime like Chainsawesome. In 2019, I also developed two comic-looking animated sequences for Mr. Sunday Movies, host of Australia's best podcast The Weekly Planet, as part of a Star Wars retrospective. It was very exciting, but it also meant completely blowing up my comfort zone since I was taking care of every single creative step.

Becoming the Artist

Getting into 3D rendering kind of happened organically. As a kid, I was completely obsessed with drawing and watching cartoons. I spent countless hours in Mario Paint for the SNES before moving on to Macromedia Flash in my teenage years, just as the DIY animation wave started hitting the newly mainstream Internet. Following those two decades of curiosity and some formal education in arts and cinema, I completed an up-and-coming baccalaureate in animation at Université Laval that covered not only the basics of 3D but also 2D, rotoscoping, stop-motion, screenwriting, etc. I used all available credits to take additional classes in photography and ended up expanding my passion for it with 3D illustrations as I discovered that — unlike many colleagues — I didn't enjoy modeling at all. My focus more than ever rested on things like lighting, mood, and composition. The appeal of 3D was clear though, but I much preferred to work with existing things and try to showcase them at their fully realized potential.

To me, this is what lookdev is at its core.

Look Development

Look development means exploring and ultimately defining aesthetic choices and technical parameters into a unified purposeful creative vision.

In more concrete terms, we're talking about establishing a visual style. This is done via a phase of exploration driven by intent, aesthetic appeal, and technical proficiency to determine the exact desired look and rendering processes of a project. If it's pure photo-realism, then, well, there isn't much to invent. Stylization on the other hand is harder to define because it covers such a broad range of possibilities. In the field of 3D, its foundations undoubtedly rest on the principles and technologies developed for realism, but it branches out in countless creative directions. Lifelikeness is shattered, as the search for fresh aesthetics becomes the primary goal.

The possibilities are, as it rarely is the case, infinite. They need to be determined, however. That's when look development comes in.

Whether it's done with assets that you've created yourself or that were provided, a ton of critical decisions remain to be made to establish a strong visual language. Those models and textures need to come together under a singular artistic vision by crafting shaders, lighting, effects, mood, rendering conditions, etc. Not only do they need direction, but also consistency throughout the project.

Establishing these qualities is pretty much dictated by four things:

• Personal creative decisions.
• Art direction, yours or someone else's. This can also include pre-production material from the project, like concepts, 2D visual development, mood boards, etc.
• Key visual references/influences (external to the project).
• Overall just what the nature of the project aims for. This relates closely to the last three points, but also broader themes or feelings that the project may benefit from.

In my opinion, this task makes the greatest use of an artist's sensibility because, in the end, things just need to look right. It requires great attention to detail as well as a solid understanding of forms, colors, contrasts, textures, light, space, and most importantly how these elements all work together to be compelling and help to tell a story. This idea can even extend to other narrative tools like framing and camera lenses. Good observation skills are definitely a plus here too as a lot of nuances from real-life might need to be included in the work, either as is or in some sort of reinterpreted way. It's fundamental to play with textures, shaders, and procedural nodes. Mix them up! Test how their numerous parameters/properties can be exploited. You can adhere to hard PBR principles or totally play mad scientist here. See where this iterative process should or shouldn't take you. Don't even hesitate to start from scratch when necessary; when that happens, it means you've already discovered an important part of the path, like an author re-writing the first draft. It’s also crucial to experiment with different palettes, atmospheres, and placement of lights to create various moods and ultimately find the right one(s). A coherent whole is being created here. Take the time to find it.

Oh and remember, all of this must be executed with great technical proficiency to ensure that things are properly developed and consistent in virtually all conditions, plus comply with the project's limitations. Finally, if you really want to push the boundaries of your creativity, mastery of your tools is essential. You need to be able to craft what you want, not have any aesthetic ambitions restricted by your abilities with those tools.

Ashli

The base model for this character was originally part of the library that the Academy of Animated Arts provides to its community. Its two amazing founders Michael Tanzillo and Jasmine Katatikarn were both senior lighting TDs at Blue Sky Studios until the sudden unfortunate shut down by Disney last February. With AAA, they are totally dedicated to providing quality online CGI courses, share a steady stream of valuable feedback, and nurture their community of students/artists/enthusiasts with, among other things, really fun monthly challenges. February being Black History Month in the USA, they took the opportunity to provide us with Ashli, courtesy of Danny Williams, to favor representation and test our abilities to light characters with darker skin tones.

