Space Game Update #3 – Water & Improved Texturing

This is a quick update post to reassure anyone who’s interested that the project continues to be in development. Most of the updates continue to be purely visual, although there will soon be a return to gameplay mechanics. The most immediately noticeable improvement is that the planets are no longer entirely static — oceans now freely move around! Although many will appear much like the short video below, the water system supports oceans of any color. This will allow for exotic planets with oceans of methane or other crazy environments.

I will update this post with close-ups of the new texturing system, but I will sum up the changes for now. Planets now support a palette of up to 64 textures determined by slope and biome. This pool of textures is drawn on to create more interesting landscapes with changing details. Up to four textures are supported at any given terrain vertex, which allows for large-scale variation across the planet.

I hope to delve more deeply into some of the updates expected for v0.4/0.5 in an upcoming post, but that’s all I have for a quick update today. Enjoy the water eye-candy!

Space Game Update #2 – Lens Flares & Orbit HUD

This week was a busy one! Both graphics and gameplay saw some new features implemented, and many bugs were squashed in the process! Visually, the most prominent of the updates are the lens flares and star glow fixes. In the past, stars would glow whether they were occluded by and object or not. Now, stars will glow and produce lens flares if you look towards the light source. Any object (planets, spacecraft, etc) can occlude the light in real time. This produces some very neat scenes where the star flickers as objects pass in front, or when you get hit with the morning sunrays.

The lens flare effect was accomplished using a nifty OpenGL feature called occlusion queries. It works like this: A quad is rendered at the star location at the end of a frame. The depth mask is activated so that the depth buffer is not written to. Depth test is left enabled and the query is sent to the GPU. At this point, we would have to wait for the occlusion result. However, for lens flares and star glow, a single frame delay is not very noticeable except at low frame rates. So, we simply use the previous frame occlusion query result. This prevents stalling the GPU to transfer data to the CPU and produces very nice results! The occlusion query provides information about how many framebuffer samples passed through the shaders. This let’s us know if the star is occluded, and therefore whether we should render a glow and flares. Once this is determined, a simple screen-space rendering algorithm is used to create the lens flare effect. Look at the screenshots for the very cool end product!

Gameplay also had some updates. As you can see in the above screenshot, there is a new HUD object available, Orbit HUD. This HUD computes your current Keplerian orbital elements and displays them in real time on a 3D globe. This is a very useful display for calculating orbits and trajectories. And now that you can easily determine your current trajectory, it is much easier to fly the spacecraft into an actual orbit. The making of this HUD exposed some issues with the PI-PR (Planet Inertial and Planet Rotating) reference frame equations that have now been patched. With the addition of a prograde indicator, it should be much easier to perform burns and control your orbit.

With these two updates, the game continues to feel more like, well, like a game everyday! It has a long way to go, but each update encourages me to add further features. As gameplay elements are implemented, I hope to start defining what kind of game this will be. The objectives, the challenges, and the general experience are all beginning to unfold. I’m excited to see where the game takes me next. Hopefully, I can come up with a fitting name!

Space Game Update #1


Hello again! I come bearing news of many positive updates to the Space Game engine. In the last few weeks, I have managed to flesh out some of the more difficult features that will pave the way forward. The game itself continues to be much more stable with every passing week. Let’s hop right into some of the most interesting improvements from the coding dungeon.

First of all: terrain generation is getting much better. It now includes a fairly accurate and efficient vector data system. Right now that just means roads/runways, but in the future it could open many exciting avenues (pun intended) for surface level details. The system conserves memory usage by making use of higher resolution terrain pages. This of course revealed a huge bug in the high LOD terrain request pipeline where precision appeared to drop at sub-meter resolutions. Luckily, this was a fairly simple fix, and the results are great! Stable sub-meter page resolution allows for much more detailed terrains without using significantly more memory. As long as your computer can handle it, sub-centimeter details are not out of the question!

