Free surfaces software for linux

Tux Racer is a very fun arcade game for Linux.
The main goal in Tux Racer is to have fun! Race down mountainous terrain as quickly as possible, while collecting herring to increase your score!
Main features:
- Fun, easy to learn
- Exciting 3D courses
- Sophisticated physics result in thrilling rides down the mountains
- Collect herring while avoiding obstacles for high scores
- Race over a variety of surfaces, including fluffy snow and slick ice
- Race in various weather conditions and lighting conditions
- Create your own courses using any paint program such as The GIMP (no 3D modeller required)
- Suitable for all ages.

Saper project is a console version of mines with many features.

It features an AI, different surfaces to play on (Euclids plain, Kleins bottle and Torus), and some basic settings.

Raster3D is a powerful set of tools for generating high quality raster images of proteins or other molecules.
The core program renders spheres, triangles, cylinders, and quadric surfaces with specular highlighting, Phong shading, and shadowing.
It uses an efficient software Z-buffer algorithm that is independent of any graphics hardware.
Main features:
- Raster3D uses a fast Z-buffer algorithm to produce high quality pixel images featuring one shadowing light source, additional non-shadowing light sources, specular highlighting, transparency, and Phong shaded surfaces. Output is in the form of a pixel image with 24 bits of color information per pixel plus one matte channel.
- Raster3D does not depend on graphics hardware. The rendering program currently supports output to files in AVS, JPEG, TIFF, PNG and SGI libimage format. To actually view or manipulate the images produced, you must also have installed an image viewing package (e.g. John Cristys ImageMagick or the SGI libimage utilities). The Raster3D rendering program can be integrated with ImageMagick to expand the flexibility of output formats and interactive use.
- Ancillary programs are provided for the generation of object descriptions based on atomic coordinates stored in PDB format. Although Raster3D is not intended as a general purpose ray-tracing package, nothing in the rendering process is specific to molecular graphics. Some of the algorithms used have been chosen for speed rather than generality, however. They work well for molecular graphics images, but possibly would produce odd results if used for very different types of image.
- Raster3D currently supports rendering six object types: spheres, triangles, planes, smooth-ended cylinders, round-ended cylinders, and quadric surfaces. It also supports the definition of material properties (transparency, bounding planes, color, reflectivity) that apply to groups of objects within the rendered scene.
Enhancements:
- changed the default output format to PNG for all programs

xmountains is a project which renders a fractal mountain range.
xmountains is a classic X11 hack that renders a fractally-generated mountain range that slowly scrolls across your desktop.
The basic idea behind a fractal landscape is to generate a continuous surface which varies in height randomly but with the random variation obeying a particular statistical law.
In the case of a fractal landscape the average difference in height between pairs of points separated by a distance l should go as a power law in l.
If you are only interested in random terrain generation rather than fractal self-similarity then you could use other functions of l that tend to zero as l tends to zero.
If the function does not tend to zero then the result will not be continuous
Enhancements:
- Added a -n flag to set the number of colours.
- This changes the same parameter as -B but you request total number of colours.

XiStrat, also known as Extended Strategy is in particular about turn-based, networked multiplayer, non-cooperative, zero-sum, abstract strategy board games (e.g., Chess, Go, Reversi variants, etc.) on 3D-visualized polyhedra and contains a server, client GUI, autoplayer engine, utilities and documentation.
Besides related recreational modern mathematics (single agent, cellular automata, graph/group/complexity/knot theory, discrete geometry, algebra, combinatorics, mathematical physics) is dealt with.
Enhancements:
- New graphs were added, including a Costa discrete minimal surface, sporadic Mathieu M(24), Kleins quartic L2(7), S(5), and A(6).
- The documentation was changed to MathML and contains links to the underlying mathematics (representations, modules, ring, invariants, groups of Lie type, second cohomology).
- The construction of a Schur cover of holonomy groups, morphing, and 2-body systems were implemented.
- Java generics, J3DBuffer, and procedural 3D texture with Perlin noise are now used.

SDL_image is an image file loading library. It loads images as SDL surfaces, and supports the following formats: BMP, PNM, XPM, LBM, PCX, GIF, JPEG, PNG, TGA.

Images provide the basic visual building blocks for any user interface. Colors and fun shapes are the stuff that we as kids looked at for hours at a time while trying to shoot down big aliens and rescue pixelated princesses. Now its our turn to make the images that others will remember later in life perhaps. Now how do we get this dang images into our SDL programs, and be flexible in the handling of the images so that we dont even have to worry about what various formats they may be in?

This is where SDLimage makes your life easier. This document doesnt help you make artwork, but it will give you the functional knowledge on how to get that art into your game. Now go forth and make your Stick Figure of Justice, someone else might fill in for your lack of artistry, at least you wont have to make much of an effort to include the new and better art into your code.

Volumevis displays pixel-based 3-dimensional datasets. 3-D pixel datasets mostly come from 3-D computer tomography (CT) and magnetic resonance (MR) images. These datasets usually come in the form of a set of 2-D picture files each being a slice of the 3D dataset.
To be displayed by volumevis, the files have to be ordered with identical names except for a number somwhere in the name corresponding to the slice number. Most image formats (jpg, tiff, dicom, etc ..) are supported.
Volumevis is a tool for visualizing 3D dimensional pixel-based datasets. Most computer 3D visualization in the world is with surfaced-based visualization. This is true for most 3D games, CAD tool, etc.
By stating it simply, surface-based rendering works by dividing a predefined surface into small triangles, and then sending those triangles along with their color to graphicscard to be rendered. For example,if you want to render a sphere in 3D, you would create a surface of the sphere, divide it into triangles, send the triangles to the graphics card, which would then turn the triangles into pixels and displaythem.
The unfortunate side-effect of this is, that if you happen to cut the sphere, or go inside it, you will see that its in fact hollow. Also, you cannot see any of the details that might be inside of the sphere.
Volume visualization is the equivalent in 3D of displaying a photo in 2D. The surface 3D visualizations on the other hand, are the equivalent of reducing a photo to a representation of lines (such as the outlines of people and houses) and drawing the lines on screen. Obviously a lot of information goes lost in the line representation.
The visualization works like this. In openGL (a graphics programming API), all fragments (3D equivalent of pixels) have an alpha value. The alpha value corresponds to the opacity of the fragment. If the fragment has a small alpha value, it will be easy to see through it and see the fragments behind it.
If the alpha value is high, the fragments behind it will be obscured. A 3D dataset is simply a block of data in the shape of a cube. Somewhere in the image there is an object which is white or grey. All the other places in the image are black. To render the object, all the pixels which are black are assigned and alpha value of 0 (meaning they are transparent). Therefore we dont see the black pixels.
The fragments belonging to the object are white and have an alpha value higher than zero. Therefore its not possible to see through them, and we perceive the location of the object fragment to be the surface of the object. Depending on the alpha value of the fragment, we might also be able to see some of the fragments behind it a bit (see inside the object).
At the moment, Volumevis only supports black/white images (which is fine for most CT and MRI datasets).
Enhancements:
- A bug with slice loading of textures which are not a power of two was fixed.
- Movie recording, cropping, and file saving for slices were fixed.
- Various feature improvements were made.

Aqsis is a Renderman compliant 3D rendering toolkit. The project is based on the Reyes rendering approach.
Main features:
- programmable shading
- true displacements
- NURBS
- CSG
- Motion Blur
- direct rendering of subdivision surfaces