The purpose of tking of is to outline the tools and techniques I used in the creation of my image, Hornedman. The image started as a sketch inside ZBrush and grew from there into the final image almost by accident.

Sketching characters in ZBrush is a great way to concept ideas, mess around with forms and come up with cool new things. After the sketch I retopologised the head section of the model and from there used the polypaint tools in ZBrush to create the textures. Rendering was handled with Mental Ray in 3ds Max. I find Mental Ray to be the most accessible of the available rendering systems for Max, and the fact that Max already comes with it makes it great. Also, having used XSI for many years, the creation of shaders in Mental Ray has become second nature to me.

Sketching

The starting point of many of my images, characters and dailies are done in ZBrush in the form of a sketch. As my drawing skills aren’t really worth writing home about, I find ZBrush to be a real lifesaver in terms of creating concepts. And the great thing is that you can visualise your concepts in 3D and even use the concept as a template to model over later on in the pipeline.
For this character I started with a basic human form mesh that I created in XSI some years ago and have been using in many projects since (Fig.01).

Fig. 01

Using this mesh at the lowest sub-division I pushed and pulled it around with the Move tool and Standard brush in ZBrush to find a base form and silhouette I liked. From there I worked my way up the subdivision levels adding progressively more detail using the Clay Tubes and Clay brushes only. An overview of the different sub-division levels can be seen in Fig.02 – 03.

Fig. 02

Fig. 03

Retopologising

For this model I decided to just retopologise the head as the loops on the body were adequate for my needs. The retopologising process was simple; I exported a mid-res .obj from ZBrush into 3ds Max and then used Polyboost’s surface snapping tools to create a clean mesh over the old one. For the body I just exported the lowest subdivision out of ZBrush and joined that onto my new head. I then exported this new low-res .obj back into ZBrush, assigned it as a SubTool of the original sketch, subdivided it up to 4 million polys, and chose Project All to project my sculpted sketch onto my clean topology (Fig.04).

Fig. 04

As this was still indented to be a full character I also modelled some base gear and straps for him in 3ds Max

Texturing

For the texturing stage I decided to use the polypaint feature in ZBrush. As the model still didn’t have a set of UVs this was perfect for me to sketch on a try out some ideas for the skin. I used a technique outlined by Scott Spencer with his Stinger Head model to paint the texture. Firstly I chose a base colour for his skin, and then painted on sprays of red, blue and green in key places. Then I “noodled” the skin with white veins and finally sprayed over it all with my base colour at a low opacity. This is a very fast and effective technique for concepting skin tones and painting textures. As I still wasn’t sure what to do with him at this stage, I left the texture rough (Fig.05 – 06).

Fig. 05

I wanted to push myself on how far I could go, so I decided to build a piece capable to make an art statement; I didn’t want to feel that it was just another 3D model.

Being an automobile nut, I, like some of you, have had the idea of building a car from the first to the last bolt. I find out that for this task a good set of blueprints is never enough. I spent around a week collecting photographs before the beginning of the modeling process, and I was still collecting references at the end of the texturing and rendering steps.

My first advice to you will be, if you are trying to make a model above average, start with planes, boats or cars that had been restored. You will easily find part catalogues, illustrations, schemes and diagrams of objects with historic meaning. I made a quick selection of some of the GT-40 pictures I found during my research (Fig. 01).

I usually try as much as possible to start from spline cages. It comes very handy to have a tridimensional blueprint of your model; it will guide you making decisions about size, position, and where your components should be organized, even before you model the shell.

Later on, you can use the spline curves to loft panels that will be the base mesh of your car body.
Once I had my cage done, I started modeling the chassis using photo references (Fig. 02).

The next step was modeling all the components that have direct relationship with the chassis.

I always started creating primitives to establish rotation and proportions, and then I went in detail using pictures. Here once more I used techniques like nurbs revolves, lofts and extrusions than later on turned into polygons.

Finally, I used lattice and nonlinear deformers to achieve the desired shapes (Fig. 03, Fig. 04, Fig. 05).

INTRO

Hi everyone reading this overview about making-of my work “DNA_priority”. This overview is software-specific as I have used 3DsMAX and Rhinoceros for making a picture and Adobe Photoshop for postproduction but there are no limitations for other software users due to most of principals of production are the same. I will try to introduce to you how I made it. So there is my workflow below:

- decisions
- modeling
- lighting
- texturing
- rendering
- postproduction

Before we start I want you to know that some techniques and solutions that I used can come in sight as something incorrect and maybe strange but anyway I will explain the process as it is. 3D as I see is a some kind of art and there are no limits and ultimate rules after all. So I hope you will like the article.

