Upscaling Technology Has Become A Crutch

Davros said:
This technique : Algorithmic anti-aliasing of specular highlights
implemented in velvet assassin

Velvet Assassin on Mac (Patch) | The Tenth Planet

www.thetenthplanet.de www.thetenthplanet.de
Click to expand...
This game was made in 2009 when there was no such thing as RT on GPUs. I'm talking about RT specular noise. Not rasterization noise that is using screenspace to solve everything.

Davros said:
Theres also : Screen Space Specular Anti Aliasing (watch the video)
www.shadertoy.com

Shadertoy

www.shadertoy.com www.shadertoy.com
Click to expand...
I know Anton K..

This technique makes derivatives of a half-angle vector thus making the specular highlight larger than it's original intent. I may *want* my specular highlight very close to zero especially if it's going to be a contribution for glass or water. It's also a 2D computation and ignores the z-component of the vector assuming things are done in screen-space. Finally this type of AA algorithm can be done using the other AA techniques like DLSS, TAA, etc.. without the overhead math operations.
 
pcchen said:
It's expensive in memory and requires special care for some rendering techniques, and it's not really performing better than other methods. Why use it when the upsides are not much better than the downsides?
Click to expand...
Whilst I can't currently see any basis of implementing the technique with deferred renderers, the idea of being able to explicity decouple sampling of visibility and shading is conceptually much more simple/elegant (no failure cases) than temporal reuse ...
DavidGraham said:
VR games are just like mobile games, performance limited and last gen looking. 95% of games on Mobile and VR have no more than a one or two light sources lighting the scene, which is why they chose Forward/Forward+.
Click to expand...
Forward rendering did see a decline in high-end real-time rendering but light count scalability wasn't the underlying reason you believe this to be. Modern incarnations of forward rendering that we know of can scale well enough with high lighting density ...
VFX_Veteran said:
What techniques? And where has it been implemented?
Click to expand...
There are countless numbers of games alone using toksvig or some combination thereof to filter their normal maps. We know recent CoD games have a custom spec/gloss AA solution and Unity's HDRP implements this method ...
 
Lurkmass said:
Whilst I can't currently see any basis of implementing the technique with deferred renderers, the idea of being able to explicity decouple sampling of visibility and shading is conceptually much more simple/elegant (no failure cases) than temporal reuse ...
Click to expand...

Well, apparently the potential upside is not good enough, and that's why most game engines prefer to use the current prevailing solutions. The fact that very few games today still using MSAA is quite telling.
 
pcchen said:
Well, apparently the potential upside is not good enough, and that's why most game engines prefer to use the current prevailing solutions. The fact that very few games today still using MSAA is quite telling.
Click to expand...
Well that's an implementation issue rather than a design flaw with MSAA. There are other contexts where it does make sense and/or much easier to exploit MSAA to directly sample visibility over methods that don't have any tight relationship to graphics theory nor fix the underlying problem itself ...
 
You must log in or register to reply here.
Back
Top