Making Of 'Doctrich - Post Apocalyptic Scientist Mechanic'

David Anastácio Ferreira gives 3dtotal a rundown of the key processes involved in creating his image, Doctrich – Post Apocalyptic Scientist Mechanic

In this tutorial I will go through my process for creating the character Doctrich from scratch. I will first start with the concept sketching, then go into ZBrush to sculpt it and texture part of it. I will then bring it into 3ds Max to create the final materials and textures for the Hi Res model and set up a basic studio lighting set for tests and for the final image.

References and sketching in 2D

The first thing to do is to establish your theme, your concept. What it is that you want or have to do?
This applies to a professional scenario as well as a personal work – the main difference is that on the first situation, the theme/script is given to you by someone else.

For me the story is very important so if there isn't any, I will make one up. And it doesn't have to be a complete narrative; it can just be a short profile/bio on the character. As Bert Dodson writes in one of his books, "Creativity likes constraints and specifics." It's also on this stage that I'll do some research and gather all the references I feel relevant.

I had stopped drawing for many years and with this project I finally got back to it. It was something that I knew I would really benefit from but was always postponing it. It really doesn't matter if you do it digitally or traditionally, as long as you really take advantage of the speed in which you can quickly experiment with different possibilities.

It is very important that you go through the iterations process, not just personally, since it will really make your final choice a better one, but also professionally since it's something employers in the industry appreciate. Keep in mind that you're not doing pretty drawings for presentation, you're just making those sketches for yourself and so long as you understand them, they're doing their job.

Here I gathered a few sketches I did in Photoshop. It's important to make them as different from one another as you can

Here I gathered a few sketches I did in Photoshop. It's important to make them as different from one another as you can

Refining the sketch

After I choose my favorite concept from the first sketches, I try to take it a bit further by trying out different angles, poses and expressions. It's pretty common for me to take weird pictures of myself in strange poses, to use as reference.

Also, try playing with the proportions a bit, particularly if it's not a realistic character. In Photoshop you can quickly change things up with the Free Transform Tool and Liquify. These really help you precisely define the silhouette and all the features and accessories you want for your character. Again, these don't have to be pretty drawings, just good enough for you to be able to read the character.

Here you can see the further development of the sketch, trying out different proportions with the same gesture, based on weird pictures of myself

Here you can see the further development of the sketch, trying out different proportions with the same gesture, based on weird pictures of myself

Start sculpting

Another advantage in starting with some 2D sketches/studies of your character is that when you get to the moment of sculpting it, it's really, really fast.

I like to start with ZSpheres. To me they're the best way for you to quickly block your proportions and gesture. I find that the topology you get from the ZSpheres is more than enough for the initial and crucial stage of your sculpt. Since the topology follows the flow of your ZSpheres structure, it's really easy to mask an arm, a leg or even a single finger. You just have to hold Ctrl while dragging your Transpose tool along the topology you want to mask.

Here you can see my initial ZSpheres structure, the Adaptive skin and 20 minutes of sculpting on that base mesh

Here you can see my initial ZSpheres structure, the Adaptive skin and 20 minutes of sculpting on that base mesh

Further sculpting

Eventually you get to a point where the initial skin just doesn't allow you to go any further, especially when you start working on the face. But thanks to Pixologic, now you can just duplicate your SubTool, use ZRemesher to get a nice base mesh and just project all the detail you had on the previous version.

As for the clothing and accessories there are several ways to go. The one I rarely use is actually the Extract option from a mask on your mesh. I find it a lot more troublesome then the other options.

My first option is to use the base mesh's geometry. I just duplicate the SubTool, go to the lowest subdivision level and just Del Higher. That way I can isolate the part of the geometry that I want and just Delete Hidden, leaving me with a piece of nice geometry to sculpt, for instance, the gloves.

In some cases the geometry of the base mesh is just too far from what you need, and in those situations I just append a ZSphere and do some really quick manual retopology.

For the shoe laces I just use one of the amazing free brushes available at Badking.

