Illustrating a cyberpunk scene: composing, paintover & post-processing

Introduction

In my opinion, the most important part of illustrating a scene is to have appeal. You want your piece to grab attention. Good composition and lighting is crucial if you want to achieve that. I will spend the majority of my time and thoughts on composing this image and less on minute details like textures, scratches, freckles, hair strand, and so on. Note that composing this scene is very trial and error – there were a great deal of experimentations, contemplating, and getting rid of things that don't work.

Generally when composing, I focus my camera on an object at an angle of my choice, then render it, then drag it into Photoshop. Here is a brief description of the pipeline of my paintover phase: I will be laying in subtle gradient lights/shadows, fixing the lights and shadows, then paint in cast shadows, ambient occlusion, and specular highlights. Then add volumetric fog, blooms, and vignettes.

Final illustration, about 45 hours of work in total, 9000px across, lots of layers

Posing and presenting the main character

There are two stages when posing my character. The first is that I pose the soft surface part of her in ZBrush, then export them into Maya and pose the rigged hard surface parts (arms, gun, and so on). When posing characters, I like to have them look graceful. To achieve this, I want their torso to twist into an arc – making a C shape. I first turn her torso in the y-axis to her right side (don't pose the legs yet). Then twist again in the z-axis towards the same direction. Then finally rotate her in the x-axis – you can see that I eventually tilted her head downwards.

One important thing in posing is that you should rotate the whole character back to where your character's head becomes perpendicular to the ground regardless of how his/her torso is posed. This makes your character look more natural (unless your character is lying down or falling).

As for her legs and arms, it's trial and error – experiment with different arm/leg poses to see which version looks the most appealing.

If you have no idea how you want your character to pose, just do research

Composing the robot

For this robot, I didn't actually pose it before rendering, it is unnecessary because I want to break them into sections and render each separately – this allows me to have better control. In KeyShot, I will then turn, zoom, and pan the camera here and there to get my desired angle.

I composed the head first, it needs to be somewhat of a 3/4 view and tilted down as I really hate anything that is presented either from purely the front or side.

I then modeled a Gatling gun really quickly and placed it in front of the head. Then added an extra platform on the side of the head for the main character's left leg to step on. Next I added the arms of the robot; one of them needs to be stepped on by the character's leg too. Then followed the body and front legs.

Composing the police and floor

Before I rendered the feet of the robot, I stopped going further because I knew that there would be two authority figures in front of the robot. They would be overlapping the feet so there was no point in attaching the feet for now. We will go back to the robot after the authority and floor are dealt with.

When posing and rendering the police, it is done similarly to the main character. I pose the foreground police to aim his gun at the robot, ready to fire. The second police figure is posed to look like he is tumbling, mostly because he was shot in the head by the main character. His torso will need to twist away from the gun in terms of the x, y, and z axis.

After I am done with the police, I proceed to compose the floor. Note that I didn’t worry too much about the floor's perspective needing to diverge into the scene’s horizon. I just placed them there and called it a day. I believe that your artwork will look decent as long as the composition is good even if the perspective is wonky.

Composing the background buildings

I compose my background based on intuition. If placing a building on one spot feels right, I'll keep it there and move on to the next. It is important to first compose the ones closest to the viewer. I also place the temples first because they are more prevalent. The skyscrapers are treated as fills.

In terms of perspectives, I don’t try to make them perfect, they only need to look okay. Again, I try to prevent the buildings to be shown directly on their front or side. They all need to be in 3/4 view.

Fill the floor

I initially wanted the scene to be on a road with cars here and there, but I thought it would look better and more immersive if set in a construction zone. I first try to hide the bad perspective between the far edge of the floor and the background by placing barricades over them. I then added a lamppost on the left side for more immersion, and then a vending machine on the right-side for more color variations.

I put a third authority figure in the background because when I create art, I try to have things come in threes. He is a copy of the policeman in the foreground. Next I added a bin and some garbage bags to the lower left corner. The plastic bags were made by wrapping two square clothes around a sphere in Marvelous Designer. I then added wires, rubble, newspapers, and traffic cones on the floor. I don’t like photobashing so these objects are modeled in 3D.

Neon signs and minute details

After the majority of composition is done, I paste in supplementary details like stickers, road signs, drawings, texts, and logos around the scene because that's how real life is. Using the skew and resize tool, I place neon signs over all the buildings. The neon signs need to be in a variety of colors except blue because this artwork is already blue by default.

