I hope the glossary below can help you understand more about video related terms.
32 bits are used to store the color of each pixel on the screen. With 16-bit rendering, only about 65 thousand different colors can be used, but with 32-bit rendering, over 4 billion colors can be used, which results in more color shades and vibrant, crisp images.
256-bit DualBus Architecture
The Matrox G400 chip uses two independent 128-bit buses operating in parallel to deliver twice the throughput rate of traditional 128-bit architectures. The Matrox G400 also employs a full 128-bit interface to memory to create a well-balanced performance design.
AGP memory refers to memory that is on the computer (system memory), not the graphics card. The graphics card, however, is able to directly access this memory as if it were its own, through the AGP bus. Data transfer through the AGP bus is also faster than transferring data through the PCI bus, since the AGP bus is dedicated solely to the graphics card, whereas the PCI bus is shared among other peripherals.
AGP 4X Multi-Threaded Bus Mastering
The Matrox G400 is the first AGP 4X graphics chip designed from the ground up to optimize AGP 4Xs 1GB/sec bandwidth. It is also capable of using Direct Memory Access (DMA) to fetch commands and data from multiple locations in memory. This feature balances the needs of next generation 3D applications.
A texture that is in AGP memory, (in system memory) can be used as the source texture for a polygon. Therefore, only the pixels that are actually read need to be transferred to video memory, rather than unnecessarily loading the whole image. With non-AGP textures, the entire texture must be loaded into video memory before it can be used as a texture. If there is not enough available video memory, then some other texture in video memory must be released. This results in texture thrashing.
Images can be rendered into surfaces that are in AGP memory. The target surface does not have to be in video memory.
Used to render polygons that are translucent. Useful for rendering effects such as smoke, explosions, water or even glass.
Even though trilinear filtering goes a long way towards improving the image quality of a texture mapped 3D surface, there are cases when trilinear filtering fails to provide good results. Both bilinear and trilinear filtering are used to solve problems that arise due to the difference in size of the source texture and the size of the surface that the texture needs to be mapped on. However, they do not take into account the difference in shape and perspective between the source texture and the target surface. Bilinear and trilinear filtering work best for surfaces that face the viewer squarely (depth across the surface does not change in this case). However, when the surface is oblique to the viewer, filtering causes loss of detail. This degradation in image quality is most obvious when the texture being mapped contains text. Text typically has areas which are as low as one pixel wide. Applying bilinear or trilinear filtering on text that has been mapped obliquely will result in the characters becoming blurry and blending with adjacent characters and therefore almost unreadable depending on the perspective. Anisotropic filtering is used to get around this problem by taking extra samples from the texture map.
These extra pixels are chosen according to the direction of the perspective (see figure 16) that the texture is being mapped on. For example, a brick texture being mapped on a wall of a hallway going into the screen will require the texels to be chosen in the direction of the surface of the wall (from front end to back end)
Technique of blurring pixels. This is commonly used to hide the staircase effect on lines or edges when scenes are rendered at lower resolutions.
When a small texture is used as a texture map on a large polygon, a stretching will occur and large blocky pixels will appear. Bi-linear filtering smoothens out this blockiness by applying a blur.
Direct3D is Microsoft's standardized 3D programming interface. When game and application developers create their 3D content they ensure compatibility with the Direct3D standard. All of Matrox's 3D accelerators support Direct3D.
DirectX is a set of APIs (Application Programmer Interface) designed by Microsoft to make Windows-based computers an ideal platform for running and displaying applications rich in multimedia elements such as full-color graphics, video, 3D animation, and surround sound. Built directly into the Windows family of operating systems, DirectX is an integral part of Windows 98 and Windows NT 5.0, as well as Internet Explorer 4.0. DirectX components may also be automatically installed on your system by advanced multimedia games and applications written for Windows 95.
Used to hide the banding of colors when rendering with a low number of colors (for example 16-bits). Banding is what happens when there are not enough shades of colors, resulting in the eye being able to see a distinct change of colors between two shades.
Double buffering is a technique of dividing the frame buffer into two areas, one "draw" buffer and one display buffer. This allows the application to display a frame of animation from one buffer while drawing the next frame into the other buffer, therefore increasing performance by doing two things at once.
This unique feature allows two physically different displays to be supported by a single Matrox G400 chip. It enables easy connectivity to an additional monitor, digital flat panel or TV. The DualHead Display also supports full-resolution, full-frame DVD decoding with hardware sub-picture support.
Environment Mapped Bump Mapping
The Matrox G400 uses environment mapped bump mapping, which is a per pixel combination of three separate texture maps, including a special 'bump map'. Bump mapping provides a distinctly higher level of visual realism, which is why many popular 3D games will be supporting it.
