The Tips in detail

 "Photo-realistic images"

Although the term "image" is used throughout this article, it's message applys to animations as well.

Table of Contents

1. Introduction
2. The Random Aspect
3. In the real world there is nothing nowhere without chaos
4. Character and Personality
5. Expectations
6. Seeing is about beliveing
7. Anchor Objects
8. Surfaces
8. 1 Too smooth to be real
8. 2 Too clean to be real
8. 3 "Reflective" does not imply a mirror
9. Scars, kinks, folds, and other subtle damages
10. Final words

1. Introduction

The first and most important fact about photo-realistic images is to understand that no image needs to be as photo-realistic as possible to be a quality image.

Many 3D artists are obviously thinking that it needs photo-realism to get a high quality image. The reason is in most cases a misunderstanding of the term "photo-realism". If understood as in "creating an image that resembles an actual photography" only one of many aspects is refered to: the whole image's overall visual effect or appearance.

Very often this leads to the artist's struggle for perfection, but this struggle is often misunderstood as the need to be perfect in simulating reality in a perfectly ideal manner. The result is many a time a sterile image.

In this article I'll try to explain what it took me to learn doing 3D images

Creating an image that is photo-realistic is a matter of understanding the priciples of nature and the human eye and their teamplay. There is already a bunch of entries in this section addressing this issue:

• "Using lights with fall-offs makes your scenes look more realistic."
• "Know the material editor."
• "Look at what things look like in the real world."
• "There is nothing like a sharp edge in the real world."
• "Learn how the human eye actually »sees« the world."
• "Use atmospheric effects - with care!"
• "Dirty is Good."
• "Use a Smoke Shininess Map to Add Visual Interest."
• "Use noise for realism."

All these entries are having several issues in common.

I'll start with the distribution of objects, an overall chaos in the cluttering. The basic positioning and aligning of objects in the scene influences the realism of your scene in a very subtle manner.

Imagine you are building a furnished room. If you are placing all the objects with exactly measured distances perfectly aligned with 90°-angles, the scene appears to be constructed, no matter what you do with the remaining aspects of your scene.

This is a basic problem with all systems generating 3D imapes. The tools provided are very tempting to do perfect aligning, 90°-orienting and even distribution of objects. But looking at the real world you will for sure recognize that there is almost never such a thing like a perfect environment. Things are almost never placed at right angles or placed centered in the real world. There is some amount of randomnes with all environments.

(back to top)

2. The Random Aspect

This randomness can be summerised as "chaos" or "entropy". Chaos doesn't mean complete disorganisation. Chaos means the lack of perfect organisation. In the real world (that's where the pizza-delivery man comes from) there are laws of physic that prove (dramatically simplyfied, I am not a physician) that creating organisation consumes energy and by doing so it contributs to entropy (aka: chaos) that again counter-effects the order.

Order means an extremly fragile balance of all factors also known as "uniformity". Breaking this balance is of course very easy, since it only takes adding literally any amount of energy to any aspect of this balance of uniform aspects. If any amount of energy is added anywhere, the exact amount of energy needs to be added to the exactly opposite point of the uniform construct to keep up the uniformity. If this doesn't happen in the very moment the first disturbing energy is added, the occuring inbalance spreads through the neighbouring aspects, causing a domino-effect of disturbances in the uniformity. To re-create the uniformity energy needs to be invested again. The amount necessarry is huge compared to the amount it took to destroy the uniformity in the first place.

All this babble means: There simply can't be such a thing like a perfect alignment of objects in the real world.

This is a law of nature you can easily observe yourself:

Example.

Get yourself three matches. Find a table with some free space, 5 x 5 inches should be sufficient for our needs. Drop the matches on the table (the free space).

How are the matches placed after all comes to rest?

Now place your three matches with the same distance to the nearest edge of the table without any tools.

Hard, isn't it?

Even if you used a ruler or something like that, you'll still find it hard to place them at the exactly same distance from the edge. Now have a look at where the heads of the matches are pointing at. Do they all point the same direction? Chances are big they don't. Are they placed tip-to-edge or side-to-edge? Are the matches aligned parallelly in respect to eachother? I bet that your answers will differ.

Now swiftly blow at the matches and simultaneously try to keep them at their aligned positions with your hands.

