Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
4K TVs Continue To Rise Despite TV Shipments Dropping
#34
Quote:Voodoo 1 & 2's did use blurry LOD's to hide the effects of their filtering and the low resolution of the textures.

You didn't read the book I linked, or if you did you didn't understand a word of it. Isotropic sampling with low resolution output based on the view frustum demonstrates pretty easily that the Voodoo boards were doing it properly- if you don't believe it just fire one of them up and turn on point filtering, off the top of my head, it's been a while- in Quake engined games that would be gl_linear_nearest.

Quote:Is it engineering or some kind of math?


Statistical analysis.

Quote:For all Bens supposed and flaunted knowledge he didn't have the first fucking clue about SG:SSAA and said it looked bad, yet all he had to do was take his own advice about negative LOD and apply some when using it.


[Image: pic11.jpg]

That is showing you isotropic versus anisotropic filtering and how it relates to mip selection. The VSA100 parts were all isotropic and as you should be able to see, adjustments to LOD bias would have a huge impact on the entire image. Now, in the anisotropic shot things change. If you had actually read the book I linked, or understood the beginner level elements of 3D rendering- you would know that using a negative LOD bias slides the calculation of where mip transitions happen. The big problem with your 'fix'- it doesn't touch the base map. Mip 0 isn't rendered in a different fashion simply because you use a negative LOD bias. The Voodoo boards *all* used isotropic filtering all of the time which means LOD bias adjustment had a huge and immediate impact. Modern cards use high levels of anisotropic filtering see a much, much smaller adjustment and don't benefit *AT ALL* on mip 0 using a negative bias.

Quote:and I'll bet he can't expand the "SG" part of SG:SSAA without googling it either!


You are right, but I'd be willing to wager not for the reason you think. Firing it up stochastic was instantly ruled out because it looked like absolute shite. It was abhorrent. That left two possibilities, scattered grid or sparse grid. Both of them are shitty vaseline style blur fixes, and without playing around with it for a while and using some test cases it would have taken me a while to figure out which shitty version of "SG" they were using. 
Reply


Messages In This Thread
RE: 4K TVs Continue To Rise Despite TV Shipments Dropping - by BenSkywalker - 03-17-2016, 03:40 AM

Forum Jump:


Users browsing this thread: 1 Guest(s)