The last monthly challenge I participated in gave me quite a nice image to showcase as well as the opportunity to create a video breakdown of my workflow. The reception was very positive and I loved to share my process with the group, so I thought I would repeat the experience as soon as a new lighting challenge would inspire me. It ended up being Ashli.

When I started playing with the asset in 3ds Max, my goal very early on became to stray away from the intended look and revamp it entirely, instead of just tackling some lighting task. More often than not, I'm tempted to approach everything as a full-fledged illustration. Furthermore, as I've hinted at before, lookdev is now closely intertwined with all my rendering work, so this project certainly wasn't going to be an exception. I was first going to reassess everything.

The two main elements I decided to depart from were:

• The face visor. I felt it looked tacked on and was preventing a pleasing lecture of Ashli's facial features. This is always a crucial point for me.
• The emissive neon highlights on the clothing and accessories. It definitely brought too much attention to the neck area. Also, I felt that having this was a bit easy and obvious considering that last year was dominated by Cyberpunk imagery. I wanted something less techno and edgy. Something a bit more warm, discreet, and mature. Something that wouldn't fight with the appeal of the face or get lost in the bright image I wanted to achieve. I also remembered some tips about concept/design from Rael Lyra, one of the most wonderful and talented colleagues one could ever have. Those insights from long ago definitely convinced me to drop the glowing bits and even to rethink most of the clothing.

In addition to those things, I proceeded to:

• Replace the provided eyes entirely with something that had more depth and realistic details. Doing something stylized is no excuse to cheapen on the eyes — they need to feel alive! Windows to the soul and whatnot.
• Edit the nose geometry so the nostrils weren't so wide and prominent.
• Edit the mouth slightly to get some more nice-looking curves.
• Add multiple new dreadlocks falling on the side of her head (a design choice to create asymmetry + something valuable for lighting. Keep reading for a detailed explanation).
• Add the round glasses to replace the visor (a design choice + something that supports the focus of the image. Keep reading for a detailed explanation).
• Edit the collar geometry, UVs, and textures to fit my intentions. I even applied new displacement maps to create some folds and waviness in the fabric so it felt less rigid. Again, even for stylized projects, it's important that things are easily recognizable by retaining some key properties that they would have in the real world.

It may seem like a lot but remember: everything must fit into a unified purposeful creative vision. Once you start exploring, who knows how many changes will be necessary for that to happen.

With those things taken care of, we're more than ready to discuss the surfacing and lighting of this piece. Those don't happen in the void, however, so let's first take a look at some inspirations.

Inspiration

Most people must have recognized in my image the influence of the 2018 animated masterpiece Spider-Man: Into the Spider-Verse. Let me level with you for a second; I cannot even find the words to properly describe how brilliant this movie is. Not only is it a rare treat to get a super-hero film full of excitement that tackles genuine human themes as well as it does comedy, but the visual style developed by director/animator Alberto Mielgo is an absolute delight that has no right to work so flawlessly as it does in CGI. Huge shout out to the art, animation, technical, story, and direction teams here. They changed the game legitimately.

The thing is, for me this movie particularly hit home in a personal way since I've always been interested in bringing 2D qualities into my 3D work. It simply blew me away and kinda made me question my whole creative endeavor. That's real inspiration right there.

More than a decade ago, I was already experimenting with toon aesthetics: flat shading, harsh lighting, different kinds of outlines, hatching, hand-made drawings mixed with 3D, screen-dependent texture mapping, etc. I was a naive lad at the time with none of the artistic or technical proficiency I have now, of course, so in retrospect let's just say that the results ended up varying from decent to questionable. Still, I really felt that 3D had the potential for so many crazy art styles and yet it seemed people were only starting to scratch the surface compared to 2D illustration or animation. This feeling never really went away despite amazing things being done in the mainstream, in independent media, or at an individual level. It's always a joy to stumble upon something that looks genuinely fresh.

