Creating a Realistic Captain in ZBrush & Marvelous Designer

Character Artist Josh Wallace talked about designing Captain Jacq Lock and shared a lot of WiP photos.

Introduction

My name is Josh Wallace, I am a character artist from Namibia, which is located in the South West of Africa.

I have been obsessed with cinematics and realism ever since I watched the World of Warcraft and Star Wars Old Republic Cinematics years ago. This project has been a huge breakthrough for me, as this is my fifth year doing 3D and I finally feel like I am getting close to achieving that realism I have been striving for.

In terms of Studies, I attended a Character Artist Bootcamp at Game Art Institute (Now Vertex School) in 2018, recently completed the Characters for Film & Game Cinematics Course at CGMA, and am currently enrolled in Marlon Nunez’s Likeness program. Other than that I am self-taught.

As of this time I have very little to no industry experience. I hope to change that soon and perhaps someday work in VFX Cinematics and Film.

Captain Jacq Lock

I developed this character while attending the Characters for Film & Game Cinematics course under the guidance of Peter Zoppi.

I decided that, in order to get the most out of the course, I would attempt a similar character to the instructor. By doing this I would be able to optimize my workflow and receive better feedback.

I drew heavy inspiration from the pirate-themed show Black Sails, where I collected most of my references. I put together a rough Photobash concept in Photoshop and began the process on my character.

Face

The face went through many changes and iterations during the project. I started with a very crude base mesh and mainly used Tom Hardy for reference as I blocked in the main forms. I tried to stay on the lowest sub div for as long as possible. This can help to keep you focused on the important forms. As the sculpt progressed I started gathering references of Ben Affleck and Brad Pitt, to further establish the look and feel of the face.

I decided to add micro skin details very early on. Once the main forms were in place, I created UDIMs and projected TexturingXYZ displacement maps in Mari. When I completed this, I extracted three maps from the RGB channels and processed them in Photoshop to reset the mid values.

In ZBrush, I created layers for each displacement map before applying them, in order to have more control over the intensity of each map.

Adding skin displacement early on can influence the forms and sculpting decisions you make down the line, which in turn helps to bridge the gap between the sculpt and high-frequency details.

Once I had the face in good enough condition, I proceeded to manually sculpt in extra details, wrinkles, and scars.

It can help to adopt a non-destructive workflow early on, as the face might need changes down the line. In my case, it changed quite a bit.

Sculpting a human face is a complex process and can be difficult to achieve when out of context with the body. To aid in this, I recommend starting the Look Development process early on.

Once the eyes, hair, textures, and shaders have been added to the model with adequate lighting, you will be able to see the face in full context, and more easily be able to spot any needed adjustments.

Splitting the beard, mustache, chest hair, nose, and ear hair, and vellus hairs into their own separate XGen Descriptions, gave me more control over the hair. I then repeated the same process of grooming on each new Description.

I started by building the XGen Descriptions on top of the head geo instead of creating a hair scalp/cap. This process can take more time to set up, especially when using UDIMs, but saves time when further changes to the face are needed.

While trying to roughly match the ZBrush hair blockout, I laid down and shaped the guides, making sure to keep everything evenly spaced. Once the guides were in place, I used the density brush to control the spread and density of the hair strands.

Clump, noise, and scale modifiers with procedural textures and settings can aid in retaining a non-destructive workflow, and by using this technique I further developed the groom. I then utilized sculpt layers to manually tweak the shape, length, and width until I was satisfied with the result.

Clothes and Accessories

I started blocking out the clothes very early on, using Marvelous Designer throughout the whole process.

Marvelous Designer itself as a tool is not that difficult to learn, the difficult part lies in the patterns and how you fit them into your model.

My process for garment creation normally starts by gathering real-world sewing patterns, either on Pinterest or Google. It can be difficult to find the exact sewing pattern of the garment you are creating. A good alternative would be to rather search for patterns of similar garment types.

