Making Of 'Abyssal Princess'

Introduction

This image started out as a lunch break exercise, but I liked where I was going so I decided to take it a bit further. In this article I'll try to reveal most of my workflow for this particular piece, covering all the main steps to reach the final image.

Concept

The main concept was something I'd had in my mind for a while, inspired by the Axolotl salamander (Fig.01).

Fig. 01

Fig. 01

I heard of a warrior challenge for Mudbox. I couldn't officially participate since I was focusing on developing my ZBrush skills at the time, but I decided to do it for myself anyway.

I chose to create a female aquatic warrior character riding a giant manta ray. Her gills would be exposed, as in the Axolotl salamander,r and would help to define the female gender of the character by giving a sense of waving hair.

I had actually done some quick sketches for this character before and that's probably the reason why things came out so easily (Fig.02).

Fig. 02

Fig. 02

Base Mesh

Since it was supposed to be a ZSketch of one hour, I started a new base mesh from a simple ZSpheres structure. Basically, I used one main ZSphere for the base of the neck, with two ZSpheres to the side for the shoulder area and one going up for the neck. Then I added another ZSphere for the base of the head, followed by one main skull ZSphere and another for the mouth/nose area.

To the main skull ZSphere I attached the gill structures and the eyes, as well as more in the back of the head. This was all done with symmetry turned on (Fig.03).

Fig. 03

Fig. 03

I used the default settings for the Adaptive Skin and at this point I wasn't too worried about the topology or the polygon count. I knew that if needed I could solve most topology issues by just dividing the model or inserting some edge loops (Fig.04 - 05).

Fig. 04

Fig. 04

Fig. 05

Fig. 05

Sculpting

For sculpting I find there are many different ways to get what you want. For me, it essentially depends on how well defined the concept/idea is. If I know exactly where I'm going then I'll probably have good guidelines for the main shape/silhouette/feel of the character, and gradually working my way up the subdivision levels will give me more control and ensure I never deviate too much from the goal. If, on the other hand, I just have very general guidelines (an intention or even nothing at all) for what I want to do, starting with a lot of subdivision gives me more creative freedom until I reach the point where the sculpt starts to tell me where to go.

In this case, I had a good idea of the character I was going for, as you can see by the quick base mesh. So I started working with the mesh after dividing it two times. It was enough to get the main shape defined.

As you'll hear lots of times in different tutorials, it's great to use the Flat Color material to check your character's silhouette with Perspective activated. At this point I mainly used the Move and Move Topological brushes, as well as the Standard brush (Fig.06 - 07).

Fig. 06

Fig. 06

Fig. 07

Fig. 07

After this, I quickly started needing a lot more geometry to work with because of my quick base mesh and its poor topology. So I divided the mesh three times and started to define a bit more than just the main shape/silhouette of the character, meaning the bone structure and some main muscles, as well facial features.

When you're creating some sort of being, I find it's very important to make sure it has some recognizable human-like features. Otherwise, people won't be able to relate to your character.

So I started establishing some of those features for my character, including the eye cavity position, the shape of the "nose" and the location of the ears. At this time I was assuming that, like the Axolotl Salamander, it would be an amphibian creature to justify some of its humanized traces, like the nostrils and lips.

After dividing the mesh this much I started using some different tools like the Clay, Clay Tubes and Standard brushes with some tweaks on the Modifiers slider (Fig.08 - 09).

Fig. 08

Fig. 08

Fig. 09

Fig. 09

One more division and I started adding some more detail. At this point I had 3.5 million ActivePoints and started using some of the most basic circular alphas with the Standard brush to get some smooth lines (Fig.10 - 11).

Fig. 10

Fig. 10

Fig. 11

Fig. 11

I also added some spheres to sculpt one of the eyes and its membranes/eyelids. After finishing one of the eyes I just used the Subtool Master to mirror the other eye. As the character was getting a sad/angry look, I increased its lines and its mood a bit to make it more empathic and a bit more feminine (Fig.12 - 13).

As strange as it may seem, at this point I decided to jump into posing to see if the sculpt was expressing what I intended it to be or if I needed to enhance some of its features in order to get there.

Fig. 12

Fig. 12

Fig. 13

Fig. 13

Posing

The pose is very important. A good sculpt may become completely void if it lacks expression, character, personality... and I think a good pose is the most effective way of bringing your characters to life.

The pose of this character is subtle and extremely simple. It was done with the Transpose Master (Fig.14).

I tried to give her a more feminine look, lowering her chin, straightening her neck and slightly turning her head over her shoulder. I think this gives her a mysterious look suggesting a fragile and suspicious character.

Fig. 14

Fig. 14

Unwrapping

I know this really isn't the best way to go. I should have done the unwrap and the painting before the posing, since ZBrush uses spatial symmetry and not topological symmetry. This could have saved me some time. But I wasn't really worried about anything, and just went with the flow. The upside is that there is no unnatural perfect symmetry. So, after a few minor tweaks to the characters proportions I did a quick unwrap in ZBrush with UVMaster so that I could start with the painting (Fig.15).

