Delving into the software that power our 3D interactive environments or Game Engines can get very technical and daunting.  Considering the number of solutions out there, and the different terminology used to describe them, it can also get very confusing.  In this article we will define what a Game Engine is, a few different types of engines and some of the most popular engines out today, along with what they can do.  I’ve also linked the images in the blog to either their source material or movies showcasing the subject of the image, so there is alot more information directly available in this post than it would first seem. 

What are they, and what do they do?

Don’t be misled by the name, a Game Engine is a software package designed for the creation and development of real-time interactive applications.  Games are just the fun, colorful, entertaining and more visible slice of the pie. Game Engines are used in marketing demos, architectural visualizations, military and medical training simulations, modeling environments, weather forecasts, accident reconstruction, material stress testing and so much more. They also serve as the proving field for new visual technologies and procedures that enhance products to their next generation.

Modern Game Engines are some of the most complex applications you’ll find anywhere.  Balancing a multitude of internal and third party systems, they create a seamless interactive staging for any content and context that feels real and  alive.  A few of the more common systems you’ll find in a Game Engine include a physics engine, artificial intelligence (AI), rendering engine, networking code, and a scripting system for controling in-game events.

Think of it as the ultimate technology in visualization and simulation, with games blazing the trail and blurring the lines of what is real and what is fiction.

In games, 3D guided tours, games, environmental demos, games, architectural presentations, games; no matter the scene, the purpose is always to make the user feel like he’s actually in your demo.   To create the illusion of being part of the world, developers use various techniques to maintain the illusion throughout the experience. In this article we will explore some of them, starting from the inside towards the outside boundary.

Naturally Blocking Line of Sight (LOS)

Almost every 3D environment starts out as a cube, or box shape.  Our job is to make you feel like you are in a fully realized world with no boundaries.   The first step in this process is to limit where the user can go while making it seem like the area stretches on for miles.  Whenever we constrain the player by blocking his progress, the player needs to feel like it’s a natural obstacle that he can go around, not the “end of the level” wall.  The realism of the scene comes in the illusion of vastness, of a complete world, not a shoe box.

Depending on your environment, there are many natural ways to block both line of sight and movement of the player.  In a city scene, you can use a car pile-up, construction, ending the lane in a large building like a library or museum,  police road block due to accident or robbery, the list goes on as far as your imagination.  All of these will make the player feel like he’s in a real environment and serves the purpose of hiding the level border and adding realism.

Perspective Range

The 3D version of “Smoke and Mirrors”.  Think of a level as having 3 stages of distance in order to provide the perspective of range and distance.   The inner level is the playable field, with natural boundaries to constrain the player within it.  The mid-level contains simplified geometry to provide ambience and a sense of scale.  The outer level provides a sense of place or location in the world.  This is mostly in part because of our perspective. As it changes, objects that are closer appear to move faster than objects that are far away.

Another great tool is distance fog. Representing the aspect of visual range, its a rolling fog that increases in strength the further away an object is.  Its also a great way to represent heat haze and smog.

Skydome vs. Skybox

The cherry on top, you can’t have a world without atmosphere, sky and clouds.  There are two popular ways to add this, Skydome and Skybox.

A Skydome is a sphere or half sphere that goes around your entire level.  Effects like lens flare and animated clouds can add dynamic realism to the scene.  It’s very easy and quick to make, but it adds more geometry to your scene and the slight spherical skewing may be noticeable.

A Skybox gives the  sense of limitless sky and distance.  It achieves this through a camera trick. In the development environment or game engine you create a cube outside of your box shaped level. To this you apply 6 seamless skyline images to each of the internal walls of the cube.  With the cube as a reference, you tell the walls of your main level to “look through” the cube walls.  Its almost better to describe it as a portal to a wall in the small cube, but since you can never reach the walls, its a visual illusion.  And as it encompasses the entire level, it feels like a seamless sky.

God rays, or light cones and light bloom effects really add realism to the entire environment. Crisp shadows take care of all the rest.

Keeping all these techniques in mind when you create a level along with game play elements or directional flow of the tour/action will give you a very professional end product.  And the entire reason we do all this work is so that no one notices that we DID do all this work, it has to feel natural.  Mother Nature is a hard task master. ;)

-Juan

In my previous blog we talked about individual vehicle design.  In this blog we will explore vehicle design for games. There are a few considerations you must keep in mind when designing vehicles for games, today we’ll cover 3: Mobbing up, Flavor, and Balance.

