Niles Doubleday shows us how to design and model epic 3D robot with some helpful hard-surfacing modelling techniques in 3ds Max, Keyshot, and Photoshop
The key to concept development is to draw all your ideas, even the bad ones, as you never know which direction your design might take. Grab the elements that you like from one sketch and then apply them to a new one in a different way. Just throw down a bunch of ideas, don’t worry about polishing them or think too hard about it. Detail can be as loosely represented as some random lines. Once you have some general volumes you like, then take some time to deconstruct it and detail those parts individually.
I usually take one of my key concepts and put it in the scene on a plane aligned to one of the axes. I always use an Orthographic camera, but it’s especially important at this stage. The Orthographic view makes it a lot easier to make shapes accurate to your concept when you start modelling. I have added some common modifiers to my Modifier box on the right. I strongly recommend setting up this UI element to support your most-used functions.
Using the plane as a reference, start blocking in your key shapes. I did the torso first and made it the most detailed. Once that was done, I started adding in other pieces such as the legs and pipes in the back. This robot is going to be nearly perfectly symmetrical for most of the process, so it’s important to only really work on half of the robot, you can mirror and weld it later. That being said, it’s not a bad idea to copy what you have across to check out the proportions on the full piece as you make progress.
Pick a core mesh and hide the rest of the model. Start to detail out the mesh and get in some supporting edges so that it holds its shape when you apply the TurboSmooth modifier. When you approach subdivision hard-surface modelling, you will spend a lot of time looking for pinching and odd shading. To help you spot these errors easily, apply a base material with a high Specular, and Glossiness setting. This material will enable you to get a better idea of how the forms will look when rendered later on.
The most challenging part of hard-surface modelling is getting interesting-shaped surfaces to subdivide correctly to get a nice clean form. The next couple of steps will cover some geometry solutions I use for some common scenarios. In this example, you can see the shading along the crease of the surface in the red box has a shadow along the edge, the surface in the green below does not. The difference is due to the extra supporting edge added around the actual cut. This provides more support and flattens the surface quicker when smoothed.
It’s always important to model cleanly, however there are exceptions to every rule. In this case, since the side of this leg segment is perfectly flat, it’s ok to leave edges ending in the middle of polygons. You can see that it still smoothes out fine in the image on the right. When you have edges end in convenient places, instead of carrying them all the way into an Edge Loop it saves a lot of time and keeps the model a bit easier to manage.
One of the most important tools at your disposal is the ability to easily move a vertex along an edge. Turn on Edge Constraint under constraints on the right, found in the Edit Polygon modifier. In the image I have a vertex that is interrupting a smooth flow on this bevel. To fix this issue I will turn on Edge Constraint and slide it to one of the end vertices, then slide it back along that edge to the correct location, It will then be perfectly in line with the intended surface and not create any odd bumps.
The Inset tool is very useful for adding Edge Loops to support edges and add detail. Sometimes when you add an inset it creates these overlapping vertices in sharp corners. This doesn’t work all the time but most of the time just collapsing that will enable you to keep your sharp corner and smooth surface without noticeable pinching. Easily navigate to Collapse by Ctrl/right-clicking and finding it on the pop-up menu.
When supporting the edges and better defining the design, I always start to add large-level detail, such as big vents or tube connections. As you add the detail, make sure the new shapes fit in with your existing large shapes. As an example, here is what the underside of the body looks like with the four leg attach points and the vent underneath. Details such as this add an extra layer of believability and display function. The more the viewer can relate the forms to real-world things, the more believable your design will be.
Now that you have all the basic major forms nailed down, there are probably some gaps still left between pieces. Now is a good time to fill those gaps. Use the same workflow that you used on the other pieces to make a sub-surface. Block in some base shapes that intersect with the large forms you’ve got. Try to make these new shapes match the design of the existing ones as much as possible. In my case I have matched the arch of the rear tubes with a rear guard.
Show your model to the client or friends whose opinion you value and ask for feedback. In my case it became apparent I needed to better illustrate a face and show more character. I started blocking out some shapes to see what would fit in with my existing structure and provide a more interesting focal point. It’s ok to make big changes to your model even at this late stage in the process. I pushed the main lens out and added some representational plating, as well as a couple of spline based hoses to add some mid-level detail and externally attach it to the body.
After I defined the face it became apparent the shape would be very time-consuming to model in its curved form and still avoid errors when subdividing. This is a perfect case for the FFD modifier. To use it effectively, model out your form flat to a world axis, where you can get the mesh to subdivide easily. Next, reference that mesh and drag it to the location it’s needed. Once your flat un-smoothed mesh is in place, use one of the FFD modifiers (I used 3 x 3 x 3) to bend it into the shape you would like. Last, go back to your on-axis mesh and TurboSmooth.
The last phase of modelling is where you really get to add some life to the model. Take elements that you think need another layer of detail, subdivide them, then apply another Edit Poly modifier. In the new modifier, start adding some insets, bolts, or seams in key places. Another method is to select a bunch of faces and Shift-drag them off the mesh to create a new sub element. Add a bit of thickness to this new element to create a layering effect.
At this point your model is all finished and you need to put it into a cool pose for rendering. To make that posing easy and avoid making a full bone rig, set the pivot points for the meshes where the joints are. Parent the meshes in the way you would expect the robot to bend. For example, in the screenshot I show the foot plate being parented to the lower leg.
Switch the Rotation to Local and start rotating the base into position, all the legs and head should follow if you parented them. Next rotate the legs and head to suit the camera and the scene. I set up a bunch of boxes to give the robot some context, but sometimes just a flat plane works. I generally play around with a few different camera angles and poses.
I chose KeyShot to render this, but any renderer that gives you easy access to a broad range of materials will work well. It’s a good idea to try out different materials and colours till you find the best result. I knew that I wanted some chipped away paint that would reveal a metal surface underneath so I made an additional render that was entirely metal. In Photoshop I took the metal image and added more contrast using Levels adjustments.
I start by copying the base render onto a layer above the metal layer and I paint a mask where I want scratches so the metal shows through. Some other important elements are Bloom and Sharpening. For bloom, copy the merged layers onto a new layer and use Levels to drag the darks almost all the way to lights to clamp it. Use the Gaussian Blur filter (6-10px) then set the blending mode to Screen. For sharpening, copy the merged layers and apply a High Pass filter with a low pixel count and set it to Overlay.