Used to add the effect of the scenery fading off into fog, or hide the far clipping plane. By hiding the far clipping plane, users won't see polygons suddenly popping into view in the distance.
The frame buffer resides in the 3D card's onboard video memory. This is the portion of video memory where the colors are stored and displayed on your monitor.
A shading algorithm that takes the 3 colors defined at each point of a triangle and smoothly interpolates them throughout the surface of the triangle. It is often used to make round objects look more round and smooth.
Multiple textures of increasing size are used to represent one texture. When a polygon needs to use the texture as a texture map, the texture whose size most closely matches the size of the polygon that is to be rendered is used. This reduces rendering artifacts such as "sparkling" and "moiré" patterns as well as blockiness.
OpenGL is an open graphics API originally developed by Silicon Graphics Inc. Historically, OpenGL was used with professional graphics accelerators for high-end design applications. Now, developers such as id Software use it to accelerate popular games like Quake II. OpenGL is supported by all Matrox accelerators
A form of texture compression. Used to reduce the size of a texture when the texture does not have many unique colors. 4-bit palletized textures can have 16 (2^4) different colors. 8-bit palletized textures can have 256 (2^8) different colors.
A technique that is used to give the proper illusion of depth when pictures are texture mapped onto polygons. Without perspective correction, warping will occur and the resulting image will not look realistic.
SDRAM is a complete remake of DRAM*. "S" stands for Synchronous, as SDRAMs have fully synchronous, and therefore very fast, interface circuitry. On every clock phase, the chip accepts and executes some command, transferred over command lines.
Synchronous Graphics RAM is a variant of SDRAM, tuned for graphics applications. Thus SG is faster in graphics applications, although it is not physically faster than SD when used "normally". The added capabilities of SG are used by graphics accelerators.
Used to add extra color to polygons. This is typically used to add a color that gives objects a shiny, wet, or metallic look.
This feature enhances scene quality by creating compelling special effects while saving the graphics engine unnecessary rendering. The Matrox G400 supports up to 8-bit stencil buffering, which can be used for many interesting and compelling effects such as dissolves and transitions, as well as for rendering plain and volumetric shadows, silhouettes and more.
The process of taking a picture, and drawing it onto a polygon. This can be used to add emblems, textures, or pictures to the rendered polygons.
Constant swapping of video memory content to replace old textures with new ones requested by the application, causing a performance slowdown. Having more video memory and AGP helps eliminate this problem.
Texture Transparency (color-keying)
If a texture has a color-key color, then that color will not be rendered. As an example, this could be used for rendering text or objects with holes in them, such as a car door with an open window.
Triangle setup engine
Special circuitry in the graphics chip that computes the mathematical aspects of what is needed to draw 3D images on the screen. This frees the CPU to perform other computations.
Triple buffering is a technique of dividing the frame buffer into three areas: two drawing buffers and one display buffer. This allows the application to display a frame of animation from one buffer while drawing the next frame into the other buffer and then start rendering in the third buffer while the other two buffering are still in use, therefore further increasing performance.
A combination of bi-linear filtering and mip-mapping which enhances even more the quality of texture mapped polygons. For each polygon that is rendered, the two "MIP maps" that most closely match the polygon size will be used to compute pixel colors that are the most realistic. This technique is superior to both bi-linear and mip-mapping.
Vibrant Color Quality
Uses 32-bit color accuracy throughout the rendering pipeline. All internal calculations are executed with 32-bit accuracy and the end result is dithered down to the original color depth from a true color palette.
Vibrant Color Quality²
The new VCQ² architecture ensures vibrant color rendering for multi-textured 3D applications by adding extra precise alpha-blending units to a 32-bit rendering pipeline capable of reading, writing and combining 32-bit textures.
The period of time when the screen is paused for a few microseconds and is getting ready to draw the next frame.
The z-buffer is a reserved portion of video memory that holds depth information (as opposed to color information) for every pixel being displayed on the screen. When a polygon is rendered with z buffering, each of its pixels depth (z-value) is compared with the corresponding value stored in the z-buffer. If the value stored in the z-buffer is less than the depth of the new arriving pixel, it is decided that this pixel is visible and should therefore be rendered. The z-buffer is then updated with the pixels depth. If however, the value is greater, the pixel is rejected and will not be rendered, as this means it is behind of what has already been drawn. Z-buffering is used to ensure that objects are rendered in the right order, that is, objects in the back should not appear in front of objects that are in the front.
*Dynamic RAM (DRAM) is based on capacitive elements. We may think of DRAM as of an array of capacitors, controlled by switching transistors. Only one "capacitive transistor" is needed to store single bit, so DRAMs have higher capacities than SRAMs.