03-17-2016, 06:56 PM
Quote:No, I just know what the fuck I am talking about, whereas you and Ben haven't the first clue!
Break it down then. So far you have given me the impression you have no clue what I am talking about- change that perspective, please. At this point you sound like you know as little as some of the viral marketers AMD uses. "It's pretty" isn't knowing a fucking thing- break it down.
Quote:Don' worry you weren't the first and you won't be the last.
To violently sodomize your..... I can't even say arguments. You haven't argued, you have pulled a straight three year old temper tantrum shoving your fingers in your ears and screaming you win. I've broken it down for you, you haven't said a fucking thing in defense outside of pretending you understand elementary levels of rendering and horribly fucking up your thoughts on those.
Quote:Baumann lost to me with his GeforceFX and DirectX 9 partial precision propaganda and responded by banning me from B3D.
Wavey banned you? He was a really laid back guy- if you were dealing with him the way you are presenting yourself in this thread @B3D I can understand- you are coming across like you are winning a theoretical astro physics debate by screaming you shit your pants- and you seriously seem to think that is valid.
Quote:What you all of you think about this?
Took me about an hour to think of a *potential* use for it that any of us would have, came up with one though. If you were running Morrowind with extreme mods and *MASSIVE* textures placing memory at a premium utilizing this approach could be viable. On a computation side this approach is kind of an interesting take on using an IMR shaders to gain some of the benefits of a TBDR.
I want to get that out of the way because, seriously, I am not trying to bash the guys that wrote this paper. Looking at what they were trying to do they succeeded, not only that but they seemed to iterate several levels out of what we would typically expect to see at a first generation technique such as this.
This method of AA is so obscenely shader intensive it will not be viable outside of some extreme case situations with *very* narrow parameters. Ray tracing is cheaper then what they are doing. To compound the issue, it is simply giving you superior edge AA(versus normal MSAA) while doing fuck all for the rest of the scene. Again, I'm not trying to bash this implementation- what they were trying to do they did- figured out a way to get very high quality edge AA while using a significantly smaller memory footprint then any of our current approaches- they get a mission accomplished and some kudos for how much evolution of the technique they did prior to publishing- it is simply a *very* narrow window where this will be useful at all.