Here you can see a turnaround of the character after 4 hours of sculpting. By now I already have the main elements of the silhouette without the backpack

Here you can see a turnaround of the character after 4 hours of sculpting. By now I already have the main elements of the silhouette without the backpack

Paint-overs

At this point I return to 2D sketching, painting over my BPR renders of the model views. I still hadn't worked on the back-pack robot design and I also wanted to try different hair styles with the helmet on.
When I'm designing the backpack I mostly worry about the overall silhouette, but also about its functionality.

I usually start with a simple line drawing and then add a base value to distinguish the different materials. Then I create a layer for light, a layer for shadows and a layer for specular. Then all you have to do is try to match the light you have on your BPR renders.

Here you can see me trying 3 different helmets and on the profiles. I also try 2 different hairstyles. I decide to go with the longer hair and the first mask

Here you can see me trying 3 different helmets and on the profiles. I also try 2 different hairstyles. I decide to go with the longer hair and the first mask

Back to sculpting

Since I know what I want for the backpack robot, I go back to ZBrush. For this type of hard-surface sculpting I usually start out with a sculpt sketch of the main forms, working with DynaMesh to fuse different basic shapes together – like the cylinder at the base of the backpack where the arms retract to.

It's pretty common for me to go back and forth between working in DynaMesh mode and ZRemeshing it to go back to a more controllable, multi subdivision levels base mesh. It's so fast to switch between the 2, and you can keep all the details you want and clean up what you don't want.

In some situations I still find it easier to jump into 3ds Max and poly model.

In this image you can see the complete model turn around and some illustrations on the different modes these characters can incorporate

In this image you can see the complete model turn around and some illustrations on the different modes these characters can incorporate

Detailing and finishing the sculpt

There's a great expression that goes something like this: ‘You need 10% of the time to do 90% of the work, and then you need 90% more time to do the remaining 10% of the work.' I think this is very accurate. The main sculpting of the overall silhouette is really fast. Detailing and doing the final adjustments is where you'll spend most of your time, especially in characters with lots of accessories or clothing.

At this stage, the background story is of great help. It is what guides the choices you make for what accessories to add, how much wear, how many wrinkles, what he needs to survive: how can I tell my character's story visually?

Here you can see I add a wrinkles and pores pass to the character, as well as a series of details like bolts and straps

Here you can see I add a wrinkles and pores pass to the character, as well as a series of details like bolts and straps

Painting the skin texture

I really like to paint inside ZBrush with Spotlight – it's fast and pretty fluid. The main downside is that the resolution of the texture will depend on the resolution of your mesh, though you can always unwrap your piece as one and then break it down for the texturing. In this case I just had part of the arms and the head, so my highest subdivision level gave me enough resolution for an 8K texture.

I used pictures from 3 different models, from the amazing 3D.SK library. On those situations I just do a quick color correction to the reference images in Photoshop so that they blend nicely.

Here you can see I add a wrinkles and pores pass to the character, as well as a series of details like bolts and straps

Here you can see I add a wrinkles and pores pass to the character, as well as a series of details like bolts and straps

FiberMesh pass

Using FiberMesh isn't that complicated, all you have to do is to lower the Front Collision Tolerance and Front Collision Variations on your grooming brushes. The default values are 50 and 75, and I usually lower them to around 10. Then it's just practice and using masks.

I also like to use Move and Move Elastic brushes to stylize the hair. But in order to use these brushes properly you have to increase the Preserve Length parameter to 100, otherwise you'll end up deforming the FiberMesh.

Another simple trick is to have different layers/passes of hair. That way you have more control and your final result will look more complex and more believable. Since I knew I was going to render the final image inside 3ds Max, I increased the FiberMesh Profile to 8.

On this image you'll find some visual guides to the techniques I mentioned earlier. On this particular character I didn't paint the FiberMesh

On this image you'll find some visual guides to the techniques I mentioned earlier. On this particular character I didn't paint the FiberMesh

Resolution in 3ds Max

Going into 3ds Max, I try to export the highest resolution possible of my most important and complex meshes. It's important to take into account all the meshes when you're thinking about how high your polycount can go in 3ds Max. I find that nowadays 3ds Max can handle a lot of geometry before you really start having trouble in the viewport. In this particular scene I reached almost 16 million triangles.