Small detail compositions like this should only be done last. After this phase, we can start doing paintovers

Gradient lights and shadows

This is the first stage of the paintover phase. I strictly will not begin this stage if the composition phase is not complete. I believe that being systematic in your workflow keeps you from being overwhelmed. My logic about gradient lights and shadows is that surfaces very close to a light source will look bright. Then as it transitions away from the light source, it becomes darker. In this case, the light source is around the top right side of the scene – making its surrounding objects brighter. Objects at the bottom-left corner will be filled with shadows due to high proximity from the light source.

So to simulate this kind of transition, I use the gradient tool set to Overlay and 'Foreground to Transparent' setting, then set the color to light blue at 10% Opacity. I will drag the gradient from the light source away on all the objects in the scene. As for the shadow, I set it to Normal and black in color.

The blue lines represent the direction of light gradient and red lines are the shadow gradient. This phase is crucial because it 'balances' the artwork

Volumetric fogs

I use volumetric fogs to separate the background from the mid-ground. This allows the silhouette of the main subject matter to pop. Because I work on many layers, I put in volumetric fogs starting from the bottom layer, then slowly work my way up. As the fog builds up, the farthest objects will get brighter and lose details. What's left is the mid and foreground objects, they are least affected by the fog – making them look a lot more prevalent than the background.

Usually when I paint volumetric fogs, I use Lighten layers for day time scenes. For night scenes, I use Darken. However, for a complex night time scene with bright neon lights such as this, I decided that the best layer to use was ‘Linear Dodge.’ I picked a dark blue color from the sky, and then set the gradient tool Opacity to 10%, then dragged that color out over every object in the scene.

Fixing the lighting

Due to the objects here not being rendered all at once, there are bound to have lighting inconsistencies in the scene. The goal of this phase is to make sure there are no shadows in the lit area, and there are no overly bright lights in the shadow area. If I spot these inconsistencies, I will fix them with my paint brush.

You will see that now there is a cast shadow on the robot’s head made by the main character. There are also cast shadows on the floor and made by the robot.

I also take this chance to add Ambient Occlusions and specular highlights on areas that requires it. Ambient Occlusion comes in very big, medium, and very small sizes – I only paint in the first two.

Try not to put Ambient Occlusion in lit areas because that will make your artwork look flat. Use specular highlights instead if you want to show form

Minor fixes, fogs, and bloom

At this stage, I try to scan for any other flaws still present in the artwork – starting from the lowest layer. As I move upwards, I'll also turn the fog layers back on and tweak their opacities if required. I also use the gradient tool with the Linear Dodge setting to simulate bloom effects. I only do this on bright surfaces and neon lights. The color of the bloom is determined by the color of the object you are blooming – but at a higher saturation.

In my opinion, the bloom effect is crucial because it lessens the stiff look of my artworks

Post processing, lens flare, vignettes, and match color

My goal of post-processing is to make my work more appealing. Usually artworks of night time scenes with small but very bright saturated light tend to look very nice. I first lay in vignettes on the left and bottom side of the scene, making the area darker, because this should be the area my viewers look at last. The vignettes are done with gradient tool with pure black color and 5% Opacity.

Then I add a large bloom effect over the main character, by shining more lights on her, the viewer hopefully will look at her first. The bloom effect's layer is set to Linear Dodge. Finally, I will go onto the internet and look for a nice photo that is closest to what I wanted this piece to look like (in terms of color and value range). I will paste that photo onto my file then make a new layer of my flattened artwork, then use Match Color to have it match with said photo. My artwork now became significantly higher in contrast and more saturated.

This cyberpunk illustration is now done, I hope you learned something. If you have questions, feel free to message me on social media or email me.

The first image is before adding vignettes. Second one has vignettes added. Bottom image is after matching color

Always finish your work

A trend I notice in the art community is that good artists focus on finishing their work – all nice and well-polished; they paint with the aim of strengthening their portfolio. And artists who improve slower tend to begin an artwork but never finish them – they lack discipline and gave up their piece when they encounter a seemingly unsolvable problem. Sadly, unfinished work cannot be put into a portfolio. I recommend new artists to just hammer through their work and get it done. Grit is important in this industry. It feels horrible spending 30 or 40+ hours on a piece you don't like but you'll improve faster by being hard on yourself.