You will for sure notice that it took far less time, concentration and endurance to destroy the order than to set it up. And you will have noticed how hard, if not impossible it is to keep them in place although you thought you knew all factors (like the strength of the airstream you created and the motions of your hand, etc.) Even if you lay your hand flat on top of the matches, what happens the very moment you try to remove your hand to check if they are still in place?

Do you now believe me? Can you follow me with what I am intending to show? Chaos is everywhere.

(back to top)

3. In the real world there is nothing nowhere without chaos.

"How does this address my renderings?" you might ask. We are talking about photo-realism. Photo-realism is about an image of the real world that is as realistic as posible. If there is no perfect order in the real world (as shown above) how can there be perfect order in your photo-realistic images?

You need to reconsider the way you are setting up scenes. Scenes are set up using tools provided by the software you are using. These tools (along with computers and their use) contribute to a very structured, linear and ordered way of thinking about setting up scenes. It is very common to construct objects and sceneries as if one was doing an "oversized" technical blueprint. This technical way to see and construct 3D images is one of the main reasons why most renderings are not photo-realistic.

Remember: Reality does not mean total lack of organisation. Reality means lack of perfect organisation! When designing scenes that you intend to be as realistic as possible remember to add small amounts of disalignment. You should understand this as a slight factor of randomnes (or "noise") within placing objects.

Look at your very own workspace. Allow the impression to settle in your mind. You will for sure recognize, that it is more or less organised, like CDs piled here and notes piled up there, pens and pencils stocked there and so on. But although organized in the whole and in the general, it is chaotic in every detail.

Now remember the workplace of a friend. Do you notice that the same applies here, too? But further you will for sure recognize an elemental difference in how it is organised. This brings us to another point: There are no two really identical sceneries or objects, as well as there are no two truely identical human beings. But what makes everything appear different? The easiest way to understand all this is to consume all aspects under the terms "character" and "personality".

(back to top)

4. Character and Personality.

Everything has its own character or personality. Things can be brand new or worn off due to heavy duty. Environments can be inviting, ugly, spooky, heavenly, and so on.

In terms of 3D images, really understanding this fact is essential. The character of a scene and its objects is what really dictates the image via its mood. This mood is the result of the teamplay of the environment's character and the contained objects and their individual characters.

This character is not only how things look like. It also reflects in the way things happen and what makes your objects do or not do anything. As an example you should compare texts you have written by hand with those written by others. Those texts do not only vary in their visual appearance but also in what they are about and how they are written. Some people are writing using their right hand, others are using their left, some can use both. Some texts are written using a pencil others are written using ink and so on. In terms of organisation, especially when speaking about organisations caused by human intervention, this individual character is very prominent. The computer nerd's desk for sure differs from the desk of an occassional typer. The jungles and rain forests differ from those forests that are cultivated (by humans.) Villages differ from downtown Manhattan. You get the picture.

The problem is that with 3D images you start with nothing. You have to construct everything on your own. As opposed to photographies there is nothing you can just look at and take a snapshot. You are starting the other way around. You are having an idea and start getting this idea to your media. Within this process you need to remember the character of each aspect. But speaking about the characters we stumble about a new aspect of the initial question of "Why does an image look real?"

(back to top)

5. Expectations.

It's the expectation of every viewer. This expectation-issue is a huge aspect in Art. Let's start simple. Stereotypes dictate the way we recognize everything surrounding us in our everydays life. These stereotypes are the result of experiences and result in follow-up expectations. If you touch the fire, you burn yourself. Once burnt, you will avoid touching flames again since you expect all other flames to be burning hot, too. In this case "fire" is the stereotype for "burning hot".

Such stereotypes are not only the result of first hand experiences but also result from conditionings of our own environment, like in example what movies we have seen and where we grow up. Those stereotypes are omnipresent. Not only with the "fire - hot" example but also with "environment - objects present" and "object - look" issues. Let's dive through these one by one.

Every scene results in expectations regarding what should be or is likely to be present and what should not be or is not likely to be present. Considering a kitchen for example, most people would agree a fridge, kitchen table and kitchen tools to be very likely to be present. On the other hand a power drill in example won't be expected there, but it might still be possible to find one here, especially if some aspect of the scene contributes to this. These aspects can be (generally) devided into three categories: obvious reasons, unknown or random reasons, intended to tell a message.

An obvious reason for the power drill to be present would be in example if there was some kind of repair or construction going on (be it soon starting, currently in progress or just finished) in that kitchen.