Fig. 15

Fig. 15

Texturing

I really like to paint my textures. Using library images for projection is great and very effective. You can quickly get an amazing amount of natural detail that way, but still I think we can get there by hand and with ZBrush tools it doesn't take long.

First, I applied a general color to the whole model. Then, using Color Spray with the some of the default alphas, I started creating some random variation in another tone. In this case, the first layer was a subdermal intense red. I usually lower the Mouse Average of the Color Spray and work with small opacity values (alpha07); then I place some blue variations (alpha22) with higher opacity levels (Fig.16).

Fig. 16

Fig. 16

While painting I made sure to follow the anatomical structure of the character, meaning, for instance, that the areas where the bones are closer to the surface shouldn't be as red as the more fleshy ones (Fig.17).

Fig. 17

Fig. 17

I export this subdermal layer so that I could use it in the SSS material development and in post-production. Then I start to build the dermal layer on top of this one, following the same process of using Color Spray with the previously mentioned alphas.

In situations where I need even less density of speckles or anything I change to one of the simple circular alphas and reduce my brush size. It's really amazing how quickly you can get such a natural amount of detail and with so much control.

For the veins I downloaded some of the alphas that Pixologic provides online and used Drag and Drop with RGB and ZAdd.
During this painting stage I increased the ZAdd values to simultaneously add more detail to the sculpt. This works great because there is an immediate correspondence between color and relief without having to do a second pass with cavity masking (Fig.18).

Fig. 18

Fig. 18

To make sure I had even more control in post-production exploring the SSS effect, I also painted a simple black and white layer that I could use as a mask to pull out more of the subdermal layer if needed (Fig.19).

All the textures were exported at 4k and with the help of MultiMap Export. Besides the painted texture I also exported the cavity and displacement maps.

Fig. 19

Fig. 19

Hair & Fur - The Gills

After finishing the sculpt, posing and texturing I exported a low resolution version of the model to use inside 3d Studio Max 2011, as the growth mesh for the gills.

Using Hair and Fur with a very simple instance node, I quickly managed to create the waving gills effect. After getting them with the density and styling I wanted, I then converted this hair into a single mesh to which I applied a Turbosmooth modifier (Fig.20 - 21).

Fig. 20

Fig. 20

Fig. 21

Fig. 21

Rendering

For the render I used V-Ray 2.0. Since this was a very confined image, limited to the character and with no geometry in the environment, I exported a high res mesh with about 1.4 million polys. It wasn't the highest level of detail, but still managed to keep most of it and the remaining would easily be recovered with the displacement map in the bump channel.

For the lighting, I used a very simple light set with only three lights and a HDRI in the environment.

I find that HDRI lighting is very effective in projecting tonal variation and richness in the model. In this case I used an HDRI from openfootage.net, an amazing free site which provides lots of high quality material. I got a small, deep forest HDRI color corrected into blue tones, as well as a high res version of the same HDRI for reflections. Both loaded with VrayHDRI, to which I applied a Color Correct (a million times better than Color Correction) (Fig.22).

Fig. 22

Fig. 22

As for the lights, the main light was a plane V-Ray light with greenish tones, a Fill light of the same sort in blue and for the backlight a standard Omni with exponential decay also in green (Fig.23).

Fig. 23

Fig. 23

After setting my lights with a basic override material, I started creating the main SSS. Instead of using the FastSSS2 I went for the basic V-Ray material and played with the refraction glossiness and the translucency effects. It really gives you a lot of control and since it was for a still image there wasn't much concern about rendering times, which were actually pretty good. After some testing I got to this configuration (Fig.24).

Fig. 24

Fig. 24

For the gills I applied the same logic. The main render was done in OpenEXR 32 bit and with lots of elements and the necessary multi mattes (Fig.25).

Fig. 25

Fig. 25

After this render I did two more passes with no lights: one with the mask texture and another with the subdermal layer inside a V-Ray light material (Fig.26).

Fig. 26

Fig. 26

Post-production

The main compositing was done inside Fusion 6.1 and took advantage of all the passes and elements I had. It was a pretty straightforward comp which got me to a good result (Fig.27).

Fig. 27

Fig. 27

Inside Photoshop, the image was cropped in order to hide the fact that it was just a bust. Here I also added some suspended particles in the water, which really help to create a sense of depth in the scene.  I rendered some cellular maps from Max with different sizes and spreads to apply in front, behind and in the same depth plane as the character, with corresponding amounts of blur. The final touch was done inside After Effects with the help of Red Giants Looks (Fig.28).

Fig.28

Conclusion

This personal work has already become my favorite, not only because of the final result but also because of the way that I got there. What started out as a one hour sketch with no expectations whatsoever became a 20 hour piece of work with a quality that gave me the chance to be in 3DTotal's character gallery, 3DCreative magazine and writing this Making of, which I hope will be of some use to other fellow artists.

Fetching comments...

Post a comment