|
Welcome back to Reality!.
In this article we’ll be covering more principles of photorealism. Yes, before we continue we’ll be
taking another look at Principles of 3D photorealism to refresh our memories.
The 10 Principles of Photoreal 3D:
- Clutter & Chaos
- Personality & Expectations
- Believability
- Surface Texture
- Specularity
- Dirt, Dust and Rust
- Flaws, Scratches and Dings
- Beveled Edges
- Object Material Depth
- Radiosity
We’ve covered the first six principles in the previous articles so we’ll be covering the final four in
this one. Before we get started, let’s revisit our example image and the background
behind it. Take a look at Figure 1.1.
'Dwellers'
is a 3D animated short film currently in production at Komodo Studio, a
3D studio in Southern California. It's about a race of cognitive thinking
robotic creatures that were created by an old toy maker named Papagaio.
Papagaio created the first Dweller, Gizmo, in his basement workshop. To
make a short story even shorter, Papagaio built Gizmo out of discarded
junk and parts he scavenged from both new and old items. The Dwellers are
built from actual real-world parts. This helps to establish their photoreal
credibility. This also helps determine the modeling and surfacing attributes
not to mention the mechanics of motion.
Figure 1.1 Photorealism applied!
Figure
1.1 shows Gizmo on Papagaio's workbench where he was created. The scene
takes place at around 1AM in the basement of Papagaio’s house. He doesn't
want anyone to find out about the Dwellers so he only uses a shop light
to illuminate the workbench. He's just finished adding the circuit board,
which is Gizmo's brain. The scene has captured the moment at which Gizmo
comes to life.
Now that we've refreshed
our memories, let’s take a look at the last four Principles of Photoreal 3D and see how they were applied to the ‘Dwellers’ image.
PRINCIPLE 7: FLAWS, SCRATCHES, & DINGS
Nothing makes an object look more artificial that a flawless surface. Even brand new objects have
occasional subtle flaws. Computer graphics have made it far too easy to create perfect objects. The
problem is that reality isn’t perfect. It’s important that you add a little wear and tear to your objects to
make them realistic. I can’t count the number of 3D wooden tables I’ve seen, yet, not a single one has
a dent or ding on its surface. All wooden objects, unless they are new, have some form of flaw. In fact,
even the new ones can be flawed since the deliver people are more than willing to help you apply
them.
Applying flaws to your objects requires that you first explore the nature of the scene. You must
consider the following questions to determine if and when to apply flaws.
1. What is the object material? This is the most important question. The material makes a major
impact on the type of flaws you’ll need to apply. Wood is the most likely surface to be flawed. Plastic
and papers, such as cardboard, frequently have dents and dings. Metal, on the other hand, usually
has minimal flaws. Hard metals, like steel, are typically scratched; softer metals like lead, aluminum,
copper and brass are usually dented and dinged. Objects that are dented and dinged are commonly
referred to as peened. You will see a lot of peened metals in industrial images. The last of the major
material types is fabric. Fabric frequently has rips and tears, though you will see knotted threads on
occasion. Take a moment to consider the objects material when you are applying its surfacing.
2. How often is the object handled? Most objects are handled at some point. It’s necessary to
determine how often the object is handled to accurately apply the flaws. Objects like household
appliances, tools, sports equipment, recreational items, and clothing are frequently handled. These
items are likely to have flaws. I’ve seen many 3D characters wearing jeans, but I’ve yet to see a worn
spot on the knees. If the item is frequently handled, it really needs to have flaws, even if they are
minor. I’ve also seen an abundance of flawless tools, which is very unlikely. Be sure to invest a little
time in determining how often the object is handled before you surface it.
3. Who is handling the object? This is a question 3D artists rarely consider when applying surfacing to
their objects. It’s important to take into consideration the personality of the individual who handles the
objects. If the person is an auto body shop worker, it’s likely the tools are very flawed. They’ve been
dropped, banged and pounded against everything in the garage. If the objects are surgeon’s
instruments, there are likely to be very few flaws. If the object is a child’s toy, you can count on some
serious dents and dings; not to mention the damage to the rest of the scene. Think about the
personality of the characters handling the object before you apply aging.
4. Where is the object located? The placement of objects will determine the magnitude of flaws. The
location of the object can have a large impact on the surface aging. For example: If you place objects
high on a shelf they are likely to remain flawless. If they are located within reach they will probably
have some minor flaws. If they are in reach of children, well, plan for the worst. This, of course, is an
obvious example, so let’s look at one that’s not so obvious. Let’s say you have a car parked under a
carport. The carport has posts that hold it in place. What are the odds you’ll ding one of these posts
when backing out the car? OK, so you’re an excellent driver. Now, what about the guy in the parking
lot that dings your car door for you? Tell me this hasn’t happened to you at least once. Get the idea?