The presence of objects due to unknown or random reasons can't be explained by just looking at the scene. Who ever looks at the image must explain the presence of such objects him- (or her-) -self, but it is not likely that the resulting idea matches the artist's intention. The power drill from the above example could of course be present in every kitchen in every industrialised society for an almost unlimited number of reasons. But why it is present in this scene must remain a mystery for the observer if the artist placed it there for no real reason. This way the whole image doesn't tell much more aside from "I did a kitchen and a power drill."

Objects present to tell a message can be quite close to the objects present due to unknown or random reasons. But by some aspect they are telling something. If in the above example the kitchen was perfectly clean and styled like in the early 60ies, but the drill was a post modern one and looked like being used a short time ago, this might give the scene a whole new look and overall impression. In scenes like this it is very likely that the artist wants to tell something more then just "Hey look, I did a kitchen and a power drill!"

What do we learn from this in matters of photo-realism? Well, as you would agree objects present due to unknown or random reasons (from the view of any observer) are very disturbing in terms of an image's character. A very good example can be found in almost every gallery of the prospering newbie: Models rendered in oftenly completely meaningless settings, somehow steril environments and somewhat strange compositions. Although the models on display in these images might be of outstanding quality the images themselves are far from being photo-realistic.

A good example are my own images "gear" and "Sat Dish".

Regarding "gear": Although the model itself is quite "ok" (for a newbie; this was my first raytraced image ever) in terms of quality, the scene's composition doesn't make any sense. The primary object in that scene has no obvious reason in its setting. Anybody looking at that image can't read any real message from that image, because it was a completely random result. Even I can't explain what the object is or what it does, since I designed it without any purpose aside from testing materials. The floor resembles some green cloth wrapped over something but it somehow doesn't fit to the red-brown background and both doesn't match what one would expect in a setting for such an object. If this object had been placed in a more futuristical environment, like in example the bridge of a starship, it would not be that distracting.

Regarding "Sat Dish": In this image all objects present do make sense in their set and setting. But still it lacks that final photo-realistic touch to make it a great image. Aside from some aspects that I will cover later in this document it's mainly the corelation of the radartower in the image's center and the plane taking off to the upper right. The plane is a futuristic one. It resembles a current plane but it lacks an aspect prominent with almost all current jet-planes. It's the tail. The rear part of that plane in the image differs largely from what anybody would expect it to be: it is too short (refer to image "Starfighter" for a more detailed view.) Same is true for the exhaust: Yellow, orange and reddish colors are expected to be found with such an exhaust, but not these tintings. They are far too intense. Although most people I spoke to did not mention this mismatching aspects as the very first issue, it was present with every statement: "The rear of that plane somehow doesn't look right..." The strange looking plane does not catch the eye as much as the tower does in the first place. It is much more likely to be consumed as "Oh, a starting plane", the focus of attention then returning back to the tower. But your mind still fiddles with the impression and starts bothering you with "Something is not right in that image". This disturbance gains further attention via the unintentionally very signal-like exhaust colors.

An example of the exact opposite is the image "Diplomat". This is a good example for objects that don't seem to fit to the scenery but are nevertheless present. The steel nuts in the left middle of the desk are not expected to be there on first instance, but they fit into the rest of the scenery since they seem unlikely but not impossible within this setting. They do not intentionally tell a message on their own, but they fit into the overall setting. Their presence is supported by the book (and its message) in the very center and the "Diplomatenspiel" and the glasses to the right.

An almost perfect example of photo-realism are the images "Chest" and "Hand", because both of them are telling a story. All objects in them are having a meaning and match closely to the individual settings. Although both images are living from their unexpected elements, those elements are as one would agree them being possible.

After you have learned what role the charcter plays for the image's photo-realism, we need to have a look at what the viewer expects objects to look like. Certain aspects are expected to be found along certain sets and settings. The spooky graveyard needs to be dark and foggy, the heavy industry needs spilled oil and paint, rust, and worn off colors, the scientist need this typical white cloth, and so on.

These expectations are again the result of being conditioned as described above. Most people would agree to expect the bridge of a spaceship being crisp and shiny, having a lot of blinkenlights here and there, and don't forget the metallic and chrome aspect. Any bridge on display would not look realistic if it lacked these elements.

