New Unreal 3.0 Engine Screenshots.

The Unreal3 Engine is gonna be so awsome, too bad it requires so much to run the thing.
 
I think more better technology will be available when the first Unreal Engine 3.0 game gets released.
 
Heh to anyone who actually pays attention this shouldn't really be a surprise.

http://www.idvinc.com/html/speedtree_for_unreal.htm (at the top are screenies for Unreal Engine 2... bottom are UE3)

They had Speedtree with Unreal Engine 2 (with Unreal Tournament 2004) and of course UE3's rendering engine makes it look great :D But I would be surprised if they wouldn't have put in Speedtree...
 
Err, I wouldn't say new, I'd say about a week old, but yeah, look awesome they do. :thumbs:
 
I think this was posted a few days ago. Its still nice to look at though. My video card is feeling up to the challenge. (X800 XT PE) However, I'll probably have a new video card by the time any games come out using the engine. ;) E3 should give us more of a glimpse of whats to come. :thumbs:
 
Iced_Eagle said:
Heh to anyone who actually pays attention this shouldn't really be a surprise.

http://www.idvinc.com/html/speedtree_for_unreal.htm (at the top are screenies for Unreal Engine 2... bottom are UE3)

They had Speedtree with Unreal Engine 2 (with Unreal Tournament 2004) and of course UE3's rendering engine makes it look great :D But I would be surprised if they wouldn't have put in Speedtree...
Of course, UE2 didn't have the crazy normal mapping stuff. I think Epic Games (they're the developers, right?) at this point should rename the franchise "Real" since that seems to be the direction they're going.

All hail Real Tournament 2006!
 
StardogChampion said:
I can't wait until someone makes a Morrowind type mod with this engine.

Err... TES4? (which also uses SpeedTree)
 
MuToiD_MaN said:
Of course, UE2 didn't have the crazy normal mapping stuff. I think Epic Games (they're the developers, right?) at this point should rename the franchise "Real" since that seems to be the direction they're going.

All hail Real Tournament 2006!

NEg. Unreal is one of the coolest game/engine names out there.
 
Meh, i think graphics are about as good as they are going to get, what they need to work on is realistic object interaction. (I.E, you shoot metal, and depending on what it is, you punch a hole in it, or it dents. That is what i want now, rather than better gfx.
 
Kebean PFC said:
Meh, i think graphics are about as good as they are going to get, what they need to work on is realistic object interaction. (I.E, you shoot metal, and depending on what it is, you punch a hole in it, or it dents. That is what i want now, rather than better gfx.

Can't we strive to get it all? :| Why should interactivity limit the improvements in image quality? Both can be achieved without sacrificing either one. However, we lack the powerful hardware to do so.
 
Kebean PFC said:
Meh, i think graphics are about as good as they are going to get, what they need to work on is realistic object interaction. (I.E, you shoot metal, and depending on what it is, you punch a hole in it, or it dents. That is what i want now, rather than better gfx.
If you think it's as good as it's going to get you're not very interested in 3d rendering technology, are you?

Spot the real shot in a group of 3D renders. Here's a group of 5 sample images I have gathered:


It's the first one. I got the rest from either nikclark.com or art.co.uk
The most obvious ways the graphics can get better are always higher-res textures and higher polycounts (neither of which are at their peak)... but the game engines of the future will also have to accurately simulate all of the properties of light in real-time to make a truly realistic image. Then, as you seem to be getting at, if things look so realistic they will have to act realistically or the illusion will be shattered for the player. That would require an equally advanced physics engine to go with the advanced renderer including real-time deformation and fluid motion (not just for water).

EDIT: It will probably be at least decade before we start seeing all of those things (incredibly high-poly models and high-res textures, global illumination simulated with photon mapping, sub-surface scattering, fluid motion, dynamic deformation of rigid bodies, wind resistance, etc) in real-time. Who knows? Maybe PCs will have a chip/card dedicated to advanced math to aid in lighting and physics that will speed up the advancement...
 
ive seen a video of the fourth image, speeded up day cycle as the light changes through the room, it uses the latest HDR lighting effects, and I want the 3DS plugin for it dammit! :D.

seriously I honestly believe in 10 years time ingame visual's will close as dammit emmulate real life visuals, characters will have an uncanny realism about them, and we will be glued to the seat's of our computers for even more hours at a time playing around in our nearly seemless virtual worlds.
 
