Currently, one of the best ways of achieving photo-realistic imagery is to render using Mental Ray for Maya. Mental Ray offers a Global Illumination and Final Gather solution, which when combined, simulates the physics of real world lighting effects. Now, for the first time in 3d, lighting techniques used by photographers and filmmakers can be applied to computer graphics. The following is a guide for setting up Global Illumination and Final Gather using Mental Ray for Maya. It is based on notes from the web, Maya’s Help manual and good-old fashion experimentation.

Setting-up

The first thing I always like to do is develop some sort of plan. As they say, ‘If you don’t know where you’re going, you probably won’t get there.’ Here is a quick set-up sketch I did in Photoshop. I have included lighting placement and some material information.

The basic idea from the sketch was modelled and setup. Maya has four basic shaders to choose from: Lambert, Blinn, Phong or Anisotropic. Lambert has no specular highlights, Blinn has soft specular highlights, Phong has hard specular highlights and Anisotropic has irregular highlights. Based on the sketch, the first material will be a white, matte background that has no shine. A basic Lambert material was used with color set to an off white (very light grey) and diffuse increased above 0.8. Next is the sphere, which will be used as a prop to study the effects of illumination. A Phong shader was used to give it a highly polished, chrome like appearance, so as to be able to better study how the light rays are being traced. The last material is the Negative fill. These are being used like large barn doors, stopping the light from spilling. To suck the light from the scene, a Lambert material was again used, with both the color and diffuse set to black.

This is the final scene set-up. Two spot-lights have been added either side of the negative fill and are pointed at the wall. Their light will bounce of the white wall, to create a soft large light, illuminating the back of the ball. Any spill from the lights is being stopped by the black boards. Later the lights will be converted to Mental Ray area lights to create softer shadows.

As the sphere subject is not receiving any direct illumination a reflector board has been added. It will catch reflected light from the screen and illuminate the sphere from the front. The same material used for the background has been applied to the reflector board.

A standard hardware render produces an almost black image as most of the illumination is to be achieved from bouncing light.

Global Illumination

The first part involves setting up the Global Illumination. Once a good photon map has been created, then Final Gathering will be added to increase indirect illumination. Global Illumination using Mental Ray, requires the adjustment of two main sections: the light’s attributes and the Render Globals Settings.

Both light’s attributes are first set-up. To get realistic lighting behaviour,

• the decay rate is set to Quadratic and
• the light intensity is initially set to 120.
• Raytraced shadows are turn on in the Shadow attributes section.
• Lastly, in the Mental Ray section, the light is set to emit photons.

To create the Global Illumination solution, the Mental Ray render is selected in the Render Globals Settings, and in the ‘mental ray’ tab, the Quality Preset is set to Draft and Global Illumination is turn on in the Caustics and Global Illumination section.

Here the initial render is created. The Negative Fill boards were adjusted to give a more even lighting. Both the light’s intensity was reduced by half down to 60 – giving a total combined intensity of 120. The spot lighting effect was also reduced by increasing the Penumbra Angle to 10.

The next stage is to fine tune the Global Illumination. To adjust the overall brightness of the scene, the light’s Exponent value and Photon Intensity attributes are adjusted. Here is Maya Help manual definition for the Exponent value:

Exponent

This is similar to decay — the intensity increases as the value decreases. To increase the chances that indirect light will reach a greater distance, decrease the value.

Visible noise can occur with values less than 1. The default (2) simulates natural (quadratic) decay, but violates the conservation-of-energy law (that happens in the natural world), so bright spots from distant light sources could occur in unexpected locations.

Phrased another way, higher Exponent values will decrease brightness. Increments of 0.1 through to 3.5 will gradually decrease the brightness.

This is the same render with an Exponent value of 3. The image is much duller.

First of all I want you to know that this image was created for a competition, the goal of which was Lighting & Rendering Atmosphere.

As this is a making of article, I’m not gonna cover every step of the whole process. I assume you have some basic knowledge of 3D in general, and basic to intermediate level for all the 3ds max users.

First I created my concept sketch to focusing better on my idea for the challenge. I wanted to play most with the lightening and atmosphere so i concentrated more on that. The baby in the scene was used to give more expression to the whole situation.

For me it’s better to start modelling from simple shapes, using the box modelling technique. So I started modelling the head from a simple box. It’s better to study your model first and find how many divisions the box needs to get the look of the face better and faster. That’s why I recommend to always start with the simplest shape, and adding details gradually.

stage 1

stage 2

stage 3

stage 4
A very important fact to consider is the modeling flow. I’m talking about the way you build the surface; edge loops/edge rings.

Ok, after dividing the box several times, I converted it to a editable poly, and started adding details. Although I was dealing with a complex organic modeling, the tools I’ve used were very few. 3ds max has a very robust polygon toolset. 90-95% of the work was done using 2 or 3 tools.( cut, extrude, bevel, weld). Cutting here and there, I’ve modeled the face of my character. I built the mask using surface patch technique. Drawing splines first, and than adding the surface modifier.

stage 5

stage 6

stage 7

stage 8

stage 9

stage 10

Using the method above, I started modelling the body. Starting from a simple shape/ turning to poly/ and adding detail with the tools I’ve mentioned before.