In order to quickly import all my meshes I use a modified version of a script I found at Scriptspot.

I have provided my modified version with this tutorial. I usually export Displacement, Texture from PolyPaint and Cavity maps from ZBrush. With these 3 I can really get some nice textures in Photoshop. In many cases my texture from PolyPaint will just have values. That works as a value guide for my whole character and it will also allow me to quickly select different parts of the mesh when working on the textures.

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Setting up the stage

After importing all the meshes, I like to link everything to a point helper in order to quickly adjust the scale and position of my objects. Very often, when you start working with ZSpheres in ZBrush you don't really know your objects' scale. No matter what units you choose to work in, it's very important to work in real scale in order to get predictable and more realistic behaviors from your lights, materials and many modifiers.

So to scale and position all my objects I just need to work on the point. When I have my feet on the ground and my scale corrected I delete the point, do a Reset XForm to all my objects and then convert them to Editable Poly.

For my light setup I use the typical: Main, Rim and Fill. I'm a V-Ray fan so all my lights are V-Ray. In many cases, such as this one, I'll disable the fill light and instead, I'll give the scene a bit more of an environment light with a nice HDRI.

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Working on the materials

When making my different materials I really like to mix things up. With ColorCorrection and VrayCompText I can get almost all the different maps (Reflect, Reflection, Glossiness) from one single diffuse texture without having to go back to Photoshop. Of course it's always better to develop your specific maps in Photoshop.

I like using hi-res textures so my range will go from 2K to 8K. When loading bitmaps with high resolutions inside 3ds Max, you should always lower the default Blur value on the Bitmap node from 1 to something like 0.2 or 0.1. A Blur of 1 in a 4K image will be extremely obvious and you would be losing a lot of the detail on the texture.

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Here you can see the maps I exported from the gloves in ZBrush and the final textures I'm using for the hi-res render

Rendering the final image

V-Ray is extremely fast and I like having fast renders, so I tweak the V-Ray settings even further to get the quality I need with the lowest render times possible. When doing still images, for GI I tend to use Irradiance Map and Light cache, both at 0.5. That way I get deeper shadows from the render to start with. In this case I don't calculate secondary bounces and go just with Irradiance Map at 1.0. I'm not a big fan of Global adjustments like the Global subdivs multiplier, but I do use it on rare occasions.

My final renders are in OpenEXR with a good collection of elements that will give me total control of the final image.

My post-production usually has 2 stages: a first one in NUKE (or Fusion) and a second one in Photoshop (or After Effects). For this project I do the main compositing inside NUKE, making use of all the elements and MultiMatte elements, for the big adjustments. In Photoshop I do the final tweaks, some hand-painted corrections and the final look.

These are the V-Ray settings and elements I use to render the 2 final images with a resolution of 2000 x 1500 each

These are the V-Ray settings and elements I use to render the 2 final images with a resolution of 2000 x 1500 each

The final image: front

The final image: front

The final image: back

The final image: back

Top tip 1: Practice your drawing

Whether you're a character artist, an environment artist, an animator… every artist can benefit from drawing. Preferably traditional drawing because you don't have undos, layers or free transforms, which makes it harder and thus more effective.

Top tip 2: Moderate your GI

One of the most common problems I find in the images of beginner artists is too much light. And one of the easiest solutions when working with V-Ray is to lower the GI contribution of your primary and secondary bounces.

Top tip 3: Know your render engine

One of the most important things to have when working with lighting, textures and materials is speed.
If you can't optimize, in order to have fast renders showing you just what you need to see, then it'll take you forever to have everything fine tuned, or worse, you'll get frustrated and won't go as far as you could.

Related links
Check out David Anastácio Ferreira's website
Scriptspot has many helpful scripts for 3D modelers
3D.sk is a free resource site for textures

Fetching comments...

Post a comment