If we consider these three projects above to be part of a single creative path evolving throughout the years, it seems that it took me 10+ years to be proficient with colour and able to create the bold unique style that was laying around the back of my head for so long. Sure, it wasn't there entirely consciously, and not attempted that often either for that matter, but it's definitely the kind of aesthetics that lived within me. Am I done with it now? Probably not; time will tell.

Obviously, not only my own journey and Spider-Verse played a part in developing the look. A ton of artists inspire me on a daily basis! Below are a few images that make up my morgue file. Whether I've been carrying them for 10 years, 1 year, or 1 week, they all contribute and have value in their own way. They help to orient my objectives. We can see in those a few things that I remixed into the illustration, including rough cartoon outlines, a harsh rim light that punches the silhouette of the character, hair that's almost entirely a dark mass, halftone patterns that accentuate shadows and highlights, hints of painterly textures, abstract shapes creating the out-of-focus environment, etc.

Shading and Lighting

Getting Started

All this part of the work was done with V-Ray. I've been a massive advocate of that renderer for many years now, and each new release just reaffirms my love for that software. I find it incredibly versatile and its array of tools just keeps getting better all the time.

After working with a linear workflow for virtually forever, I switched last year to an ACES workflow. Developing projects in that particular color space turned out to be a revelation; light seems to transport and bounce way better, with colors and contrasts giving a result more vibrant than ever. I'm not exactly sure why, probably due to its very wide gamut, but it seems like the lights can be cranked a lot stronger than before without the surfaces becoming harshly blown out. These over-exposed areas just seem to have more energy now, which looks particularly pleasing when the bloom is introduced or when they are out of focus. Chris Brejon called ACES "his personal obsession of 2019" and I 100% understand why. I highly recommend this article if you want to dig deeper on the subject.

The 3ds Max, V-Ray, and ACES parameters of my working environment are rigorously set to be physically correct at all times by the way, which I think is the best starting point to create things with accuracy whether they're a cartoon or not. Doing stylized 3D imagery certainly means knowing when to be creative and when to adhere to important established principles, and for me, having a proper well-calibrated workspace is crucial. It also makes everything a lot easier to troubleshoot when you don't mess around with core settings on a project-by-project basis. My advice is to let the creativity happen elsewhere!

Shading

Shading started very casually with a SSS material (VrayAlSurface) applied to the face just to get a better feel of the asset and see if I could easily obtain some nice-looking skin. I did get just that, but as I said before I wanted to question the intended look, so for some reason going for a slick Pixar-like aesthetic just didn't seem right this time around. I then started experimenting with something more graphic, closer to cel-shading, along with versions of the textures that I edited with obvious brush strokes to accentuate the 2D intent. This very flat look sure was interesting, but at the same time, it didn't feel quite right either as some of Ashli's main features like the nose and lips didn't read so well anymore. By blending a few test-renders in Photoshop, I came up with a middle ground that looked quite promising. The cartoon vibe was definitely there, but I also had really nice sub-surface scattering and specular properties showing up to better define the shapes. I then reworked my shading graph in 3D to reflect that visual result I had just obtained. After again a few iterations to fine-tune the blending and to tweak things like the density difference between lit and shadowed areas, I felt I had something really good in my hands. The trick was now to perfect the visual language of my illustration by gradually refining it while developing all the other elements surrounding that initial work.

The exploration process:

The main shader for the head (simplified a bit here to better convey the idea).

The primary trick for the toon part is the FallOff node in light/shadow mode, which basically acts as an on/off switch with its behavior depends on the illumination received in the scene. Being lit activates strong emissive properties mixed with more SSS while being unlit activates low emissive properties mixed with less SSS.

The threshold to filter between lit and unlit was manually determined with a curve inside the node.

Lighting

Before moving on with anything else, I decided to make sure that my current lighting was more or less what I would carry until the end. I had come up with it while developing the main shader and I thought it worked pretty well. It's barely a variation of a 3-point setup, yet its simplicity was giving something soft and pleasing that I was sure wouldn't cause the piece to become overloaded considering the unorthodox lookdev. In that sense, I approached lighting more like a portrait instead of something utterly dramatic or dominated by the context of a scene. Don't get me wrong, I sincerely believe that lighting is a conscious act that needs to evoke/support a narrative, but in this case, showcasing the totality of the lookdev was more important than forcing something too complex into the equation.