Mobbing up

No man is an island, in games, no one vehicle can win a game. Sure there are mega/epic units but by themselves they will get over powered and destroyed. Build vehicles that can accomplish their one particular role well.  Don’t add weapons and abilities “just in case”.  You want the player to build all the wonderful units you made and use them all as one cohesive swarm. Each vehicle performs one role well, and all others badly, but in a group, it can’t be defeated.  The best real world examples are Navy Battle Groups. Large formations of ships supporting one another with specific combat roles. 

Flavor

Here is where we get creative.  All games have factions or races, each one with a particular flavor.   You can’t give a vehicle with laser gatling guns to a nomadic faction that cobbles together their technology from various junk piles.  When building for a particular faction, write down their strengths, weaknesses and what makes them unique from the other factions.  Keep this in mind when you create a new vehicle.  That laser gatling gun vehicle could be an excellent anti infantry unit for a techie cyborg based race.

Balance

Here is where we make sure the game play remains fair for each faction.  The main point of balance is to keep the game fun but challenging.  No one likes a game they can’t win, and no one really enjoyes a $50 game they can beat in 15 minutes with 2 units.

Vehicles for each of the armies in the game usually have a direct corresponding vehicle in each of the other armies and a suitable counter vehicle.  In this respect, you see many armies have an anti aircraft vehicle. You use the flavor to make them really different from each other, but they perform the same function, thereby creating balance.
 
A popular balance method is the old Paper-Rock-Scissors. Where class A vehicle defeats class B vehicle – class B defeats class C vehicle – class C defeats class A.  

You’ll find some games that don’t follow these general rules, and go so far as to create units that one faction will have more trouble defeating than another, attempting to create a 3 way faction balance. These games usually give you other alternatives like super weapons/units, or the ability to pair up with another player using diplomacy rules to defeat a stronger army.

Recently some games have allowed you to customize special modular vehicles, allowing you to alter the role of the vehicle to meet your needs. These tend to be a unique vehicle, and allow you to create some real power houses adding an element of unbalance in to gameplay but creating a really fun (albeit unfair) experience as a trade off.

No mater the method you chose for balance, always keep in mind that each vehicle you create has a role within its flavor/faction/race, and it is best supported be an army of specialist vehicles covering each others’ weaknesses.  All working together to defeat the latest evil nasty nameless army that has threatened your people with ever lasting elevator music!

Till next time,
Fight well, honorably and often
-Juan

-Models in the Tittle Image where created for a GI-Joe mod by Brendan Fennimore, Chris McCune, Jonathan Millman, and Juan M. Del Rio.

 Often once you build a vehicle, you want it to do everything well, to handle any obstacle and take on entire countries by itself!  What started out as a troop transport, is now an amphibious Monster Truck with 8 gun turrets, 18 missiles, 40 rocket launchers and 2 miles per gallon that has the seating capacity of a small stadium!  We get carried away in our excitement and lose focus on what the vehicle’s role is.  In this blog we’ll look at some of the things we usually forget when creating a new vehicle, and talk about the design.

Whats Missing?

 Each vehicle is designed and created to play a particular role or fulfill a tactical need: Recon/Scout, Anti Personnel, Anti Armor, Troop Transport, Main Tank, Long range support, Infantry Support, Bombers, Assault.  Keep in mind that your vehicle is not all of these at the same time, but it can be a hybrid of two.  For example a Fighter-Bomber or an Assault-Tank.

Another thing we usually forget are the things that our vehicle needs:  Engines, weapon power sources, ammo drums or compartments, crew space, emergency escapes, armor panels.  Sometimes we stack so many things in a vehicle and then realize we only left space for a AA battery or 2 sleeping quarters in a space ship with a crew of 30.

You may think I’m crazy but all of these details matter.   These are things you may not notice, but you feel something is missing on the model.   Models that are functional, are more exciting than models that make you guess at how many chickens are powering the 2 square foot engine.

Design

Who’s driving the vehicle?  What environment is the vehicle most likely to be in? Whats the daily routine of the vehicle?   The more you know and figure out about your vehicle the more detailed and alive it becomes.  For example, a large cammoflauged tractor with a mini gun might do great against that alien bovine invasion but do nothing for the wheat field the farmer is supposed to be using it in.  Change it to a bright red tractor with an air conditioned cab, a cooler in the back, a shotgun in a golf bag and a large radio antennae on the top and its the farmers dream vehicle.   

