Bilinear / Trilinear question

entropyentropy Yah-Der-Hey (Wisconsin)
edited September 2004 in Hardware
Well, everything I've seen on the internet goes into insane detail, using all sorts of thing I've never heard of. So, could someone explain it in layman's terms? Like, what should I use for the best graphics/performance ratio?

What I think it is right now, just from guessing, is that bilinear passes things over twice (smoothing them out) and trilinear does it three times? Just an idea, which is probably wrong.

Comments

  • RWBRWB Icrontian
    edited September 2004
    I was explained this a few times from my lecturer in RT3D a couple months back cause I couldn't seem to retain what the deal was with this. But I think it has something to do with mipmapping, making textures appear more crisp meanwhile also trying to keep the texture from getting a miore effect or something along those lines.

    Trilinear is the best of course... but I don't think it's that much of a difference from Bilinear, I don't know what kind of performance hit it takes though. I wouldnt think it is that much however.
  • SimGuySimGuy Ottawa, Canada
    edited September 2004
    Point Filtering:
    In layman's terms, point filtering is used to determine the color information of a pixel that is too stretched to look any good. (IE if you've got a texture that is 128x128 pixels wide and you run a 1024x768 screen resolution, that texture could be stretched to 512x512 pixels and be severely reduced in quality).

    Point filtering takes a single texel from the original texture and copies it into the empty space around the original texture, filling the gap. However, it leaves heavy image distortion.

    Bilinear Filtering:
    Bilinear filtering is identical to Point filtering, however the primary difference is that it uses 4 adjacent texels from the original texture to fill the gap instead of the 1 used in Point filtering. All 3D cards today utilize Bilinear filtering and it is considered the bare minimum for acceptable image quality.

    Trilinear Filtering:
    Trilinear filtering is the next step up from Bilinear filtering. It takes 8 adjacent texels from the original texture and fills the gap instead of the 4 used in Bilinear filtering. However, Trilinear filtering also blends the mip map levels (the different texture quality bands you see) and smoothes them together as well by reading 4 texel samples from each side of the mip map border and blending according to that information.

    All together, Trilinear filtering results in 8 texel samples being read from a minimum of 2 different mip map levels to give the smoothest image quality.

    What gives the highest image quality? Trilinear.
    What gives the best image quality / performance? Bilinear.
    Should you use Point filtering? No.
Sign In or Register to comment.