Project Breakdown: CFA Gatehouse
Getting everything into MAX
On this project, we were lucky to get quite a comprehensive Sketchup model from the client. Now, while additional detailing was needed, this took out a lot of the errors that can occur when translating a set of working drawings. Getting a ready built 3D asset from a client, even if it is only in massing form, speeds up the process immensely!
While things like furniture, foliage and people were already in the Sketchup model, they must ALWAYS be replaced for a high-end 3D image. Sketchup models usually also use texture to show detail, often using opacity mapping. Here at BURN, if we can, we model it. Proper use of textures is definitely half the battle, but actually modelling the needed geometry almost always yields better results. While you can’t always model individual grout lines or actual links in a fence, there are certainly some things that you CAN model that will give your image a much better sense of believability – and when used in conjunction with dirt maps and occlusion passes, will give your image much more ‘punch’.
Another advantage of receiving a ready built 3D asset from the client, is that they usually have a good idea of camera angles and of what they want to see. This helps in streamlining the process of ‘only model what you see’. Like the back of this building wasn’t going to be seen on camera, so we never touched it in the modeling process. We were also aware of how close or far the camera came to the building for the shots so that the amount of detail that need to be added could be easily judged. For example, even though there are large windows on the rear of the building, the camera is far enough from them that low resolution models can be used.
When exporting files from Sketchup, we export as DWG, as all of your layer information is saved.
Fire up MAX and go to FILE >> IMPORT
Click OPEN,
Adjust any settings you need to adjust. I find that the default settings work 99% of the time.
The file has been imported, complete with all its layers intact.
Landscaping
The contours already inside the Sketchup file can’t be used in the final render, but can work as a guideline when remodelling it in MAX. There are literally hundreds of ways to model effective landscapes – from joining contour lines and using patches to using ZBrush. Because these contours were imported as solid geometry we used them as a base, and using FFD modifiers and soft selection, edited the box to cover the landscaping.

Create a box covering the entire landscape, and set its Length and Width segments to something that will provide you with some density when modeling. Dont set it to high, as you can always smooth the result later, and smoothing a super dense mesh can cause issues.

Use soft-selection to sculpt the smaller sections of the mesh, raising, pushing, or pulling the faces to provide the contours.

At BURN we always do camera specific detailing, and thats no difference when it comes to landscape sculpting. Using Soft Selection, tweak the mesh to best show off the landscaping per camera view.

A quick render, showing the imported building and new landscaping. Notice the additional detailing on the metal screens.
Once the landscaping was done and small additional detailing to the building added, we can begin texturing. Initial texturing tends to be ‘broad strokes’, with no bump/displacement maps (or even spec/dirt maps). Doing the initial texturing as just DIFFUSE textures allows for fast tweaking from the client’s side (enlarge a certain texture, make something darker, adding in shine, etc). Once the basic texturing is approved then go into further detail; adding in dirt maps, bump, specular maps and anything else that’s needed for the image to ‘pop’.
An advantage we have in the archviz field? We can usually get away with clever shader work in the place of complex unwrapping. A square texture and a box UVW Map is enough, in most cases. Although there have been times when we’ve unwrapped and painted models, but it’s usually overkill.
The metal grating was textured using THIS TECHNIQUE.
Grass!
In keeping with the philosophy of keeping as much physical geometry as possible, we actually model our grass. Yup, model it! This allows for an excellent degree of control over the final placement of the grass and also direct feedback on the ‘look’ of the image.
To create the long African type ‘veld’ grass for these particular images, I took inspiration from a fish’s’ scales. Using a ’scale’ system of circular grass plates, I quickly populated the undulating landscaping with long grass. Sounds confusing? Let me explain!

