In order to address concerns raised by revelations about its adaptive trilinear filtering algorithm, ATI yesterday dispatched a pair of reps to talk about the issue in an online chat. Bjorn3D has the chat log. The chat was conducted by ATI, but the questioners aren't just blowing fairy dust at them:
Q: Can you give a more detailed explanation as to why the use of coloured mipmaps shows the use of full trilinear, which doesnt correspond to what seems to occur in a normal, real-world situation?I should address this issue just briefly, because some of you guys have called for us to retest everything in our Radeon X800 review using new IQ settings in light of the revelations about ATI's new algorithm. I don't think that's entirely appropriate given the fact that NVIDIA's drivers weren't doing full trilinear on the GeForce 6800 cards, either, in our review. Also, frankly, we don't have enough time or information yet to do so.
A: Coloured mipmaps naturally show full trilinear as our image quality analysis reveals that there could be visible differences in the image. It should be noted that trilinear was originally invented to smooth transitions between mip-levels, and in the original definition mip-levels should be filtered versions of each other, as coloured mip-levels clearly are not. Despite this, we understand that people often make use of hardware for purposes beyond that originally envisioned, so we try to make sure that everything always functions exactly as expected.
However, we can't ignore the new information, and we'll have to take it into account more fully in the future. Truth is, both ATI and NVIDIA have new algorithms for both trilinear and anisotropic filtering in these new cards, and we are still trying to get a read on how things have changed. NVIDIA's new isotropic filtering algorithm appears to be superior to most all of the competing ones in terms of coverage at multiple angles. ATI's adaptive trilinear appears to be damn near impossible to catch with the naked eye. NVIDIA's brilinear is less so, but not by much. You can see comparative IQ, sans colored mip maps, right here and judge for yourself. You can also see how NVIDIA's various IQ settings affect performance on this page if you want to see the sorts of speed gains that are at stake generally here; we tested three different NVIDIA filtering settings. We'll keep an eye on these things and likely retest performance and IQ as the GeForce 6800 approaches stores shelves or shortly after it arrives.
|Some 840 EVOs still vulnerable to read speed slowdowns||28|
|Details leak out on AMD's first Zen-based desktop CPUs||30|
|Nvidia: the GeForce GTX 970 works exactly as intended||62|
|Report: 4GB of RAM coming to GTX 960 in March||87|
|Early deal of the week: A 27" G-Sync monitor for $480||31|
|Gearbox's Homeworld remake due February 25||39|
|Nvidia admits, explains GeForce GTX 970 memory allocation issue||239|
|Here's my guest appearance on tonight's Alt+Tab Show||12|
|HA. AMD in the red and nVidia in the green. Thats funny cause you know... *cough* oh forget it.||+82|