You need to be creative with your aging. The more creative the more likely it will appear realistic.
Now it’s time for a little fun. Let’s ask these questions of an object in the workbench scene. Take a
look at Figure 1.2. Here we have a close-up of Gizmo’s body. Let’s explore the aging of the Swiss
Army knife. Gizmo is made from many discarded parts. We will assume the knife was an object
Figure 1.2 Flaws, Dents and Digns are necessary tools of aging.
Papagaio found in a city street gutter. It probably fell out of
someone’s pocket. This gives us the location. This also tells
us the object was handled frequently, though in this case,
most of the damage would come from falling in the gutter.
We also know the case is plastic so it tends to dent easily.
This covers the object material. The last remaining question
is who was handling the knife. Well, in this case it’s a
question of what, not who. It was the gutter that handled the
knife. Yes, it’s unusual but we must assume that it spent a
while in the gutter before it caught Papagaio’s eye. The water
running through the gutter pushed the knife for some time
before it was discovered. This would add quite a few flaws to
the plastic case. Take another look at the knife in Figure 1.2.
You’ll notice it is littered with dents and dings. This adds a
great deal of natural photorealism to the model.
There is another great example of surface flaws in Figure 1.2. Take a look at the hip joint on the left
side of the image. This object is showing quite a few flaws in the form of dings. I wonder how they got
there? Actually, this object is a new item. So why does it have dings? Well, it was stored with several
other metal items, such as screws and bolts, in the tin can on the left side of the workbench scene.
Since it’s a softer metal, it was dinged by the other harder metal items when they were dropped on it.
It also received damage when Papagaio would shake the can to find the object he was seeking.
As you can see, it requires some planning to determine the proper use of aging. It’s a bit of work, but
the time is well spent when you consider the final result is a truly striking photorealistic image. Just
remember not to get carried away with aging items. If you apply too much aging, the items will tend to
look unrealistic. Just apply enough to break-up the surface. Nobody will buy into an object that is
completely mutilated; unless it’s a post-nuclear war environment.
We’ve covered all the staging and surfacing principles, now it’s time to take a look at the modeling
principles.
PRINCIPLE 8: BEVELED EDGES
What is the one feature that is missing in nearly every 3D object? Beveled edges. Very rarely will you
see beveled edges on a 3D model. This, of course, is a problem since nearly ever real-world object,
that’s manufactured, has beveled edges. In the real-world beveling is done to shave off the harsh right
angle that can cause injury. In the 3D world beveling becomes necessary for 2 reasons. The first is
that the objects need to mirror the features of their real-world counterparts. The second becomes an
issue of specularity. As we discussed earlier, specularity is the reflection of the light source on the
object’s surface. What this means is there will be very subtle specular highlights on the beveled
edges. These highlights become particularly noticeable when the object is animated. We are
accustomed to seeing these highlights on real-world objects. Without them, the 3D object will seem
artificial.
Let’s take a look at how beveled edges were used in the workbench scene. Look at Figure 1.3. You’ll
see a close-up of several hex-nuts. This is a simple yet effective use of beveled edges. You can see a
small bevel on the outer edge of the nut. Notice how the beveled edges closest to the light source
have a specular highlight. These are fairly obvious bevels. Now let’s look at a less obvious, but no less
important bevel. Take a close look at the edge surrounding the blade lever on the exacto knife. You
can see a very fine line of specularity on this small beveled edge. You can even see the specular
reflection in the brass lever grip. This gives the exacto knife a refined and manufactured look.
Ok, those are the more obvious bevels. Now it’s time for a really subliminal bevel. Take a look at the
edge of the sheath around the copper wire that's facing you. You’ll notice a smooth look to the edge.
This is an extremely small bevel, but it helps make the object realistic. When you cut a sheath with
trimmers, you’ll compress the edge where it was cut. The beveling on the 3D object creates that
compressed look. If it didn’t have the bevel, it would be flat. This would cause the end of the sheath to
be washed-out by a specular highlight, making it look unnatural. Even though it’s a minor element in
the scene, your eye would be drawn to it because of the irregular specular highlight. You can see how
even the smallest of details can have a large impact on the photorealism of the scene. While
beveled edges are important, they are not
necessary for all models. Only objects that appear
manufactured require beveled edges. Beveling is
only done to three material types: metal, wood
and plastic. If your model is made of another
material it’s likely you won’t need to add beveled
edges. Now it’s time to explore the other modeling
principle: Object Material Depth.