OCybrManO said:
EDIT: It will probably be at least decade before we start seeing all of those things (incredibly high-poly models and high-res textures, global illumination simulated with photon mapping, sub-surface scattering, fluid motion, dynamic deformation of rigid bodies, wind resistance, etc) in real-time. Who knows? Maybe PCs will have a chip/card dedicated to advanced math to aid in lighting and physics that will speed up the advancement...

Meqon physics have shown a demo of realtime rigid body destruction where tiles could be shot off the wall and into pieces, not something out of prefab models as still happens today. Not public yet though, just read about it.

SSS is already possible, the X800 XT demo shows it off, so that's with relatively old dx9b. Reality engine also shows it with plants.

And I've already seen a short movie where they used photon mapping to create caustics on a ring with a GF3. All small scale, but we will come a long way the next few years.
 
Yes, in the next few years some games will start to use the destructible physics as seen in meqon 2.0 but it's not as advanced as the kind I'm thinking about and still with nothing anywhere near real fluid dynamics and advanced particle physics as used by professionals to make water effects in commercials/movies. They will most likely have something like a more advanced version of the meqon demo of ripples on the water's surface... but that's nothing special. I think I got something like that with a demo of my old GeForce 4. Oh, I forgot about hair simulation, too. That's a really tough thing to get right. Look at 3D Mark 03... the hair looked like crap in the troll cave demo.

Yes, faked (still dynamic but not totally accurate) SSS is possible in real-time on DX9 hardware.

Yes, if you limit it to just simplified caustics, photon mapping can be done in the general area of real-time on extremely simple scenes, usually with only a single light, and with a relatively low photon count. In a few years it would still probably take several minutes, at least, to render a complex scene (with lots of objects, several lights, and a polycount in the millions) worthy of a next-gen game with full GI and caustics.

What I'm saying is that making one aspect of a game true to life would take so much of a sacrifice in other aspects of the game that we won't see all of the advanced features in one engine... being used at the same time... at a realistic (as in "like real life" not just "better than current games") level of detail for quite a while. It's going to be baby steps in each field until it reaches a point where it's impossible to distinguish screenshots from photographs. Either I'm underestimating the development of PC technology or you're underestimating the complexity of real life.
 
Imagine WoW with those graphics...then imagine the system needed to run it.
 
OCybrManO said:
Yes, in the next few years some games will start to use the destructible physics as seen in meqon 2.0 but it's not as advanced as the kind I'm thinking about and still with nothing anywhere near real fluid dynamics and advanced particle physics as used by professionals to make water effects in commercials/movies. They will most likely have something like a more advanced version of the meqon demo of ripples on the water's surface... but that's nothing special. I think I got something like that with a demo of my old GeForce 4. Oh, I forgot about hair simulation, too. That's a really tough thing to get right. Look at 3D Mark 03... the hair looked like crap in the troll cave demo.

Yes, faked (still dynamic but not totally accurate) SSS is possible in real-time on DX9 hardware.

Yes, if you limit it to just simplified caustics, photon mapping can be done in the general area of real-time on extremely simple scenes, usually with only a single light, and with a relatively low photon count. In a few years it would still probably take several minutes, at least, to render a complex scene (with lots of objects, several lights, and a polycount in the millions) worthy of a next-gen game with full GI and caustics.

What I'm saying is that making one aspect of a game true to life would take so much of a sacrifice in other aspects of the game that we won't see all of the advanced features in one engine... being used at the same time... at a realistic (as in "like real life" not just "better than current games") level of detail for quite a while. It's going to be baby steps in each field until it reaches a point where it's impossible to distinguish screenshots from photographs. Either I'm underestimating the development of PC technology or you're underestimating the complexity of real life.

I say in about 10 or so years we will be very close.

Look how far we have gone since doom.
Doom
Half-Life 1
Half-Life 2, Doom 3
about 1 year later from Doom 3/hl2 we have unreal 3 that looks 100x better.
 
Back
Top