It is also extremely important to have your character model on the real-world scale before importing it into Marvelous Designer, as this can save time and prevent scale issues when switching between software.

Once I had my character model imported with patterns sewn in place, I started adjusting the patterns to better match the reference. I spend a lot of time on this stage, as I prefer to push the garments between 70% to 80% completion, inside of Marvelous Designer, before taking the model to ZBrush for final touch-ups and detailing.

For the jacket, I created a general pattern and then cut that pattern into strips to better match the look of the reference. This posed quite a challenge during the retopology stage, but gave more control over the mesh in the ZBrush sculpting and detailing phase.

As for the accessories, everything was built in Maya using basic modeling techniques, with a detail pass in ZBrush at the end. For the necklace twine, I used curves to shape the geometry with curve warp, twist, and bend modifiers.

Texturing

I used a combination of Texturing XYZ maps and hand painting techniques on the skin. Using Mari for texture projection and Substance Painter to hand paint extra warm tones, veins, and liver spots.

While painting skin, I mainly use Dirt brushes with a low-flow setting. These brushes give a more organic look and really help to blend in the layers.

The clothes and accessories were textured using Substance Painter. My texturing workflow mainly consists of fill layers and heavy use of masking.

I normally start with three base layers, which consist of a base color, a darker shade, and a lighter shade.

In the masks, I use two fill layers with Grunge Maps. The first as a base and the second set to multiply. Finally, I add a paint layer on top for any needed manual adjustments.

This method adds a more realistic breakup to masks and helps to remove that “procedural” look.

From there I build up the layer stack, using different subtle color, roughness, and height variations until I am satisfied with the look. The entire process can be very iterative, often going back and forth between render and texturing.

This iterative workflow proved to be extremely useful while texturing the leather jacket.

I struggled to get the right leathery feel and ended up having to redo the texture a few times.

Being able to see your textures in a well-lit rendered scene, can definitely give more perspective and help point out certain errors and needed changes.

Rendering and Post-Production

The render and lighting setup were quite simple.

For the main cinematic shot, I only used a single HDRI with a 135mm camera lens.

HDRI Haven has thousands of really great, high-quality HDRIs and best of all, it's free. I ended up using the Old Tree in City Park HDRI, which can be found on the website.

Depth of Field can really add to the realism of the shot, but it can be quite challenging to set up correctly.

An easy way to combat this is to use the Measuring Tool in Maya. By parenting the start of the Measure Tool to the camera and pointing the end to the Eyeball of the character, you can get the exact distance needed to set up the Focus Distance for the Depth of Field. Then in Maya’s Node Editor, you can bind the Measure Distance to the camera’s Focus Distance, giving your accurate real-time updates to the focus, whenever you move the camera.

For the portrait shot, I used a basic three-point light setup, with a studio-lit HDRI set to a low intensity in the background, for ambient light. I set the camera to 85mm for a more portrait look and used the exact same Depth of Field trick.

On the Post-Production side, I used Photoshop and Lightroom. In Photoshop I overlaid some subtle dust mote and scratch grunges, with a lite chromatic aberration effect on top to add to the cinematic look.

In Lightroom, I cropped the images and used Photography color grading LUT presets to add more dynamics and contrast to the images. I made sure to tweak all the settings, adding subtle grain and vignette to match the look I was after.

Advice for Beginners

My advice to beginners would be to focus on the fundamentals. A good understanding of shape language, silhouette, human anatomy, texturing, and lighting are crucial to breaking through the uncanny valley. Creating a believable character is no easy task and requires thousands of hours of practice and hard work.

In my opinion, it is extremely important to look at the right learning materials. Luckily, nowadays there is an abundance of great tutorials and courses out there explaining and teaching each process.

Some good sources to look at are Anatomy for Sculptors, Flipped Normals, CGMA, and Pixologic’s Youtube Channel.

Josh Wallace, 3D Character Artist

Interview conducted by Arti Sergeev

Keep reading

You may find these articles interesting