Never underestimate how we are mislead by our expectations. Most people do expect dinosaurs to be huge, having mostly those common reptilian scaly skin with more or less earthy colors as seen on TV. When presented an image showing a small or colorfull dinosaur, most people would consider it being artificial. Telling those people the latest facts about how dinosaurs look like before showing them such an image would result in more people to agree to the image being realistic. If you'd tell those people these facts after showing them the image, it is very likely that they won't believe in any of those two.

Expectations are having big power in everybodies life. Lean back, relax. Do you hear that hissing sound the fan of your machine is producing? Do you hear that faint ticking sound your harddisc produces? And what does that periodic click mean?

Even if your system is perfectly quiet, many would agree to suddenly hear those sounds I just mentioned. Some might even suddenly "hear" an unusual periodically ticking sound.

As you can see expectations do play a keyrole with photo-realism. You need to conform to certain expectations when creating photo-realistic images. If the viewers expectations are met chances are high the image at least appears to be photo-realistic.

(back to top)

6. Seeing is about beliveing.

Photo-realism lives from the fact, that the viewer is able to believe in the image. Believing does not mean the viewer to actually accept the rendered objects to be existing in the real world. This would be the best possible effect to achive, but not always possible. The best one can try to achieve is a silent agreement with the viewer in the point that the items and the scene itself being on display could possibly happen to exist somewhere.

But what makes somebody believe in an object? The objects themselves need to be believable. The environment needs to be, too. The teamplay of both must be believable as well. What carries the believability? What contributes to and what reduces believability? Basically it's all about objects, textured surfaces, light and visibility.

Objects are basically shape and textured surfaces. The power drill from the earlier kitchen example becomes believable via its shape. The textures add to the realism, but the object is defined via its shape and what the viewer has stored in his mind along with this shape, again a stereotype. The textures play a somewhat lesser role here. But by changing these textures the artist has a vide viarity of options.

By making all the textured surfaces looking like toy-plastic the whole object would very likely start looking like a toy, although the shape itself did not change at all. By keeping the "standard" power drill textures and twisting the shape slightly it suddenly resembles a ray-gun or a power grinder. This clearly proves that in this case shape dominates over texture in means of as what something is recognized. The textured surfaces on the otherhand are playing a fundamental keyrole in the object's believability. A power drill made of glas is less likely to be believed in being a real tool compared to one having the "right" textures.

To make an image photorealistic is to combine the expectations with believability. There need to be certain points in the image that allow the viewer to hook his believing to. Such hooks are objects in the scene that are having undeniable photorealistic attributes. These objects are very often simple in terms of the model itself with texturemaps taken from the real world.

An object needs to be familar to the viewer to be a working hook for his believing. Objects can vary dramatically in shape and texture - such as cars - but it needs certain aspects for the viewer to recognize it as what it is. If an object has wheels at the right places, doors, windows, front and rear lights and an overall shape that fits to the viewers expectation of the shape of a car the viewer would accept the object being a car regardless of most details (like color, sharp or rounded edges etc.) of the actual model. If the keyaspects would be altered massively - like i. e. the car not having wheels and windows - the viewer would possibly not recognize the object being a car. This also applies to the objects textures. If an object is not recognized then it can't add to the viewers belief in an image.

The proportions of an object in respect to the remaining scene are playing a keyrole for the viewers belief. Let us again take the power drill in the kitchen. Imagine the kitchen being perfectly modelled and textured. Imagine the power drill being placed on a table. What happens to the scene if the proportions of that power drill change? If the power drill becomes too large, the scene would obviousliy resemble a dollhouse at some point. If it becomes too small, the viewer would either not recognize it being a drill or consider it being a micro-toy. This will disturb the credibility of the overall scene although there is quite some tolerance. But if these tolerances are stressed too much and add up one to many, the viewers trust can't hook to the image at all and the image looses its photo-realism.

In most scenes the models are having quite the right proportions, but there are very often either proportional mismatches in respect to other objects proportions or their own textures.

Imagine a can of coke. You will for sure be able to do the model quite easily by using ie. loft- or box-modellig techniques. Now imagine having the texture as a scanned image. This real-world texture applied too this model would dramatically add to the objects credibility. But: now it shows if you met the real proportions. If the object is to long or its diameter to wide, it would still resemble a can of coke, but it would not look real. It would not look real, too if its proportions would have been correct but not matching to those of other objects in the scene.