Selecting the CONE, use the SCATTER OBJECT and SCATTER the cone over the cylinder. 100 instances should be enough but you can always adjust this later. Set the rotation of the cones to 2 (this would be higher if your grass was a mess... if you wanted it to look 'trampled on')

Make duplicates of the 'scale'. Use INSTANCE when duplicating the scales, so that if you want to change the grass length or density, you just have to change one patch. The pattern I have above works in most situations. Rotate each scale a little to make the scales appear more random. It's important to set the PIVOT POINT of the scale to the center of the cylinder. You'll be dropping these on the Z Axis to the landscape surface using the MOVE TO SURFACE script. Set the DISPLAY PROPERTIES of the CYLINDER to INVISIBLE TO CAMERA. This way when you render, only the grass will be visible.
Save this patch as a separate MAX file that you can access for any other projects. Using this technique you can create any sort of grass – from short manicured lawns to long high grass. It also renders very quickly, often faster than hair based solutions or displacement mapping.
Placement of the grass was done using the methods discussed HERE.
Additional rocks, trees and other low level planting were VRay proxies, placed individually by hand to best maximize the impact of chosen views.
Render Setup
Initially, I wanted to go for a dusk shot with the terrain being quite dry and arid. Eventually this changed and the preferred feel of the image was a mid-morning light, with the vegetation being much greener and the grass slightly shorter. This was easily altered because of using instancing.
The Vray setup for the render was very simple. I find that the default settings are default for a reason…they work!
The scene was rendered using Irradiance Maps and QMC with colour mapping set to Exponential.
For Antialiasing, I used Adaptive QMC and AREA as my filter.
Lighting is a single VRAY Sun and a VRAY Plane light for the interior lighting. I prefer to limit the amount of lights I have in each scene to keep render times down as low as possible.
Here are the renders with the grass and the dusk lighting:
From this point on, we worked hand-in-hand with the client to tweak the image lighting and landscaping. There were some smaller modelling updates that happened as the project went forward, but that is to be expected. Setting up the file using LAYERS meant that changes didn’t take up too much time and the model kept constantly up to date.
Here are thefinal images with Photoshop foliage seamlessly blending in with the modelled geometry. Having a good base render is essential to getting a great final result.
Some techniques used on the final post work for the images are discussed in these articles:
Post Production: Chromatic Aberration
Post Production: Defringe















![CFA Gatehouse 1 [1024x768]](http://www.burnstudio.co.za/blog/wp-content/uploads/2010/05/CFA-Gatehouse-1-1024x768.jpg)
![cfa gatehouse 2 [1024x768]](http://www.burnstudio.co.za/blog/wp-content/uploads/2010/05/cfa-gatehouse-2-1024x768.jpg)
![cfa gatehouse 3 final rev E temp2 [1024x768]](http://www.burnstudio.co.za/blog/wp-content/uploads/2010/05/cfa-gatehouse-3-final-rev-E-temp2-1024x768.jpg)
![cfa gatehouse 4 [1024x768]](http://www.burnstudio.co.za/blog/wp-content/uploads/2010/05/cfa-gatehouse-4-1024x768.jpg)




Jonathan May 7th
Nice tutorials! thx for posting. I would be interested in seeing your methods of post producing the foliage, particularly the grass and water.
davis May 7th
Thanks for your tips.. i enjoy reading you blog.
Davis
Nic B May 7th
Thanks Jonathan, I’ll definately do a post production tutorial in the near future.
Davis – thanks!
Nikhil May 12th
Cool , can u plz breakup how to get detail in gray scale image and when i try vray cam the image gets dark corners on the edges of image and in center the image to bright
Nic B May 12th
Hey Nikhil,
The ‘dark corner and light center’, is caused by camera Vignetting. This is an ‘undesired’ effect in physical photography, that Vray simulates for the sake of realism. Our images always have a slight vignette applied to them, sometimes even done in Post.
For grey renders, if you are looking at getting something quite ‘clean’, the easist way is to set your COLOR MAPPING to EXPONENTIAL, as this will give you a nice, soft feel. You an also turn off Vignetting in the BASIC PARAMETERS of the Vray camera. Something that works in practical photography to reduce a vignette is to inrease the amount of FStops on your camera…this may also work with the Vray camera, although I haven’t tried it!
About Author
Chris B
Co-founder of Burn Visual Illustration and the lighting magician.