Generally speaking, lighting for characters must fulfill three requirements:

• Create a mood, to support an implied narrative via color, tone, and intensity.
• Sculpt visual shaping, to accentuate the silhouette and volume of the subject.
• Direct the viewer's eye, to heighten readability and the impact of the composition.

Knowing these, I felt confident to keep going and took a mental note to mockup a background as soon as possible to complete the mood and composition.

I really wanted this image to be bright and colorful, and I made sure that the harsh orange kick light was there to make that statement. Its strength, saturation, and sharpness are unequivocally intentional and the style as a whole would arguably not work as well as it does if that single element wasn't there. It's not something that looks particularly natural; it belongs more in the world of graphic novels than anything else. It almost becomes another type of outline by itself, and I needed that to accentuate the shading work. Such a stylistic light is a very effective device to shape a silhouette and instantly bring the viewer's eyes along that line. It essentially fulfills all three requirements mentioned above at the same time! In this image, it also acts as a counterforce to the sharp shadow on the opposite side of the face to frame the entire lower half of the head in a well-balanced manner.

Like I mentioned in a previous section, the new hair strands I put in place ended up playing a crucial role to support how effective the lighting was. Sure, treating them as a dark mass with very few bright details was contributing to the visual style of the image, but most importantly it created a dark region right next to the strong orange kick to ensure that this light would have the maximum possible impact. This is in essence a principle sometimes called counterchange, or light over dark, or who knows what else. It basically means that readability is at its best when a shot is consciously lit/composed so the brightest areas of a character contrast with darker portions of the background and vice-versa. Obviously, the hair isn't part of the background, but the theory can still apply between the light and this darker secondary element.

In fact, the whole illustration was designed to have either this type of juxtaposition and/or color contrasts between as many elements as possible. It was made to be more pronounced at the focal point of the image and gradually toned down as we move away from it.

More Shading

Shading everything that remained once the main recipe was finalized happened relatively swiftly. A few things received their own relatively simple material, but all the clothing just got a variation of the skin shader. Its SSS component was replaced with a standard material though, to make sure the fabrics showed some core differences with the face that also made sense with their real-life counterpart.

The flat part of the shader also received an edit to create an additional bright fresnel in the illuminated areas, which was introduced to enhance the impressions of sheen and folds. Contrary to good PBR practices, the metallic details were given a non-black base color and almost mirror-like reflections, giving a result similar to the way they could be approached in a 2D illustration. The shiny leather got a bit of that particular treatment as well, although in a more subtle way. We can see in some areas how that created sharp and somewhat geometrical highlights, which I thought were quite interesting.

Focus

By approaching this lookdev exercise as an illustration, as a portrait of the character, it was critical for me that the focus on the face was flawless. This can be done successfully by making everything (color, contrast, composition, DOF, etc.) work together to bring the viewer's attention to the eyes. If at first glance they instantly become the central appeal of your image, then you've got it right. This won't always be what you want, there are exceptions, of course, but we, humans, are wired to be drawn to faces, to read their expression, and interpret meaning, and most of that happens through the eyes. We are naturally seeking them, all the time, even by mistake on inanimate objects or in random shapes. So I would say that whatever you do, it's definitely one of the most important aspects to consider when working with characters. Lighting, as always a tool for storytelling and visual shaping, can also play a significant part in how they are interpreted.

Deep, glossy, gem-like eyes are incredibly appealing. Eyes hidden in the shadows create mystery. Only one visible eye pinpoints the focus even more. Two different eyes tell a story. Closed eyes break the human connection in favor of internal emotions. Having too many eyes is unsettling. Having no eyes is horrific and literally alien-like.

Considering the direction I took with Ashli, the choice was clear to make her eyes all big, bright, and pleasing!

Since I got rid of the tacky visor of the original asset, the eyes already had all their chance to shine, but I thought I could support them even more effectively. I've been reading a lot of Spirou from the Tome & Janry era and noticed how the eyes of female characters were always cleverly supported by numerous curves. I shifted a lot of decisions around that idea, which also justified the inclusion of the sharp round glasses. This meant that the dark and vivid iris, framed by the bright white of the scleras, themselves surrounded by the stylized pitch-black eyelashes, were now each reinforced by one more well-defined ellipse. If we add to that the shape of the eyebrows, the arrays of dots painted on her face, and a few more curves put in evidence in the image. The eyes then became consistently supported by this scheme of multiple concentric curves, all enhancing each other's contribution to the desired focus.

