Mipmapping where levels are 1+2^n in size and texels stack up directly over each other must Shirley have a name! http://pastebin.com/HYfF8dch
-
-
Replying to @tom_forsyth
@tom_forsyth Hopefully@cmuratori and I are talking about the same thing, otherwise there's TWO ways, neither with names :-)1 reply 0 retweets 0 likes -
Replying to @tom_forsyth
@tom_forsyth Figure 5 here are the two things I was talking about: http://pages.jh.edu/~dighamm/research/2004_01_sta.pdf …1 reply 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori Oh - this is a use for the Offset parameter of Sample() - see "Applying Texture Coordinate Offsets": https://msdn.microsoft.com/en-us/library/windows/desktop/bb509695%28v=vs.85%29.aspx …1 reply 0 retweets 0 likes -
Replying to @tom_forsyth
@cmuratori Ah crap - except it's an integer, and we want to feed in 0.5. Bollocks.1 reply 0 retweets 0 likes -
Replying to @tom_forsyth
@tom_forsyth Well you also need to scale it per LOD as well - it's a bias _and_ a scale that are dependent on the texel size.2 replies 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori That's the nice thing about that bias - it's in texels, whatever layer you use. Unfortunately, it's an integer, so useless.1 reply 0 retweets 0 likes -
Replying to @tom_forsyth
@tom_forsyth That'd still be fine for the 1-texel pad case - you always add 1,1. The problem is you still can't do the scale.2 replies 0 retweets 0 likes -
Replying to @cmuratori
@tom_forsyth They just needed to add a texel-relative scale and it would have worked!1 reply 0 retweets 0 likes
@tom_forsyth Of course, that is a much more expensive endeavor, for obvious reasons :(
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.