How any object is looking like depends on scene's style. Comic-style objects are not likely to be expected in highly realistic scenes and vice versa. Antique scenes are not likely to have ultra-modern objects.

Another point is that viewers tend to believe in objects and sceneries they already know somehow. This again is a matter of conditioning. The more objects (or more generally speaking: sceneries) you are using you believe in yourself, the more likely it is somebody else believes in them, too.

(back to top)

7. Anchor Objects

Very closely related to the believability of your scene is the presence of so called "anchor objects". Anchor objects themselves are objects easy to recognize along with undeniable photorealistic attributes. The viewer usually knows such objects by first hand experiences. The complexity of those objects and their suitability to be an anchor object are directly related: the simpler the overall object, the easier the recognizability.

The more anchor objects are in the scenery, the more they will add to the scenes photorealism. Anchor objects can considerably disturb your scenery if they missfit the set and setting (see above.)

Don't confuse "the need for presence" with "the need for focus" in your photorealistic scenes. Even though anchor objects themselves can be the center or main aspect of your sceneries, they don't have to. They can be quite small in size and being placed literally everywhere in your scene. The important point is that they are "there".

(back to top)

8. Surfaces

All surfaces are having their unique characteristics. Several of these can be explored by touching them (like i.e. temperature, sharpness, softness, etc.) some of them will be determined by just looking at them (i.e. reflectivity), others by sound (i.e. vibration) or even smell (i.e. wood). Combinations of these make an object to appear "real" to any human. The problem with 3D graphics is that you can't make use of any features addressing smell or touch, instead you have to rely solely on the optical aspects and properties.

Animations can (and should) include sound to address more sensual aspects and by doing so gaining a much higher acceptance from any viewer. Within the most 3D software products all aspects of any surface are summarized under the texture, but remember that this is actually sort of an abbreviation (if not limitation) of what "texture" is about in the real world.

So what is wrong with the basic objects produced by your outstandingly expensive top notch 3D modelling and rendering software? Basically, there is nothing "wrong"... it's more like that there is exactly nothing wrong with them. And that's a problem, as you will see soon.

(back to top)

8.1 Too smooth to be real

The vast majority of rendered objects are way too smooth to be real. Sure, there are perfectly smooth surfaces out there in the real world, but they are rare. Extremely rare. Look around and count those surfaces in your immediate surroundings that are perfectly smooth. Now do the counter-test: How many surfaces are not perfectly smooth?

As you can see the vast majority of objects is having not-so-smooth surfaces. Even those surfaces that you would agree to be smooth when touching them very oftenly don't look smooth or vice versa, like i.e. polished wood, marble, steel. All materials and objects are suffering from weathering (due to the environment it is in) and deterioration (due to i.e. being used or due to aging) in the real world. All this results in scars, punctures, little damages but also in traces of being grown or used. The aspect of being grown is especially true for any natural object / component, like i.e. wood.

Spend some time examining those materials you are intending to use in you scenes. How do they look and how do they behave? How does iron rust? How smooth is smooth wood? How does that look like that disturbes the smoothness?

The aspects of the materials usually change very often due to the way they are being exposed to light and due to the way the material itself is being stressed, i.e. by being bent, torn, exposed to great heat or cold.

The best way to learn about any given material is by observing it in it's intended environment. Most materials can be found outdoors, some might be a bit difficult to find. But trust me: it's worth the time spent staring at some withered stone, a rotten tree, or some polished metal plating to get the idea about it's consistence.

And while you are at this: take a camera with you and take pictures. This way you will

1. learn about perspective, lighting, and stuff
and
2. grow your very own and very individual mapfile libary.

The most common way to add some kind of irregularity to a surface is by using Bump-mappings. Those Bump-mappings do not add any real irregularities to the object (or mesh) itself, but they rather make the light hitting these objects behave as if the object had an irregular surface. In other words: Bump-mappings are trying to fake the eye. And in the most cases, Bump-mappings are doing a great job actually doing so. But from time to time you will recognize that you can achieve considerably better results by actually deforming the object's surface, especially when it comes to flat or narrow angles or extreme cloese up views.

Have a look at the Crystal image.