Figure 1.3 Using beveled edges to create specular reality.
PRINCIPLE 9: OBJECT MATERIAL DEPTH
One of my major complaints about 3D objects is
their lack of material depth. I’m not talking about
their depth on the 'Z' axis. I’m speaking of the
physical depth of the material. I’ve seen far too
many objects where their material was thin as paper. This problem is mainly seen in character
clothing, where the clothing resembles paper hanging off the model’s body. While this is great for
keeping the model cool during the summer, it’s not a very photoreal look. All real-world object
materials have depth. There are only a few objects that have a depth that would resemble the
thickness of a single polygon. This, of course, would be paper.
Take a look at Figure 1.4. Here you’ll see a close-up of the camera box from the workbench scene.
This is a great example of object material depth. Notice how the cardboard tabs have thickness. If you
take a closer look you’ll notice that the surface texture of the box label doesn’t go all the way through
the material, making it possible for you to see the cardboard surface. This is a critical aspect of object
depth in regard to box labels. For a 3D box to appear photoreal, you need to show both the label and
cardboard surface. Real-world labels are printed on the surface of the cardboard, not as a part of the
cardboard. If you ran the label surface all the way through the depth of the material, it would look
artificial.
Now take a look at the other end of the box. You’ll see that there are actually folded tabs that give the
box the look of being assembled. We have all seen way too many 3D, product, boxes that were
merely scanned images mapped to a simple cube. This obviously doesn’t look realistic because there
is no depth to the material. In fact, there is no visible way to open the box. This makes the box lose
all photoreal credibility, in spite of the scanned texture maps. For 3D product boxes to be realistic,
they need to have material depth and signs of being manufactured. They also need to have slightly
rounded edges that display specularity. The hard edge of a simple cube is completely unrealistic.
As you can see, material depth has profound
impact on the realism of a model. Adding material
depth is relatively simple but you must plan
ahead. It’s far easier to add the material depth
when you are just beginning the model that waiting
until you’ve completed it. Take a moment to
consider the material of the object before you
begin modeling it.
Figure 1.4 Object material needs to have depth.
Well, that does it for the modeling principles. Now
let’s take a look at the final Principle of Photoreal
3D: Radiosity.
PRINCIPLE 10: RADIOSITY
Radiosity is the most critical of the 10 Principles of Photoreal 3D. It was saved until last because few
3D programs have the capability to render radiosity lighting. What is radiosity? I’m glad you asked.
Radiosity is defined as the radiation leaving a surface per unit time per unit area. Did that clear it up
for you? Well, it didn’t work for me either. Simply put, radiosity is the indirect light that is distributed
between objects. Most real-world objects reflect light.
You’ll be surprised to know that the majority of illumination in any room comes from the light reflected
off objects, not directly from the source. For example: Let’s say we built a room with four walls,
painted with semi-gloss paint, and a floor tiled with linoleum. We’ll call it a dining room. We then
place a table in the middle of the room and light it with a single light source directly above. Can we
see the legs? In fact, we can. You see; the light was reflected off the table, then the walls, and finally
the floor, thereby illuminating the table legs. This is radiosity. Now try the same test with a standard
point light in a 3D scene. Can you see the legs? Unfortunately, you can’t. The objects in the scene
did not reflect the light. This is why radiosity is necessary for creating truly photorealistic scenes.
Now that you understand the concept of radiosity, you can see why it’s so critical for photorealism.
Unfortunately, nearly ever 3D scene suffers from a lack of reflected light. This has nothing to do with
the artist. It’s a limitation of nearly every 3D program. Why, you ask. It’s simple really. Radiosity is
the most complex lighting formula to create. Not to mention that it also adds considerably to your
render times. The good news is that many of the more popular 3D programs are working on radiosity
solutions. So what do you do until radiosity is a feature in your 3D program? You fake it.
Well, there you have it, the 10 Principles of Photoreal 3D. See, it wasn’t that bad. Well, not too bad
anyway. Now the fun begins. You’ve seen how these principles were applied to the workbench scene
to make it photorealistic, now it’s time for you to apply these principles to your own images.
Get more Bill Fleming - look for his books in the cadmonkey.com bookstore.
Part I - Principals 1 - 3
Part II - Principals 4 - 6
|