Combat vehicles follow the same line of thinking.  You figure out its combat role and give it 2-3 weapon systems that will help it fulfill its role.  Don’t add things “just in case” you run into an aircraft carrier!  If its a troop transport, its main role is to carry infantry safely to  where they can be best used.  Its main armament is actually armor, second would be a machine gun to provide cover fire,  and third maybe smoke launchers to protect itself and its passengers as they disembark.   Another great addition would be to allow the infantry to shoot through gun ports adding to the fire power of the transport.   Its definitely vulnerable to heavy weapons but it has the tools to fulfill its role.

 Look back over some of the vehicle’s you’ve created so far, and see if there is anything you can add to make it more life like and functional.  Use them as a stepping stone and make the next vehicles you make that much better.   In my next blog, we’ll talk about creating units for a game, faction influence and game balancing.

-Juan

Most of us realize that 3D is an abbreviation for three-dimensional— a concept we apply to everyday life and understand. We live in a 3D world where we can walk around objects, pick them up and see them from all sides. But how does this relate to computers, movies, brochures and all the media we see everyday claiming to use 3D? Perhaps the best way to understand 3D is to start with 2D.

Below we see a 2D image of a bowl of water with floating tea lights. In this image, the artist has given the illusion of depth through shading techniques. It looks as though the bowl is three dimensional, but it in fact is not. The bowl and its surrounding environment are flat, and for eternity will only be viewed from the perspective they were painted (unless the artist sketches from another angle).

Here we see several images of a house created using 3D software. Next to the house is the 2D image described above. As we move around the house we can see that it has depth, whereas the 2D image is flat. Unlike 2D, 3D objects can be viewed from infinite angles, as well as moved from place to place, just like objects in the real world. The images we see below are merely “photographs” (known as renders in the 3D industry) where a 3D artist has navigated around the house on the computer and taken snapshots from interesting angles.

Hopefully, the above example better defines what 3D is, but perhaps you’re now wondering why even bother. Why spend the time to build a 3D object and render it when you can take two seconds to snap a photo of its real-world counterpart and get the same result? The reason is control. It’s true, 3D can produce the same photorealistic effects a digital camera can; however, with 3D every variable can be controlled and manipulated. Imagine a photographer who could control the weather, add more branches to a tree, change the color of the sky, etc., and you would have a 3D artist. In addition to realistic objects, 3D software has the ability to create believable, non-realistic objects as we see often in movies (e.g., the fantasy characters in The Lord of the Ring movies, the cities in the latest Star Wars movies, etc.).

When trying to understand 3D, it is best to think about it in the context of real life. While 3D objects and their 3D environments may not exist outside of computer confines, they behave in the same way real-world objects do, and for all intents and purposes, can be thought of as such.

I recently posted a 3D Model on 3DVIA that asked you to spot the fake crate from the “real” crate.  I am going to go through the process of how I made it in this blog.

Modeling the full detail Crate

This was fun to make! A nice simple crate that everyone can make at home.  I started with a cube in Maya.  Made a few cuts around the sides and pushed the faces in.  I then did a similar step to the top and bottom by selecting the face, extruding, shrinking it a little and pushing it in.

Next I made a long cube . Make it just the same length as the pushed in sides.  I then used the bevel tool to smooth the corners of the this cube along with the corners of the crate. I then duplicated the long cube and scaled it so that it became thinner, and used it as the wood paneling on the sides, top and bottom.  Crate Complete.

 Texturing
This is where all the work happens.  In order to make the simple box look like the real box, I used an occlusion and a normal map.  I’ve spoken about occlusion before.  Its a process by which you create realistic shading based on distance to other object, giving everything a real weight. 
A Normal map, is a process that records the surface normals of an object as a texture, so when light passes over a bumpy object, it changes accordingly. Now because this object is the same on each of its sides we can either render the occlusion and normal maps to a texture (this process is called “baking”) or we can do an occlusion and normal render of one side of the crate and repeat the texture on each side by stacking the uv shells.  I chose to stack my uv shells for this demo.

Layer the occlusion with the wood texture in Photoshop and then apply to both the simple crate and the real crate.  They should look pretty much the same now except for one thing.  You can still tell the box is flat compared to the real crate.  This is where the normal map makes its magic. After I apply it to the box, the light will react to the flat surface as if it had all the groves and bumps of the real crate, making them really difficult to distinguish.

