Graphical effects rarely seen this gen that you expect/hope become standard next-gen

AlStrong said:
I'd be curious to see how it handles Doom 3, now that I think of it.. :) (or Tenebrae!!)
Click to expand...
Or any game with horribly shimmering water. ;)

The demo has hardly pathetic case for specular aliasing, but shader AA on speculars would be a most welcome addition for most games.
 
Yes 3 samples is enough for me even at 1440x900.

There is very little IQ difference between 1 and 11 samples when viewing at a steep angle, even in motion.





But the perf hit for even 3 samples is 22% on my GTX260. (1266 vs 982 fps)

So perhaps the technique could be made more efficient if it were angle dependent, e.g. forgo the supersampling when the angle is >60 degrees.

Also this stuff was done in SM3.0 four years ago.
 
What I'd really like is a game like Black (though I never played it) that concentrates on what your bullet can really do. Shoot in a window and it'll break. Have a couple of enemies standing below a lamp? Shoot the lamp which fall down and damage them. Or have more interaction with your surrounding. Shoot something that can fall onto another object that in turn can damage anything on the playing field. I don't mind if there are bounderies to what can break, but it would be fun to have any object within that playing field fully destructable that in turn can also interact with other objects around them....
 
corduroygt said:
Can shader aliasing be alleviated via an image processing algorithm, for example, as an enhancement to the MLAA algorithm?
Click to expand...

On a related note, do these "MLAA" techniques rely on the depth buffer at all or are they completely luminance/contrast based. If they need the depth buffer then obviously they'd be useless for specular aliasing.

Also, specular aliasing is most obvious on near horizontal or vertical objects. MLAA seems to fail even harder than multisampling in those cases.
 
homerdog said:
On a related note, do these "MLAA" techniques rely on the depth buffer at all or are they completely luminance/contrast based. If they need the depth buffer then obviously they'd be useless for specular aliasing.

Also, specular aliasing is most obvious on near horizontal or vertical objects. MLAA seems to fail even harder than multisampling in those cases.
Click to expand...
Usually MLAA edges are detected from color/contrast, but it is possible to get edges from g-buffer if you want.
It cannot do much for specular aliasing in any case, just blur the mess already in the image.

Only way to deal with shader aliasing are supersampling or designing a 'proper' shader which has LoD/AA in itself.
Problem is that having something like surface with a complex lighting model and making it alias free within memory and processing budget while maintaining constant appearance, is not quite as easy as it sounds. ;)

It's like the classic texture aliasing problem with the exception that the traditional mip-maps get the result wrong and aliasing is usually in HDR.
 
Last edited by a moderator:
its kinda a minor quibble, but i just cant stand other peoples flashlights in games.

like in l4d or crysis.

your flashlight looks great, but everyone elses looks like a cone of fog.
 
Are the flashlights ok in Alan Wake?? It seems like all the lights in Awake are rendered using the same method...and that's on the measly 360....
 
snarfbot said:
its kinda a minor quibble, but i just cant stand other peoples flashlights in games.

like in l4d or crysis.

your flashlight looks great, but everyone elses looks like a cone of fog.
Click to expand...

The STALKER games don't suffer from this (at least at max settings):D
 
snarfbot said:
its kinda a minor quibble, but i just cant stand other peoples flashlights in games.

like in l4d or crysis.

your flashlight looks great, but everyone elses looks like a cone of fog.
Click to expand...

I found Crysis' flashlights to be pretty good also, when using high detail levels. Not sure what the 'lower' detail levels even look like though, to be honest...
 
Albuquerque said:
I found Crysis' flashlights to be pretty good also, when using high detail levels. Not sure what the 'lower' detail levels even look like though, to be honest...
Click to expand...

On low it's a bit of a joke, but I'm not sure if it's just a glitch with the demo or not:

http://img94.imageshack.us/i/crysis2010050417180438.png/

Doesn't seem to be much of a difference between the higher ("volumetric effects") settings:

http://img263.imageshack.us/i/crysis2010050417193880.png/
http://img218.imageshack.us/i/crysis2010050417192392.png/
http://img256.imageshack.us/i/crysis2010050417190812.png/
 
I disagree. I recently bought R2 and through and through it's been a disappointment, and after Uncharted's awesome watr shading, what R2 is doing with surface displacement, while kinda funky, is far from aesthetically pleasing. Also regards water displacement, PS2 was capable of some impressive feats. Good examples of propagating waves, interference and displacement are present a couple of minutes in...


TBH I see nothing to R2 that shows an advantage Cell offers.
 
AlStrong said:
Not so sure it'd be affordable in a real game; depends I guess.
Click to expand...

Heavy rain had suport for supersampling in the shader at the material level (obviously reserved for special cases).It's good for specular aliasing or cubemap lookups on high frequency NMaps.
 
You must log in or register to reply here.
Back
Top