In this redesign of PAX TV’s on-air look, the primetime promotional end caps are filled with light and translucent shapes. Simple flat characters and shapes were exploded and hit from a number angles with different-colored lights. Multicolored light rays add dimension and energy. (creative director: John LePrevost of LePrevost Corporation; art director/designer: Wendy Vanguard of Manna-Design; realization: Chris Meyer of CyberMotion)
Because of either lack of time, or simple Fear of the Complex and the Unknowwin, many editors and 2D graphic artists resist learning how to use a 3D program. And that may be unwise. More graphic design is incorporating 3D elements – from the ubiquitous extruded flying logo, to cool lighting effects, to wireframes of simple geometric shapes added as visual spice. Your clients may not even know this is “3D”, but they know it’s a look they want…and if you can’t supply it, they’ll look for an artist who can. Don’t worry – you don’t have to create Toy Story 3 single-handedly – but some basic skills will more than pay back the moderate effort invested.
There are many solutions that give you limited variations of 3D inside your normal 2D environments, such as plug-ins that add 3D text or objects to After Effects or various editing packages. Some 2D compositing applications such as After Effects, Motion, and combustion, also allow you to animate planes of graphics in 3D space and to move a camera around them. These are indeed useful, and provide enough 3D for many jobs, but rarely do they offer the power and flexibility of a full-blown 3D program.
For those new to the field, the first two pages of this article will cover some of the basic concepts that make 3D different than your normal 2D programs. I’ll then go over a trio of real-world case studies from our vaults that feature relatively simple 3D: They don’t feature any dinosaurs or pod races, but neither could they have been pulled off as effectively in 2D alone. The concepts explained here can be carried over to virtually any 3D program.
3D: The Lay of the Land
I don’t want you to be intimidated by 3D, but I do admit that the field is deeper and more complicated than can be explained in a single article. Fortunately, there are literally hundreds of books out there that can help – either in general, or with specific lessons and advice for the program of your choice.
To get you started, the following is a list of concepts of how 3D varies from the 2D or editing programs you’re already using:
Depth and Movement
The primary advantage of 3D over 2D is that it has an additional dimension: depth (sometimes referred to as Z space, for the third axis in the traditional X/Y/Z coordinate system). In addition to left/right and up/down, you now have front/back, with objects automatically scaling themselves as they get closer or farther away.
Working in three dimensions initially takes some mental calisthenics, but a lot of advantages naturally fall out from it. For one, objects at different distances will automatically multiplane as you move past them. They also have depth: you can turn them on their side and they don’t disappear, unlike many psuedo-3D implementations in 2D programs.
The other mental exercise required by 3D is that not only can the objects move, but so can the virtual camera that “sees” the object. If you’re only used to moving objects around a video frame, you can get yourself tangled up real fast if you try to animate both the 3D objects and the camera at once, with the objects seemingly wandering out of view. Try to restrict your first animations to either keeping the camera stationary and moving just the objects, or arranging the objects into a stationary arrangement you like and then moving the camera around them.
Moving in three dimensional space also means you often use more than one “view” – such as from the front, the side, and the top – to really see what is going on with your motion paths, especially if you’re moving in Z space. Most 3D programs tend to be flexible in letting you set up these alternate views. An example of this may be seen in the figure above.
Timeline & Keyframes
All 3D applications we’ve used have a timeline and keyframes, similar to what you’re using in 2D compositors or to keyframe effects. There’s just a lot more parameters that can be keyframed: Z position for starters, but also all kinds of information about the surface of the objects, plus how they may scale or deform. Many of these parameters may be hidden initially from the user. The figure at right is an example of this: Only the location of keyframes are shown; you have to twirl down the arrows to the left to reveal the individual properties.
Those used to After Effects’ easy access to velocity curves will dislike how 3D programs often hide them in a separate “function curve” window; those moving over from programs like Discreet’s combustion will be more used to this concept. Some 3D programs also lack the option of entering precisely the amount of ease in/out influence you want; you might have to add more keyframes to get the speed curve you need.
Another area of frustration can be in attempting to trim and fade objects. In some applications, a 3D object will hang around the entire duration of the project, unless you explicitly edit its visibility to turn on and off at certain frames. Fading objects can be trickier too: sometimes just changing its surface color to black or turning down the lights works, but it will still be there in the alpha channel, which can cause confusion later. There’s also a difference between “transparent” (as in glass) and “invisible” (as in gone), again with the result often being visible in the alpha.
If you need to do traditional video-style editing on 3D objects, it is often better to render them as individual elements, and then edit and fade them in another application. Indeed, it is good to employ scene-based storytelling techniques with 3D; too often, beginners go for overly-long, continuous camera swoops simply because they’re so easy to do in 3D.