These techniques are the foundations of great game modeling, able to keep the quality of a high polygon model with the speed of the low polygon version.  Go ahead, try to Spot the Fake!

http://www.3dvia.com/jdelrio/media/9159758799ABBD8F

UV’s

UV’s are 2D representations of vertices on your 3D Mesh.  It’s best to think of your 3D model as wearing a skin tight wrap, and in order to put a flat texture on a 3D wrap, you are going to have to make some cuts and flatten it out evenly. After you apply the texture to the 3D model, any wrinkles or overlaps on your UV’s will show up as distortions on your texture. Luckily you have some tools to help you in this process.  In the image above you can see the UV’s on the cylinder on the left, and the UV’s in the editor on the right.  3D in a 2D space.   Notice also how I’ve moved some of the UV’s and it hasn’t affected the shape of the cylinder. UV’s only affect the texture space on the model.  Great UV’s mean great textures.

UV Editor/Unwrap UV Modifier

The UV Editor is your first tool for sorting out the unique problem presented by a 3D object in a 2D space. Your first step to a great texture is to apply a checked material to the model as you fix the UV’s.  This way you can get rid of the big distortions and make sure each shell (sections of uv meshes) is relative in size to the rest of the model. Next up you want to make sure there are no overlapping UV’s.  In the image above you can see the top set of UV’s has overlapping, and the checkered texture comes out garbled and stretched.  The lower set of UV’s removed the overlap by setting the top and bottom caps separately, and by having the vertical faces all laid out flat and uniform. UV’s are normally a longer process than actually creating your 3D Model so make sure you give yourself the time to do it correctly.

Pelting

Like the hunters of old, pelting is a process by which you can take your mesh and flatten it out almost automatically, by selecting seams along your UV mesh.   If you had a 3D model of a bear, and you used a pelting tool, your UV’s  would looks like a bear rug after the process.  Seams are UV shell borders.  Pelting works by taking the seams you marked and unfolding the UV’s through a mathematical process. This images are from the pelting tool from Hydralab.  There are other pelting/UV unwrap programs out there to make your job easier.  Another great one is the Headus UV Layout.

 After all this, the good news is you get to paint canvas like a traditional artist and wrap it on a 3D Model, like the modern artist that you are.

-Juan

D is for Dimension.

You may notice after you start creating in your favorite 3D Application, that you start encountering odd behavior or pinching when you smooth, bevel, or even while still modeling! This is due to the way the vertice’s and edges are arranged to create the model. The following images show a pair of letter D’s.  One is created automatically by Maya and the other is a duplicate of the first, but I’ve redone the Polygonal Flow.

You can see how the interior edges on the right D follow along the curve of the geometry in a natural motion.  The left D, has the same shape when rendered but as you continue editing, you will find that its poly flow will cause a lot of problems.  Also notice how the top left corner is held up by one edge on the right D, this will be come important in the next image, the Bevel.

Beveling.

It is said that in nature there are no hard edges, when looked closely enough even a sharp pin’s tip looks like a ball point pen.  You don’t go around life constantly worrying if you are going to cut yourself on that ridiculously sharp counter edge.  When reproducing real objects in 3D, it becomes necessary to bevel the edges to create a realistic feel and reflection of light.  The Bevel Function takes the selected edges and replaces them with  number of edges in a 90 degree arc, creating a smooth edge look.

Here we start seeing the benefits of a good poly flow.  You notice the pinching and bad corners on the left D,  while the right D looks nice and smooth.  The pinching is created by multiple edges converging on the same point, or multiple vertices too close together.  The bad corners are created when the corner vertex is not “anchored” by an edge. The benefits become more pronounced on the next step: Poly Smooth.

Smoothing.

Smoothing is something we 3D guys love to do, no matter how many polygons it creates.  Its a function by which the 3D application exponentially increases the polygon count, interpolating the distances between vertices and edges.   It usually takes a low poly object and makes it look really refined and smoothed if you have good poly flow.

Ach!  It looks like someone hit the left D with an ugly stick!  The uneven poly flow becomes difficult to smooth because of the uneven distances of the vertices.  This creates bunching, flat surfaces and pinching.  The right D smoothed nicely into a cylinder shape.  Notice how evenly the edges are on the right D.   Its sooo pretty!  You can find some nicely done models through our model search on 3DVIA.