While the crystals themselves look quite beliveable at their surfaces, you will agree that this is not true for their edges and corners. This is due to the fact that the surfaces of crystals have been composed using bumpmaps. It would have been wise to use real deformations here to achieve that final twist towards photorealism. On the other hand those crystals are having quite complex surfaces, and adding deformations to any surface is about adding verteces and polygons to the object. For getting this particular scene done I decided to go with the Bump-mappings, since "true" deformations would have pushed up the vertex- and polygoncount to incredible amounts, directly resulting in really long rendering times.

Basically it's a matter of balance to know when to use Bump-mappings and when to actually deform an object. Bump-mappings can solve a big deal of situations, actual deformations can appear more realistic. Bump-mappings are generally cheaper in terms of resourcess you will need to render the image, real deformations will not only need more RAM and HDD space, but it will also take some time to get them done as well.

As a rule of thumb: The more a surface is exposed, the flatter the angle of view towards that surface is, and the closer the surface is to the viewer, the more likely it is that "true" deformations will achieve better results regarding photorealism.

Deformation is not the only aspect of surfaces. Another truely important one is closely related to this and it is called "the sharp edge". I already asked you to read "There is nothing like a sharp edge in the real world.". If you did not do so yet, do it now and return here afterwards.

Now, having in mind that there are (almost) no perfectly smooth surfaces nor perfectly sharp edges an corners, we need to cover the remaining aspects of how surfaces are appearing. Let's start with the clean and tidy sterility that sticks to most computer-generated objects.

(back to top)

8.2 Too clean to be real

As I already asked you regarding smooth objects and surfaces, I now want you to have a close look at any object you want. Does it appear to be "clean" to you? If yes: How clean is "clean"?

Again you will agree with me, that many objects appear to be clean, but at a certain level, there is some dirt. Or at least some dust. Or fingerprints. I've described an easy way to apply some dirt here.

The point about dirt is that dirt is a matter of fact in the real world. With a photo realistic image you are about to build some scenery that can be agreed to be real only by looking at it. The lack of dirt or dust (etc.) directly results in a somewhat sterile looking image that viewers will find hard to believe in.

Again there are some rules as how dirt applies to a surface. The way something gets (and stays) dirty depends on many aspects, like i.e. the angle the surface has in respect to the floor. Something standing straight up is less likely to collect dirt, where something laying flat is more likely to gain some over the time. A rough surface will be more adhesive towards dirt than a smooth one.

Another point is the dirt itself. Mud can be amazingly adhesive and sticky. Windows will blind over the years by gathering dust (and other stuff) from the air. But where does the dirt apply to? And why? Dust will pile up in brackets and spread out from there. Splashed mud will cover areas and pour downward (toward the gravitational "down".) And dirt will always add up. One kind of dirt might completely cover another one, but generally they won't wipe out eachother. Again: looking at objects "in the real world" helps big time getting the picture.

(back to top)

8.3 "Reflective" does not imply a mirror

Having learned about the main issues regarding what happens to surfaces in the real world, we need to address some crucial issue: The way all this affects the light.

While the general lighting of a scene is a matter that's best discussed in a seperate entry (to come, bear with me), the way light bounces off from any surface is important within this context. The light is carrying the information regarding any object and its properties. Images are a matter of visual aspects and "visual" means "seeing".

Objects, that is surfaces, are reflecting light - with the exception of black holes. The amount of reflected light varies due to the material's and the light's properties. The human eye can get an impression about some of the properties of a material by examining how light is reflected by an object. We are using the overall reflections for measurement.

Light is influenced by it's environment. Simple, sheer white and clean light is pretty rare. Look outside. Is the light you can see there really white? I bet it is not. When Light is hitting some object, the light is affected by the materials physical properties. Light can be reflected, bent, scattered, fractured, or absorbed. Or combinations of these, depending on the actual material. Each behaviour will result in a visual effect.

We usually don't preceive an image as an accumulation of light. It is more that we are "seeing" light in those bright spots, where it is reflected with maximum brightness. These spots of maximum reflection are refered to as "Highlights". We usually do not pay close attention to those highlights when looking at any given object but nevertheless we do see them. Highlights are carrying sort of a sublimal message telling us "this object is soft" or "this object is wet" or "this object is brand new" or something like that.

By bouncing off a surface the light adopts information about the surface it just hit. Whatever else happens to a ray of light it will at least get some modification in terms of color: the light is tinted. The exact tinting color depends on the surface. But aside from the tinting, the surface determines wether or not a highlight will be sharp and clear or dull and vague.