Texture & Light
These are the areas that I personally enjoy most about 3D. In most 2D applications, every object or layer is usually “lit” evenly, and any text or simple shapes you create default to perfectly smooth, even color. But that’s not what happens in the real world.
If you don’t have experience using real lights, virtual lighting can be one of the biggest challenges to learning a 3D program – but if you do, you’ll find 3D far more flexible than most lighting tricks available in 2D programs. By using lights, you have a much wider range of moods you can set by focusing on individual areas, or just “brushing” an object with light. The figure at right is an example of a simple, flat logo that has been made far more mysterious by focusing a pair of spotlights on relatively small regions of the semi-transparent letters.
Virtual lighting in 3D encourages artistic experimentation in setting moods, even on these three simple solid-colored letters. The “P” also has a slight bend deformation along its face, to allow the light to play off it in a more interesting fashion.
Most 3D objects default to the look of smooth, white plastic – not a big improvement over the solid colors in 2D programs. However, all 3D programs allow great latitude in how you “texture” the surface of your objects. You can either apply a texture map, which can be a still image or movie of the type of surface you want your object to have, or use a “shader” which automatically calculates building-block style textures, such as clouds, bricks, or surface grain. Both texture map libraries and shaders are available from third parties, akin to stock footage and plug-ins for 2D programs.
Avoid the temptation to use textures that resemble shiny gold or blobby day-glow psychedelics; they’re sure marks of a beginner. Quite often, I choose only a simple color for my objects, and then add a “bump” map which modulates and diffuses the light that hits the surface to make it look like it is rough or has a pattern. Another underused, understated trick is to manipulate what’s usually called the “diffuse value” of a surface, which decides how much light is absorbed or bounced back at various points across the surface. Applying a grime map or even a simple cloud shader, you can age a surface nicely for added realism. (For more on this technique, we particularly like the Surface Toolkit from dvGarage – it’s the first texture library we load up for every job.)
Perhaps the scariest thing about learning 3D is learning how to build the objects that will populate your virtual worlds. Some dedicate their careers to the art of building 3D models. But even though knowing how to model is a definite plus, there are shortcuts.
A large amount of simple 3D consists of bringing in text or EPS outline art, and extruding it into 3D models. Fortunately, most programs have this ability built-in. If not, you can buy a plug-in or application that does this; Zaxwerks’ Invigorator Pro is one of the most popular for After Effects users, while their ProModeler is a stand-alone application for other users. Most 3D programs also make it easy to create primitive shapes, such as boxes and tubes.
A secret weapon to add to these is to learn your program’s deformation tools, which allow you to bend and warp your models. The letter “P” in the figure above shows a common trick of adding a slight bend to the surface of an object: light plays across it in far more interesting ways than perfectly flat objects. Some of these techniques were covered in a previous column.
And just as you can buy stock footage of shots too expensive or impractical to get yourself, several web sites act as clearing houses for 3D models, such as TurboSquid and 3D Cafe. Also don’t rule out hiring someone who has dedicated their career to the art of 3D modeling – this way, you end up with something custom, and most appropriate for your job.
The biggest gotcha of 3D is the time it takes to render a scene. If you thought a few seconds to render an edit transition was painful, wait until you experience a few minutes per frame for a high quality, moderately complex 3D render. This means you will probably need at least one more computer, so it can render while you get work done on your main workstation. In the meantime, plan your renders for overnight.
A related gotcha is the plethora of render settings you have to understand. Virtually every 3D company claims they have the fastest renderer, but most of their comparisons are at their default lowest quality settings – which are often unacceptable for broadcast work. If you see moir© patterns in the textures and crawling dots along the edges of your objects, you’ll need to increase your oversampling and/or anti-aliasing settings to smooth them, with a corresponding hit in render times – indeed, the idea behind oversampling is to create an image two to 32 times larger than needed, and then scale it down to the requested size in an effort to average out these artifacts. These setting usually have much finer increments than the typical “draft” and “best” in a 2D program; run test renders to see the best balance between quality and speed before rendering the entire scene.
Most 3D programs also have a choice between different rendering algorithms. The “Phong” algorithm is a great balance between quality and speed. The downside is that reflections between objects in a scene do not automatically occur with Phong – you will need to step up to “raytracing” for that (again, with a corresponding hit in speed). Different programs also have different subjective rendering qualities, even when using the same settings and algorithms – this is one of the reasons we personally use multiple 3D programs. Use your eyes, and not the manufacturer’s spec sheets, to decide what works for you on a particular job.
Parts of a Whole
The last area of initial confusion is that the separate tasks involved in completing a 3D scene – modeling, animation, and rendering – are often separate programs that are part of a suite. This is initially even more frustrating that having the CG program separate from your editor; there certainly are advantages to integration. However, this can be a strength down the road, because you can use different modeling or even rendering packages if you are uncomfortable with the tools provided by the main program you bought.
But enough theory – let’s get on to some case studies!!!