Now that I’ve finished the modeling, it’s time to skin it. Doing this I can make the pose, like my concept drawing.
I built a fairly simple skeleton rig with max’s standard bones and added some simple controls for the limbs.
I added the skin modifier on top of my poly surface first, and then all the bones of the skeleton rig.

As soon as the skin phase is finished, it’s time to do some texturing work!… I added Unwrap UVW modifier and started pulling and pushing the UV’s, trying to make them as flat as possible. It’s good way to be in sub object mode, so you can see directly in the view port, all the polygons you select in the UV editor.

After I opened photo shop, I imported a screen shot of the UV’s, and started painting the diffuse map. I wanted the skin to be smooth and translucent, so I don’t need to create a bump map for the skin.

This image was done for a competition at a Swedish community for 3ds max, www.animate.se. I will be focusing on the materialsetup.

Modelling:

I started out sketching on a couple of ideas and I quickly made a mockup of the final model. It is going to be a robot.

Then I started to model the different parts, beginning with the face. I started out with a polygon and extruded edges to form a face. I didn’t have much time on this model so I decided to create the face in its final state right away.

And here’s the finished model with all of the different parts put together. The same process was used for the other parts as the face.

Rigging:

Character Studio was used to make a skeleton for the model. Since I had modeled joints and other parts separately, and I didn’t want any deformations, I just linked each part to the corresponding bone instead of using a Skin modifier.

Materials:

I really didn’t want to spend a lot of time unwrapping the different parts to texture them so I decided to do it procedurally. But I still wanted control of where the different materials would appear. And to accomplish that, I used VertexPaint.

The idea is to paint a black & white map in VertexPaint and use that as a mask in a Blend material (above image).But to be able to paint fine detail onto your model, it has to have a lot of vertices since the information is saved in them.

To get past that I added a TurboSmooth modifier and set both iterations to 3 as you can see in the image. It is important to have the VertexPaint modifier after the TurboSmooth. That allowed me to paint a lot of detail and I was also able to lower the main iterations while keeping the Render Iterations at 3.

Remember that you must have the same amount of iterations when you render as when you painted your vertices. So if you painted with 2 iterations you must have 2 iterations at rendertime. It is pretty obvious really since you’re changing the amount of vertices with the iterations.

The base textures I am using here are the ones that come with 3ds max 7. I’ve used Noise and Mix maps together to try and hide the repeating textures you get from the base bitmaps. Although you can easily spot the tiling of the textures it doesn’t show up on the model unless you look really close.

Lighting:

I used a basic three point lighting with the keylight coming from the right. All three mainlights are arealights which basically means you’ll get soft shadows. Small omnilights were also placed out to get light on certain key areas that I wanted to emphasize. A Skylight was used as an ambient light together with Final Gather.

Rendering:

I rendered using mental ray 3.3 which is available in 3ds max 7. It gave me the control that I needed make this image. Final Gather was used.

After having collected some pictures of the famous actress Angelina Jolie, I decided to use some of them for giving life to my homonymous creation Angelina Jolie. Before starting the phase of modeling it was important to take a moment to observe properly the features of this beautiful actress, above all of her cheekbones which are very particular and pronounced.

Starting from the two references, front and lateral view, I modeled the lowpoly face, imported into ZBrush and increased the subdivision level by 2 steps.

Then, I could concentrate on adding details to critical areas of morphology like mouth, nose, bowed eyebrows, eyes and ears. During the working process, I pay attention to maintain a clean and morphologically correct mesh and to avoid disharmony, in order to permit an efficient modeling in ZBrush without too many difficulties.

I added loops only on the more important areas, to have a more manageable mesh for the modeling.

As you can see, I added only one additional iteration, obtaining a density which satisfied me. I obtained the bump map from the diffusion texture.

Having completed the model, I imported it in 3ds max, where I rigged the head and neck to pose the subject with a simple rotation of the head. In the end, I’ve put the camera and chosen a typically portrait view, I added a leather string in order to render the idea that the woman is wearing a dress.

For the shading, I was inspired by various tutorials I had found on the web. I also studied how the various slots work in order to obtain a realistic skin. Here are my sss fast skin settings:

For the illumination I considered various photographical techniques, and put the primary light between camera and subject (45°), slightly moved to her righter side. Then, I added two backlights to emphasize the silhouette of the face and to obtain more depth for the final image. Finally, I added a fill light to compensate the dark areas.

For the final render I decided to use Mental Ray as my render engine, already integrated in 3ds max and using only final gathering and not the global illumination. In the settings of the FG, I used 0 bounces to obtain a low render time, compared to the SSS fast skin complexity.

So, I rendered a beauty pass of the scene. I created a composition in Photoshop that permitted making last substantial changes to the final image. As background I added an image of a curtain that I had turned black before. I made the hair with the Hair and Fur modifier, and rendered it separately. So, I composed the hair and the beauty pass with Photoshop to obtain the final image.

Hope you like this image and thanks a lot for reading this overview of the image. If you have any queries then don’t hesitate to ask.