|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ |
LOSTCIRCUITS |
|
| MSI StarForce 822 (GeForce3 64 MB DDR) Per Aspera Ad Astra | |
| (Review by MS, June 21, 2001) |
Complex Character Animation
The nFiniteFX engine allows for up to 32 control matrices. Each matrix is equivalent to one bone or muscle in a given object or, by extension, allows the game programmer to extrapolate the data to animate several hundreds of bones per character.
KeyFrame Interpolation
I have discussed this feature already in my review of the ATi Radeon but here is the short of it. If there is a sequence of movements of a given object or character, it is possible to predict certain intermediate states of the character without having to import the data for each intermediate step from the CPU and memory. Basically, two frames are taken and the probabilities of the intermediate output are calculated by the vertex shader. Depending on the power of the GPU, it is possible to insert more or less intermediate steps without altering the temporal characteristics of the transformation. The simplest example would be to have one specific point moving between two different locations. One possibility is to load each intermediate step in all its glory and just output the data onto the screen. Another possibility, however, is to predict the path of the point and only use the beginning and end coordinates while the GPU internally fills in the rest without need to access the CPU or memory. Again, the key feature is that the sequence of events is not running any faster with more powerful hardware but that the granularity is increased by adding more intermediate steps.
Morphing
Morphing is similar to key frame interpolation. The difference is that, with a higher power graphics engine, key frame interpolation will resolve in higher granularity without changing the overall speed of the sequence, whereas morphing speeds up the sequence. In other words, two shapes of an object, e.g. a swimming dolphin with its tail either up or down are the reference points and, morphing fills in the intermediate steps to generate a fluid motion. Depending on the power of the GPU, the dolphin will do more or less tail strokes while the granularity is not altered.
Motion Blur and Lens Effects
These two effects that need no further explanation are also standard within the nFiniteFX engine and can be programmed for hardware execution by the game programmer.
Pixel Shader
If a vertex shader is a black box that allows transformation of the shape of objects, the pixel shader is the complementary function that allows modification of individual pixels. A further development of the nVidia Pixel Rasterizer (NSR), the pixel shader allows an infinite number of modifications of each pixel according to the DX8 API. One difference between a vertex shader and the pixel shader is that Vertex Shaders can be software emulated, whereas this is not possible with Pixel Shaders that require hardware implementation.
The nFiniteFX texture pipeline supports up to 8 texture-blend operations, dependent texture reads and programmable pixel shaders

Animated, bumpy, reflective surfaces are some of the most sophisticated techniques possible
HRAA and Quincunx
HRAA or High Resolution Anti Aliasing is another term for what has been known for some time as FSAA (Full Scene Anti Aliasing). Because 2 x AA is not really too thrilling and 4 x FSAA is using up too much bandwidth, nVidia has come up with a little trick called Quincunx using multisampling. An excellent article on this subject has been posted on Fullon3D and I couldn't possibly add to that one.
Next Page: => On to the StarForce =>
If you enjoyed reading this article and found it useful, please consider making a small donation to LostCircuits.