My post wasn't specifically directed at you, but you did say Killzone 3 was "ugly". I suppose that could be interpreted as unimpressive.
What about Helmuts?
Range of graphics effects in console games *spawn
- Thread starter Cyan
- Start date
My post wasn't specifically directed at you, but you did say Killzone 3 was "ugly". I suppose that could be interpreted as unimpressive.
What about Helmuts?
Note how the bottom lowres cloth texture looks a lot more like the KZ3 screenshot than the top highres cloth texture.
Thanks for the comparison shots
Lucid_Dreamer
Veteran
I see. Isn't that where high frequency comes in? It saves the texture size from being to large and brings additional detail, but it can, supposedly, be computational expensive. Right? BTW, do you consider a 1K texture to be low rez? Don't PCs regularly use 2K textures with 4K being the highest in games?Out of curiosity I threw a texture map onto a generic character. The UVs encompass the whole body, including legs and feet. Every polygon has unique UVs.
http://www.opengl.org/resources/code/samples/sig99/advanced99/notes/node88.html
It's a very different mather talking about either character textures or game world textures.
I don't have the time or patience to search back, but I recall calling the clamped highlights ugly and not the entire game.
It's even less straightforward because you can use multiple textures on a single character or any other object.
Most movie creatures actually use 2K textures - but up to hundreds of them are needed to fully cover the model. But it's easier to work with both during texture painting, and during rendering (only accessing those maps that are actually visible can save a lot of disk access and network traffic).
Games can also use anything from a single texture page (for example 2K), or two of the same size (1K + 1K), to any other combo (2K for the body, 512k for the individual head)
Texture size on its own says nothing. Total texture space used per character does.
This says much more concisely what I was trying to say in reference to the Halo Reach/Killzone 3 character comparisons.
Not really, because texel size can be compared based on screenshots taken from a similar viewpoint. It's still evident that KZ characters' main texture layer has less less resolution compared to Halo, even without knowing what sizes are used. The detail texture does a more or less effective job at hiding that.
Lucid_Dreamer
Veteran
Like Kratos, in GoW3, has "at least 20 textures on him now", right? They said they used 2048 size textures on his head, upper body, and lower body. "Each character gets a normal, diffuse, specular, gloss (power map), ambient occlusion, and skin shader map. We also use layered textures to create more tiling, and use environment maps where needed." GG should let us know about the techniques they used on each main character.
Frankly, I have no idea because I don't work in the games industry. In my line of work I tend to use 4K and up for character bodies, so I was simply curious how 2K vs. 1K compares from close viewing distance.
Well, you can always add them up to get an idea of how much detail there is. Although I doubt most characters are split up beyond head and body (except for external accessories of course)?"
I've meant it in the context of "game X is using 1K maps so it can't have as much detail as game Y which is using 2K textures". This statement on its own is not enough.
If it's like "game X is using a single 1K map per character and game Y is using a single 2K map per character so game Y has more detail" - then it would probably be true.
And yeah, most games don't usually split up characters to more than two maps. It's either one for the head and one for the rest, or two if they find a rectangular layout works better (so they use two squares).
Although for example the original Halo used many separate textures for the human characters to be able to generate variations on their appearance. Different headgear, legs etc. - can't seem to get Google to find me the image right now...
If it's like "game X is using a single 1K map per character and game Y is using a single 2K map per character so game Y has more detail" - then it would probably be true.
And yeah, most games don't usually split up characters to more than two maps. It's either one for the head and one for the rest, or two if they find a rectangular layout works better (so they use two squares).
Although for example the original Halo used many separate textures for the human characters to be able to generate variations on their appearance. Different headgear, legs etc. - can't seem to get Google to find me the image right now...
I have no idea where you are getting at. I have already pointed/agreed that the particular model is made out of low res textures in previous posts. Its underneath the "textile" layer that you showed as the offender which hides the low res textures.
Perhaps this is best illustrated with megatexturing!
The shadow casting dynamic light sources are very limited in Crysis 2.
(((interference)))
Veteran
Yeah, they should learn from Epic, Gears looks great despite not having particularly noteworthy tech.
Bungie's imposter LOD system was quite impressive though, I could never spot any LOD shifts while playing and can't tell at all the difference between an imposter and actual geometry.
Open world games like RDR, AC, Skyrim etc would benefit greatly from a similar system
as you can generally distinguish the lower LOD models in those games (especially if you've zoomed in the camera).
Well, impostors themselves aren't a new tech, though I suppose it's not often we see them used on characters or vehicles. Mass Effect 2 used them on Illium in order to display the crowds off in the distance for example. It's been awhile since I've checked out other games with larger environments and crowds though. Not sure what Dead Rising 2 does.
LOD transitions are certainly something that could use some work in general at some point. I suppose next gen when we've got more robust tessellation routines for all platforms, it'll be taken more seriously.
They're pretty limited in most games for a couple reasons. Performance, limited resolution, artistic control. You certainly don't want a shit framerate or lots of noticeably low res artifacts. Also, just enabling it for tons of lights would make lighting the scene extremely chaotic and artistically defunct. Hard/soft shadowmaps aren't really appropriate a lot of times in well lit areas with lots of ambient lighting, which is the main focus of their SSGI anyway. Considering Crysis 2 is going for fancy realism, it doesn't make too much sense to make it shadow heavy ala Doom 3 for example.
Multiple shadowmaps from independent light sources is a hefty resource drain because you're rendering individual shadowmap buffers (heavy z-fill and memory and bandwidth requirements).
If you've ever fired guns in real life, it's not like you're going to see much in the way of hard shadows unless you're in the dark, again, low ambient lighting. Or just take for instance, a club scene with lots of lights.
Multiple shadowmaps from independent light sources is a hefty resource drain because you're rendering individual shadowmap buffers (heavy z-fill and memory and bandwidth requirements).
If you've ever fired guns in real life, it's not like you're going to see much in the way of hard shadows unless you're in the dark, again, low ambient lighting. Or just take for instance, a club scene with lots of lights.
L. Scofield
Veteran
Not really. All the shadows in the game are dynamic. From those of small rocks to skyscraper's.