DECISIONS

Generally before the project have to be started it is important to decide to yourself what your work will look like. I mean the moodand impressions of the final image. I think it is not a good idea to changeyour mind about the style in the middle of the process. But if you sure that your newstyle or mood is better than you are workin on, so do it – change everything. In this case during the processI have spent a lot of time sitting infront of my monitor and just look “inside” viewports because Ihave no idea how to make it more expressive. So if I didn’t like the result I delete all the files ofmodifications.

>>>Tip: pressing SHIFT+DEL is a best way to make your work better.
If you don’t like what you do, who will?

So before I started I decide to show a symbiosis and make the picture non-stylised and realistic as I can. And I thought that filling the scene with abstract and incomprehensibleobjects will solve a good fake effect (as for me I like artworks that fake me out and looks real.As one of my teachers said:”You know that it is fake, but you don’t know where it is”).So all theese was my guidance in the process and I followed it.

>>>Tip: If you know what you want the half of the process is finished.

MODELING

So the next step is modeling. All the geometry in the scene was created withNURBS and POLY with NURMS subdivision techniques. The mesh is quiet simple,so I don’t see any reasons for deep explanation of modeling but however maybesomeone will find it useful if a few samples will be shown…

This is an overall mesh view at the end of modeling process. Well… Actually when I see this and the final image I arrive at a conclusion that the mesh is “overloaded”. In my oppinion the rule “you shouldn’t model it if nobody can see it” is completely correct. But sometimes it is impossible to know what does the final image look like. And in this case exact positions of objects wasn’t determined yet. But anyway it’s always better to realise a proper distribution of mesh density: the closer the object to camera the higher density should be provided and vice versa. Of course this is not a strong principle especially when you are going to use effects (DOF for example) but it will sufficiently increase the perfomance and rendering speed.

OK. To create these objects (except of wires and other small details) I’ve used Rhinoceros.I found this software very useful when I need to create smooth surfaces that do not require of “clean” meshes and normals. As you see all of this objects consist of triangles withdensity distributed regulary over the surface depending on it’s curvature.

>>>Tip: Such a geometry is still editable but Idon’think this is a good idea to do it directly.So it would be logical if we have final meshes inNURBS modeling output.

Rhinoceros has many powerful tools and is very flexible in modeling. Except of allexported meshes contain UV coords (in some cases it really helps in mapping). In additionit brings a lot of fun.

This figure shows some of geometry was made in 3DsMAX. So everything is very simple as you see.There are input meshes on the left side and final meshes after NURMS subdivision applied to iton the right side. Priorly it is Editable poly with subdivision surface turned on but thereare no global difference if MSmooth modifier would be applied to low-poly later.After completing a low-poly model it is very helpfull for mapping to apply polygon ID’s and assign Milti/Sub-Object material with differentdiffuse color attribute. And besides it helps to distinguish objectsin viewports to get a proper composition.

>>>Tip: To meshes with extremely high number of polygons on output it’s better to applydifferent subdivision values for viewport and render due to high valuesfor viewport may sufficiently slow down VGA adapter perfomance (for figure abovethose values are the same with isoline display option turned on).

After all the geometry is completed and have a proper position in the sceneI have started the next step “lighting”.

LIGHTING

Talking about lighting it is necessarily to bear in mind that everything should be in conjunction withrenderer you have chosen earlier. This time I chose Light Tracer as Advanced Lighting (Default Scanline Renderer). So due to this I am able to use Skylight and IES Sun and don’t have to worry that something would be wrong with rendering. You can see lights positions on the figure below with IES Sunparameters.

As for Skylight I left it’s params by default. After some renderer adjustments that’s what I hadon output:

Ofcourse it is only a first steps to finat rendering but it is necessary to decide an overall light ina scene before texturing and then calibrate textures to lighting. As you see it is time to go on with texturing.

Hey guys!
I did this image for the GNOMON Gallery Show: COGNOSCO. Feel free to find out more information about the gallery here:http://www.gnomongallery.com/current_show.php

So I used my old “Angel” model and just re-rendered it. Here is the final image:

I’ve also created a showcase tutorial article about how to create cool renders using ZBrush and Photoshop.
Let’s get started.

Preparing model and document.
Create new document with size twice as bigger as image you want in result. I do that every time I render any model in ZBrush because in the end I shrink the image at half size to get rid of “jagged” lines. I also press AAHalf button to see a preview for smooth-rendered antialiased image.

Set up your model on the canvas as you want to see it rendered. You will need masks for easier selections at further comp stage. So just go through all subtools one by one and export image with Flat render turned on.

As result you get bunch of images with masking for each separate subtools.