Terrain page selection has also seen some minor updates. Loading into the world was pretty painful for the last few minor versions because of the way terrain pages were requested and selected. The previous algorithm was aggressive; tiles were selected before their height data was parsed. A maximum bounding box extent was assumed for these pre-loaded tiles, which caused a cascade of tile selection in the quadtree. The new algorithm will abandon selection if a tile is not yet loaded. I had avoided this in the past as it does not guarantee that quadtree nodes are always within one level of each other as CDLOD requests. However, my terrain LOD scheme has become a kind of CDLOD/Chunked LOD hybrid with skirts, which allows me to hide most of the terrain issues while also taking advantage of CDLOD’s morphing on the bulk of the terrain.

Another huge improvement was in the GPU-side terrain generation shaders. Terrain page data is generated on the GPU on a 32-bit float texture with four components (RGBA32F). The height was stored in one component and the normal was stored in the remaining three. Using 96 out of 128 bits on normal data was pure insanity. I can afford to use much lower precision when processing normals, so I decided to pack them into a single component. The X and Y components of the normal each use 16-bits and are packed into one 32-bit float. The final component, Z, is determined when unpacking by working backward from unit vector normalization. In normal space, the Z direction is always positive, so it is easy to infer the sign. After these changes, I have 64-bits available in the data page for accessory terrain data. I have not quite worked out how I will be using these components, but it will likely be in generating a more visually interesting terrain.

Physics feel pretty stable. There are some open issues with the physics system, but overall, you can take off, land, crash, do flips, or whatever and you probably will not notice the deficiencies. Although the collision system will have to be written before a major release, it is fine for now. I have some plans for an orbital dynamics HUD display that will help visualize your current orbit. This is where things will start to get interesting. It’s hard to do anything in space without knowing your current orbit. Hopefully this will lead smoothly into prograde and retrograde guides among others to aid the player in making adjustments.

There are many other performance, visual, and stability updates that have been made in the past few weeks. I am working hard to put out a playable v0.2 in the near future. This will feel a lot more like a playable game/simulator than previous iterations. I am super excited to share it with everyone. Please stay tuned!

Progress Update

It’s been quite some time since the last update. I am happy to say that I have made several crucial improvements to the engine in this time, and a new version release (v0.1.4 — by far the most stable) is imminent. I will briefly discuss some of the more exciting updates in this post.

Most of my latest code has been dealing with the stability of the collision sub-system. Previous versions have had a very rudimentary terrain collision detection that operates only on the landing gear. Worse still, the existing system used terrain normals which were in the wrong transformation, resulting in unpredictable corrective impulses. This usually means being randomly accelerated into the ground. Not good.

The new collision system currently uses three individual collision checks. Firstly, high velocity collisions must be checked so that objects do not phase through one another. This works by casting a ray along the velocity vector for the duration of the next time step. If this intersects with the terrain, an impulse is computed to prevent penetration.

Next, a Bullet collision mesh is tested against the terrain mesh to search for spacecraft body collisions. If any are detected, the appropriate corrections are computed. Finally, the existing landing gear spring-dampening system is used to allow the player to land and move along the ground. But this time the normals are right and you don’t get pushed into the ground.

The next feature has to do with providing predictable locations on randomly generated planets. There was previously no way to ensure that a certain location on the planet would be guaranteed to have a good place to land. With the new “Bases” feature, locations with predictable terrain can be generated for any planet. In the pictures below, a crater with a bottom at 400 meters is generated to provide a landing zone in an otherwise rough location. These are the only planetary terrain features currently supported, but others are in the works.

Along with this, I have begun working on a “decal” sub-system. This will provide ground textures for landing strips and roads that will be important in future releases of the game. These can be applied to arbitrary terrain, but will frequently be coupled with Base locations.

There have also been many bug fixes along the way that fix all aspects of the simulation. There are still some quirks to the engine, but I hope you will find that the latest version is an order of magnitude more stable than previous releases. But as always, please report any bugs at the BitBucket issues link.

As critical bugs are resolved one by one, progress seems to be picking up again. Although I won’t make any promises, I anticipate launching v0.1.4 within the next few weeks. I think that v0.2.0 will follow soon after that since I believe many of the latest features are worthy of a minor version increment. Stay tuned!

Space Game v0.1.3 Released!

A new version of Space Game is now available for download! This is not a major update, but it includes many performance fixes and gameplay improvements over the previous versions. The main performance update corrects the sphere vertex distribution (as explained in Cube-to-Sphere Projection) for a consistent and uniform frame rate. The bounding box algorithm now uses Oriented Bounding Boxes (OBB) to reduce selection in planet “corners.”