Keep this in mind when you model, or create something from an image and extrude.  Take a few minutes to go back and fix the poly flow of your models;  make sure there aren’t multiple edges coming into one point. Remove any edges that don’t affect the shape of the model and you’ll end up with professional and efficient models.

-Juan

Baking Textures

One way to do it is to fake it, by “baking” the high poly information into a texture and then using that texture on the material of the model. Baking is the process by which render information is transferred to an image file like a jpeg or targa format. To achieve the quality of models you see today in games like Gears of War, Crysis, and Unreal Tournament III each character model has a high polygon and a low polygon version. The high polygon version is used to bake all the detailed information into images, used in the textures of the low polygon version or game model. A combination of renders, including: Normal, Occlusion, Specularity, Diffuse, Bump, Displacement, Glow, and Reflection maps are created from renderings of the high poly model. When used on the game model, they reproducing all that super detail at a fraction of the processing cost to the game engine.

This is done for every object, from a massive space monster to the toilet in the dungeon. The end result is one really beautiful game, with really simple, low poly models, pretending to be super high detailed models. Many of the house models on 3DVIA follow this example, texture over geometry.

Level Of Detail Models

One other way to achieve a smooth 30 frames per second is to use Level Of Detail Models. Have you ever seen objects in a game that look fuzzy from far away but as you get closer they look better and better. This is done by swapping models at certain distances. For example, you have a dwarf that is 2000 polygons, up close and up to 5 ft away he looks great! 10ft away he’s now 1000 polygons and he’s looking ok. 20ft away and he looks like a bop up inflatable punching bag at 500 or less polygons. The texture sizes also decrease per model so that the furthest away the model is, the less resources it uses to be displayed. This is used most often in Massively Multiplayer Online Games. In order to give the player a rich experience they give the area around him all the attention, and anything further away from the player gets less resources.

I’ve had to make a few LOD models in the past and its painful to take a beautiful 2500 model down to a 100 or less LOD model. But in game you really don’t get to see the 500 polygon model as its always the size of a peanut on screen, and as you get closer to inspect the bop up doll, it magically turns into this awesome crazy dwarf!!

-Juan

We have a wide variety of content on the site now, and I wanted to go through and detail what these models that we are all making could be used for. This is so in the near future we can start sections dedicated to each discipline. 3D Models are used in 3 main mediums; Real time render, Offline render/Batch rendering, and still images.

Real Time Rendering

In real time rendering, the computer has to render everything on the screen 30 times per second, including all AI computations, special effects and in-game cinematic. For this medium, efficiency and loading speed is the key for these 3D Models. Conservative polygon use and smaller texture sizes all contribute to a smooth running application, be it a game or Virtual Earth. As an example, a GI-Joe Skyhawk, a game model I made under 2000 triangles. Thanks to increasing technology, today’s game models range in the 2,000-10,000 polygon neighborhood. Another good example of a game ready model is this Boat. For more game ready models in this polygon range visit this 3DVIA page.

Game Model of a GI-Joe Skyhawk, occlusion render.

Pre-Rendering

Pre-Rendering is the opposite, this is used in commercials, movies or animations. All the rendering is done beforehand and then compiled into a compressed movie format. For this medium, the sky’s the limit on polygon count and textures, the main rule it to make it look GREAT. This is the movie version of the same Skyhawk seen above. Its 25 times more complex than the game model, at 55,000 triangles. Any computer would have a hard time rendering this vehicle, 30 times per second, but it would look fantastic in a movie. Another good model like this would be this Spaceship. For more movie models in the 100,000 range visit this 3DVIA page.

Movie Model of the GI-Joe Skyhawk, occlusion render.

Digital Art
Lastly there are still images. These are models you start in a 3d application and complete in sculpting or image editing programs to create a final image. For example, you can start a simple horse in 3Dmax, add all the vein and hair details in Zbrush or Mudbox, add the eyes, nose and other details in Photoshop, then paint in the grass and composite all the different layers to create a beautiful piece of digital art.

So when modeling for a particular use, a game, a 3d world or a movie, keep in mind their medium, and find out any polygon and texture limits beforehand. Then as you make the model you can optimize it for any medium. Not only will this save you work later on, but it will be easier for others to work with you and your models and create fantastic worlds and movies.

-Juan