donderdag 22 september 2011

Level of detail

This week I worked on some bugs of the robot shooter. There was a descent list of fixes, tweaks and optimalizations after the first QA session. The last one, optimalization is what I want to talk about today. In this case I want to look into simple, or high level optimalizations.

First some theory to start with. Games in general are rendered realtime to your screen and displayed using 'pixels'. Imagine having a 3D model displayed close to the camera. It fills up a lot of pixels with its shape, texture and polygons, because it is relatively close to the camera. We could say that how closer the 3D model is towards the camera, the more pixels it uses to display itself. Since it uses so many pixels, we are able to see all of the details of the model.

Now, when moving the 3D model away from the camera, it eats up less pixels to draw itself. Sounds logic right? This means that a single pixel on our display covers up some of the detail we have in the 3D model. We could say that the model is less detailed displayed than it actually is. While it is still being processed the sameway, with all of its details. Which actually costs performance. Therefor we could replace the model by lower resolution version of itself, if the model is on a certain distance. A system like this could be called a 'LOD', or 'Level Of Detail', system. 

Model and image by Owen Shepherds
In our robot shooter it is possible to turn on pixel shaders, which is render bit slower than the regular hardware shaders. In fact, when playing with a descent amount of enemies displayed on the screen using these pixel shaders, there is a noticable delay. We had the same problem in 'Army of the Damned'. Pieter wrote a little 'LOD' system which switches shaders on enemy models. They switch between the regular shader and the pixel shader. This speds up the game a lot and is barely noticable (visual wise) for the user. Since we had the same issues in the robot shooter I took Pieter's system and made it work with the robots. This, as expected, improved the peformance.

There are some drawbacks though. The system in this example uses 'distance to the camera' checks to know what to do. Distance checks can be expensive too, and sometimes can be even more expensive than just rendering extra polygons or models with a bit more complex shaders. But keep in mind that these LOD systems don't require to be updated every frame, which can also speed things up. It is all about finding the balance between direct optimalizations and LOD systems.


vrijdag 9 september 2011

Xform Games - Gotta play 'em all

Hi all,

The last couple of weeks we are very busy completing some of our bigger projects.
And because I think they all deserve a lot of attention and everyone should play them, I like to update you on our latest releases.

Army of the Damned
First of all there is Army of the Damned, the first person zombie shooter that many of us already have written about. The game was released on August 10th in two editions: a less gory version on and a bloody version on 

Check the Army of the Damned official trailer:

And you can play the game here. (Don't worry, it's a link to the bloody version!)

Red Bull Formula Face 
Secondly, last week Red Bull Formula Face (our super secret project) was launched.
A game totally different then our zombie shooter, because you won't see any blood. And it's not just a kart racing game. It's a kart racing game with mimic tracking! Yes, you have to
 control the game with your face. 

Of course, we alse made a trailer for this one:

And don't forget to play at

Robot Shooter

As third, we are working on a Robot Shooter. You can scroll down this page for some preview images. This game isn't released yet, but we work hard to bring you this awesome free online shooter as soon as possible. You may expect this new game in a few weeks.

Between all our busy schedules we still found a little time for a trip to Cologne, to visit Gamescom. It was a long, long, but really great day! With a lot of games. And beer!

Grüße aus Köln!

- Shirley

Particles again!

Dear readers,

The last post by Pieter was also about particles, but as I'm working with different software we have to redo some of the particle work. I'll go a bit deeper into the details of the implementation.

First of all a particle is a simple quad (but in general doesn't have to be) which always faces the camera. To make particles always face the camera we could calculate the transform needed for it to exactly face the camera. But doing this for all particles on screen is a bit expensive. Instead we simply take the rotation values of the camera and apply it to all particles. The effect of this is that not all particles are exactly facing the camera but they are roughly and the difference is not really noticable and it's a lot cheaper CPU-wise.

Second, the spawning and deleting of particles is expensive. Instead we're using a particle pool. In this particle pool there are only unused particles (not in the scene). Whenever we need to spawn a particle we take it from the particle pool, assign properties and add it to the scene. When the particle has to be removed, we're not going to delete it, but we return it to the particle pool. In this way we avoid the expensive operation of spawning and deleting particles.

Third, particles often fade out by using changing the alpha (transparency). Currently the only way we can change the alpha is by changing the material. In this case it would mean we cannot share the material across particles as changing the alpha of one particle would change the alpha of all particles. Using one material per particle would be possible but is very expensive as per material it requires shader program switches on the GPU. So we want one material for all particles but still change the alpha per particle, what now? Well, we can still change the UV coordinates of the particle (quad) and have different images with different alpha settings. We do this by using one large image which has several smaller images in it, see image below. For the first image we change the UV Coordinates to be between 0 and 0.125 (on the 'x' axis), for the second image between 0.125 and 0.25, etc. This functionality is also useful for animation.

Until next time!
-- Stijn

maandag 5 september 2011

Faking a complex 3D background

For our latest game we wanted to create an impressive looking start screen where a camera moves around the main character in a "frozen" 3D environment. We wanted the environment to look detailed with lots of enemies and action going on, but we also wanted to keep the file-size small and complexability low to keep the downloadtime short and system performance high.
We tackled these issues by creating a scene where only the main character and some small effects are true 3D models and where the rest of the environment is displayed as a panoramic texture mapped onto a sphere.

A panoramic texture can easily be rendered in 3ds Max using the Panorama Exporter which can be found in the Utilities panel. The result is a distorted looking image which looks normal when mapped onto a sphere and when viewed from the spheres center. The images below display the rendered panorama image and the Photoshop edit which was used for the final scene.

 The next image displays the final 3D scene for the start screen. 

Here the panoramic texture has been mapped onto a sphere, the spheres bottom has been flattened so it won't look like the 3D character is hoovering above the floor. A plane with the characters shadow is placed below the character to make it fit better into the scene. There are also some extra objects like bullet streaks, impacts and a big muzzle flash coming from the characters gun which exaggerate the 3D effect when moving the camera. The cameras cannot be moved outside of the sphere and moving the cameras closer to the center results in less distortion of the panoramic image.

Here are some screens of the final result.

Cheers Matt