Shadow mapping is off by default in this release. Since this feature is still experimental and has a big performance impact, I think it is better at this point to have the player manually enable it. To enable shadow mapping, open config.ini and append the following lines:

shadow_mapping = 1

Known Limitations:

The new frustum-OBB culling algorithms are not yet perfected. There are cases where terrain will not render though it is visible on-screen. I am currently looking into a fix for this issue.

Cube-to-Sphere Projection

As the release of Space Game v0.1.3 approaches, I thought I would discuss one of the more challenging obstacles that was blocking its release. I first noticed something was wrong when traveling to certain “corners” of the demo planet in the game. Performance would drop, and the simulation would attempt to request and render thousands of terrain nodes. Towards the spawn region — and most of the rest of the planet — I never noticed this performance hit. For a while, this seemed like a minor issue and I just left it on the backburner. Well, it turns out that the issue was a bit more sinister.

To give some background, the planet renderer originally used the naïve cube-to-sphere projection of simply normalizing the quadtree cube coordinates and scaling to the planet radius. This method can be seen in the image below. The red dots indicate vertices placed at equal intervals on the cube’s surface. These vertices are then “pushed” along their normals relative to the center point until they reach the sphere surface (marked with green dots). Notice the green dots are no longer at equal intervals on the sphere surface. This introduces a distortion with more vertex density towards the face edges.

In the terrain renderer, each pair of dots constitutes a renderable tile. This means that as the player approaches the edges of the former cube faces, more terrain tiles would be required and performance would drop. To solve this, I implemented a projection scheme that would warp the position of the cube face coordinates such that when they were projected, they would create (nearly) equal area regions on the sphere surface. Terrain tile density would remain mostly constant and gameplay should stay smooth. The desired projection is shown below.

The cube face coordinates (red) are now warped towards the center of the face such that after the sphere projection they are at equal area intervals (green). As it turns out, this is a projection that is easy to compute an exact solution for in 2D, but difficult in 3D. My goal was to find a projection that minimized computation, but also provided an equal area solution to an acceptable margin of error.

After some research, I found a paper that went over several options for cube-to-sphere projection. The approach outlined in the paper that I found most desirable is the Odd Polynomial Univariate method. It had minimal area error, but was also very fast to compute. The method forms a 5th order polynomial with 3 coefficients that are optimized to minimize distortion.

    \[ K_1 x + K_2 x^3 + K_3 x^5 + \ldots \]

This method further reduces the error beyond the similar tangent function, and it is a simple computation. So with this equation in place, I thought I had a sphere with nearly equal area distribution. I start up the game, head to a planet corner and… performance dropped to a slow crawl! After investigating for a few days, I found that the warping scheme was applied on my UV coordinates which were in the range [0, 1]. This caused the projection to push vertices into the (0, 0) corner of each face, resulting in extreme terrain tile counts in this region. After remapping the UVs to a [-1, 1] range, my performance was as I had predicted.

This was another one of those issues that seems simple in hindsight, but the processing of discovering the problem and tracing back its cause can be exceedingly difficult. The resulting mapping is now fairly uniform and has good performance all over the planet. Now that I’m done wasting time fiddling with renderer basics, I’m hoping I can get Space Game v0.1.3 out soon!

Space Game v0.1.1 Released!

I am excited to announce the first public demo release of the Space Game engine! The goal of this demo is predominantly to start testing the game on a wide variety of hardware. There are many quirks to the GPU-side double-emulation code that have been exposed on new hardware in the past. It is better to eliminate these issues now while the code base is relatively small, than to discover them later when many more lines have been written based on the current assumptions.

There are a few new features relative to the previous update. I had mentioned adding HDR and exposure support. Well, I jumped the gun and implemented a basic system without eye adaptation. The atmosphere and sun are now at more realistic brightness levels relative to the rest of the scene, although this will still require some tweaking in future updates. The output is also now gamma-corrected for completeness.

The Space Game engine still has a long way to go, and has plenty of yet undiscovered bugs. But this first release is a big milestone in its development. I hope you enjoy experimenting with it! Please feel free to share any ideas, and don’t get too upset at the many deficiencies at this point.