The two materials of the glasses ended up being the only "standard" shading solutions in the piece. The metal of the frame was kept entirely PBR (black diffuse + high reflection value) which created the desired mix of contrasting dark and bright arcs around the eyes. As for the glass, I made it slightly tinted and gave it a bright purple hue that would contrast with both the warm skin and orange kick light, yet would harmonize well with other elements in the piece. I cheated and made its IOR value higher than physically correct glass though, which created a magnifying effect in the refraction that not only made the eyes a bit bigger but also warped the harsh lighting on the face into even more supporting curves.

Effects and Post-Production

Finally, now, there are two of the most important aspects of the visual style that I haven't mentioned yet: the outlines and halftone effects. They were achieved in a few different ways, both in 3D and in 2D.

For the outlines, four sets of them are working together:

• Those, generated by the shaders I developed. They are based on a FallOff node and vary in width depending on the curvature of geometry. The hair shader also has a special screen-dependent hatching effect (where the strands are occluded by other strands) to simulate a fuzzy quality in a graphic way.
• Those, generated by the renderer (V-ray Toon effect). They are essentially simple continuous lines occurring based on geometry and the parameters I've set.
• Those I've drawn by hand and included on geometry (cards). This turned out to be a bit trivial so I gave up the idea after doing the eyebrows. However, it proved that the same result could be achieved entirely in the 3D scene instead of relying on post-production, as long as the cards never received light or cast shadows.
• Those I've drawn by hand in 2D / post-production. For the time being, it just saved some time to draw directly on the 3D render.

The four sets of outlines. From left to right: those from the shaders, those from the renderer, those on cards, and those drawn in post-production:

The halftone effect on the other hand is there to accentuate the highlights and shadows of the character. I decided to add it in post-process instead of going with a screen-dependent texture in the 3D scene because I wanted its outer borders to feel uneven as if they had been naively clipped. I achieved this by masking the patterns super quickly with a cheap mouse instead of using a tablet, which gave the wonky look I was looking for. This gross execution in the small details mixed with the hand-drawn lines and the stylized surfacing work just felt totally cohesive.

The bright and dark halftone patterns complementing their respective areas of exposure. The bright one was actually tinted yellow in the final version to better suit the mood, while the dark one remained as shown here.

All the effects put together (exaggerated):

After being done with the final render (noise-free, unclamped, and uncompressed 32bit EXR), post-production didn't end up being too fancy aside from the usual compositing of render passes (AOVs) and some color corrections. All the important decisions had been made prior to this step — including in various mockups — so honestly, not much remained to do. A great deal of consideration was once again put into the eyes/glasses, I added the abstract background (a mix of photo editing, maximum filter, painting, and colour grading), and I just made sure that the exposure and overall readability of every element were perfect.

The last touch was to generate some chromatic aberration in the areas of the character that were out of focus. It was kind of easy to integrate it nicely into the image: the green hue (offset to the right) balanced well with the hair beads and the left side of the background, while the magenta hue (offset to the left) fitted just as naturally with the glasses, rim/backlight, and right side of the background.

A few final tweaks later, and just like that, I was done!

Final Words

For most of my life, by many standards, what I created wasn't great. There's no shame in that. That's just how it works: it mostly sucks until it mostly doesn't.

Then someday you check your inbox and maybe on that particular day the job opportunities are pouring in, or your followers are exponentially growing, or the head of content at 80.lv wants to conduct an interview with you, because somehow, with time, your hard work got noticed. It just never happens right away. In retrospect, you realize that it's worth the hassle and that you're truly blessed if you can make a living with something you're also passionate about.

So if your guts tell you to keep going on the days when you doubt yourself, if you're truly dedicated to your art, then there's no other path but to constantly improve and be proud of what you achieve. Keep learning. Find inspiration. Curiosity, perseverance, and resourcefulness matter more than anything else.

That's what I want to conclude with. Thanks for reading all this. Hopefully, I was able to bring a little something positive into your craft.

Cheers, dear reader!

Mathieu Mathieu Grenier, 3D Character Artist

Interview conducted by Arti Sergeev