A rule of thumb: Rough surfaces will have tiny highlights. Soft materials will tend to have dull or no highlights at all. Rounded or curved surfaces will have smaller highlights than plain ones.

Since every object is reflecting light, we are able to see them. Due to their highlights we are able to tell something about their materials respectively their properties. You need to pay attention to have the right highlights with your materials. Keep in mind that highlights are affected by liquids and the light's angle of attack.

But when we do speak about bouncing off lightrays, why doesn't that imply a mirror? Well, basically everything is a mirror. But there is always a factor of loss in the mirrored image. If you have ever played with raytraced materials you will agree that it is quite easy (though timeconsuming) to get realistic reflections using this technique. By adjusting the amount of reflectivity you can vary the way the object appears. But not all light is reflected from every object. That's why we are not living in a world of mirrors. The higher the finish of any given surface, the more likely it is that the surface acts like a mirror, reflecting it's surroundings.

There is a wide range between a real mirror and soft rubber or charcoal. The key is to know when to use mirror-like properties and when not. You will again have to have a look at the real world. To get things running, try modelling and texturing a poolball on a pooltable. After that try texturing a wheel (rim and tire) of a car. Then try a window and a TV-set (switched off). These exercises will teach you the difficulties related to highlights, reflections, and mirrors.

(back to top)

9. Scars, kinks, folds, and other subtle damages

As stated before: surfaces need to be deformed in some way or another to be belived as being "real". An utterly flawless surface will look sterile and artificial. Remember that objects are exposed to some environmental effects. Good examples are wooden surfaces and metal objects. Have a look at some examples you should be able to find in your household, like your kitchentable, some canteens, etc.

By looking at these you will notice some important facts: The way these objects are showing their respective subtle damages directly refers to the way they are used and where they are placed. This again relates directly to what you have to keep in mind when modelling these objects.

The main issue dictating the signs of aging and use is the material the object is made from. Soft materials will be more likely to have folds and wrinkles. Stiff materials might have cracks, bucklings and harsh kinks. The thickness of the object plays a keyrole here, too. Tinfoil i.e. will look different than 2 inches thick stainless steal plate regarding its signs of use and aging. Same is true for the color coating of a brand new car compared to a (usual) car that is several years old.

The amount of use and the way of being used plays a keyrole for damages. Old cups for example are likely to have cracks, chopped out pieces, or even broken off handles, especially when being used on a daily basis. On the other hand the signs of use will differ a lot from those you will find on a wrench used by a mechanic to work on truck engines. Roadsigns are usually damaged in a different way, compared to framed panitings on the walls of your livingroom.

By considering the use, make up your mind about who is actually using this object for what purpose? This dictates big time regarding the overall condition of the object, not only regarding its actual shape, but also its placement. As soon as you made up your mind about the "who" and the "what for" you will be at the point where you will need to consider what the given object is used for in your image. This again might add to the way the object appears and looks like.

Talking about damages: Perhaps some damge was repaired? Almost every repair leaves traces in the appearance of an object. Try to figure out how an object might be repaired, if at all. Perhaps the bent blade of a knife could be repaired by bending it in the opposite direction, but that would almost certainly leave some kind of visible folding or bending line. If an object was damaged that seriously that it comes to either repairing or dumping it, be real. What is likely to be repaired and what not? If something is repaired, how likely is it that this repair is done the "DIY"-way? A cup - i.e. - would rather be dumped then repaired if damaged seriously. A diamond collar would most likely be repaired, but almost certainly by some professional jeweller. Ladders can be found fixed DIY-style, ranging from "close to not fixed at all" to "almost new". You get the picture.

(back to top)

10. Final words

This article has been rather long and I hope I got some main issues covered and provided something usefull for you. A really big issue, the light, has only been touched very slightly (not to mention the radiosity), and regarding "sharp edges" I would like to point you here (I would repeat myself if I would cover that issue again.) There is also the huge area of "realistic modelling" regarding objects, that I did not cover: humans, animals and other living creatures. That is a problem on its own which I will eventually cover later. Generally speaking, I'd like to stress you to practise. Do not expect your first attempt being the all time greatest image. Ask others about what to do in another way or what irritates them about your images. Artists become blind for flaws in their own works very often (yes, that is true for me, too.)

(back to top)