Skybox Update

I have been holding off on this one for quite some time. The graphics, especially in space scenes, have always seemed to be devoid of…. soul. There were only two visible bodies: the star and the main planet. Not a star to be seen in the sky!

I have held off on adding a background for a few reasons. Firstly, it’s not totally unrealistic to have the background empty when viewing very bright objects in the foreground. Just like how there were no stars in the background in the moon landing images, there should not be visible stars in these scenes. However, it turns out that the blackness of space is very unwelcoming. The game feels friendlier when you can see some stars.

Now with 100% more dust!

Second, the game will ultimately support a whole universe of stars, galaxies, and nebulaes that the player will be able to visit. Once this feature is in place, the background will be the true sky around the player rather than a skybox. Any star in view would be a potential destination for the player. But, this feature is some ways off. So in the meantime, I have decided to breathe some life into the game with a static skybox.

Atmospheric scattering and color blending.

In the future, I hope to add an “eye exposure” that may help replicate your eye adjusting to the light level. This will help strike a balance with realism without making the universe seem boring. Stay tuned for developments with this feature.

Particle System

I have added many features to the space project over the last few months. But one feature that I have been looking forward to adding for a long time has finally come to fruition. It’s a feature that adds a small touch of life to the game and is useful for countless dynamic graphics effects. By now you probably have guessed it (not to mention the blog title)… Particles!

Set the controls for the heart of the sun!

The space project’s Particle System faces many of the same precision issues that have made conventional approaches more difficult. These particles must be able to be emitted at very high relative speeds, and in many different reference frames and coordinate systems. By making use of double precision on the CPU-side, and emulated double precision on the GPU-side, I am able to accurately render many billboard particles at a wide range of velocities and distances throughout the entire Star System.

The Particle System itself is based on a pool (underlying deque) of IDs that point to Particle objects in a constant-size array. Active particles are moved from the pool to a list where they are updated every frame. An active particle has a life cycle that is directly related to its current “alive time,” which measures how long in seconds the particle has been alive. During this life cycle, particle parameters can be adjusted based on functions or simple interactions to create interesting graphical effects. Once a particle has reached its end-of-life, it is transferred back into the pool where it can flag that it is available for use again.

On the graphical side of the house, the particle position and parameters are stored in a dynamic Vertex Buffer Object (VBO) which is updated after any change is made. Like other graphical objects in the game, positions are assumed to be relative to eye (or camera) to conserve precision close to the viewpoint at the expense of precision far from it. This results in a very stable simulation with minimal apparent jitter. The Particles are sorted by the square of their relative to eye position for correct alpha blending of translucent objects.

This is the most stable and efficient launch profile for this spacecraft.

The Particle System is far from compete and currently only supports smoke trails. However, the bulk of the implementation is out of the way. New Particle types can be rapidly added to support all kinds of graphical effects. But for now, what other than smoke does a space game need to support? Yeah, yeah, fire and… explosions of course! Stay tuned.

Space Simulator Progress

Over the last few weeks I have made significant progress on the core parts of the space simulator engine. Although some changes were made under the hood (more sensible organization and rendering), the most noticeable are the addition of model material support and shadow mapping.

Model materials are a feature that have been a long time coming. Models used to support only one mesh and a clunky and simplistic shading model. With this latest update, models can have an arbitrary number of meshes (performance-limited) and can compile flexible shaders on the fly. Although the current demo shader continues to use the Phong shading model, I am looking into implementing a more complex Physically Based Rendering (PBR) model. Textures are also implemented, but have a few bugs associated with them. Check out a rendered scene with multiple materials below!

A new system was needed to provide high fidelity shadows both on the ground and at altitude. I had started the skeleton of Cascaded Shadow Maps (CSM) nearly a year ago and was finally able to get the feature implemented. Although it still has some bugs that will need to be ironed out, shadows give the rendered scene a greater sense of depth.

These are just a few of the most obvious updates. A rudimentary cockpit system is in place that will allow the player to have a first person view from the spacecraft. I am currently working on improving the rendering model a bit, and then I will move on to adding fidelity to the planet and surrounding solar system. Stay tuned for some exciting updates and a possible official public demo release soon!