New Unreal 3 Videos

Wish they could be in better quality. 4 was impressive I must say, but in three the water on the wheel just looks.....eh. Unimpressed by that. Water that's displayed like that in general though, I've yet to see an engine that really does something of that nature well. Flowing water or falling water to clarify.

All in all, cool stuff. Sadly though unless Unreal 3 does something innovative and fresh, it will fall to the wayside for me. I haven't even played HL2. I'm just not driven enough to even go buy it and play it.
 
QHalo said:
Wish they could be in better quality. 4 was impressive I must say, but in three the water on the wheel just looks.....eh. Unimpressed by that. Water that's displayed like that in general though, I've yet to see an engine that really does something of that nature well. Flowing water or falling water to clarify.
Click to expand...
I imagine something as free form as water is would be very difficult to simulate.
 
Bane said:
I imagine something as free form as water is would be very difficult to simulate.
Click to expand...
Oh I bet it is, but it seems like designers are making strides in lighting and movement...that just seems to be an area that I've never seen a stride in really. I mean the water in HL1 and the water in HL2 is significantly different yes, but like you said freeform water is probably very difficult to make. I'm not an engine designer or programmer/coder whatever. It was just an observation though, the engine does look pretty cool. :)
 
QHalo said:
Oh I bet it is, but it seems like designers are making strides in lighting and movement...that just seems to be an area that I've never seen a stride in really. I mean the water in HL1 and the water in HL2 is significantly different yes, but like you said freeform water is probably very difficult to make. I'm not an engine designer or programmer/coder whatever. It was just an observation though, the engine does look pretty cool. :)
Click to expand...
I would like to see some of the calculations and code that goes into makeing even the simplest water effect. The day that someone does make water, act like real water in real time will be the day I have a new hero.
 
With the advances in physics in games, I think we will soon see realistic acting water. If im not mistaken water is one of the toughest things to emmulate. When watching behind the scenes of Terminator 2, when the T1000 goes through the metal security gate, they used water as an example of how his metal structure should flow around the bars. The computer power required to produce such an effect was very high. Granted this movie is old, and computers have advanced since then, but it still is a difficult effect, to achieve.

Heck the 3D program I use, Solidworks still cannot emmulate water flow accurately.

Next thing you know the Geforce 7200 will include a LPU Liquid-dynamic processing unit.

:D
 
Oh trust me I'm sure that whatever it takes to render it is out of this world otherwise we would have it. Water effects always get a rise out of me in games. I loved the water in Farcry and HL2, from the pics I've seen looked like it was decent too.
 
The water in Half Life 2 is incredible if you can see it in reflect all mode. Amazing. Doesn't do everything, but just looking at it is awe inspiring to me.
 
I would really like to see water effects improve. The other night I was playing WoW, and I was underwater hunting crabs near the dock in ferralis. A boat passed by overhead, which was kinda cool, but I just thought to myself how much better it would be if there would have been a wake, and water disruption (bubbles) trailing the ship as it went by.

And the water in HL2 with relect all is very nice. Looking down on it anyway. If you are underneath the water looking up things do not distort as they should as the light passes through the waters surface.
 
QHalo said:
All in all, cool stuff. Sadly though unless Unreal 3 does something innovative and fresh, it will fall to the wayside for me. I haven't even played HL2. I'm just not driven enough to even go buy it and play it.
Click to expand...

There is no hope for you then my son.........
 
bonkrowave said:
I would really like to see water effects improve. The other night I was playing WoW, and I was underwater hunting crabs near the dock in ferralis. A boat passed by overhead, which was kinda cool, but I just thought to myself how much better it would be if there would have been a wake, and water disruption (bubbles) trailing the ship as it went by.

And the water in HL2 with relect all is very nice. Looking down on it anyway. If you are underneath the water looking up things do not distort as they should as the light passes through the waters surface.
Click to expand...

There are pleanty of realtime physics demos that sport a simple water model, which have realtime deformation of a triangle soup type surface. I think a completely particle based water system that would splash etc is a bit too complicated for realtime.
 
Lord of Shadows said:
There are pleanty of realtime physics demos that sport a simple water model, which have realtime deformation of a triangle soup type surface. I think a completely particle based water system that would splash etc is a bit too complicated for realtime.
Click to expand...

They just need to get ID on it. ID whored out the particle effects for quake 2. Im sure they have a lot of R&D put into it.
 
IceWind said:
There is no hope for you then my son.........
Click to expand...
:(

I know, I watched my roommate play it some when I was playing WoW, and it looks cool and all, but I just wasn't that impressed. Farcry was the last game I played a decent amount of to be honest. All in all, I'm waiting for Fortress Forever as some of the guys from my clan are helping create it. TFC will live again!
 
what water needs is real time dynamic geometry displacement

unfortunately currently graphics card's aren't powerful enough yet to do it

but one day they will be, either them or a separate physics engine, however the future works out, we'll get there eventually

i really liked the city in demo 4, absolutely incredible, can't wait to play some games using this engine
 
Water is **really** difficult because it's such a dynamic object. It keeps it's volume while it can break appart, split, wrap around an object, ect...

This is how it's handled in a non-realtime environment. The feedback is realtime after the simulations are computed. Calculating millions of particles interactions with other particles and object surfaces is a little difficult ;)

http://www.nextlimit.com/realflow/index.html

Realistic water is the Holy Grail for CG artist/programmers. On the film side, it looks REALLY nice, but it's still a ways off. On the game side, surface deformations are being handled via shaders and surface deformations...

T2 fluids were very hard to do at the time, but pretty easy now with the current generation of 3D software.
 
Metallica_Band said:
Fire>H2O :p
Click to expand...


I urn for a meaty realistic flame thrower that can knock over objects because of the strenth of the projectile flame. I could run around all day in an imagineary world and burn stuff all day
:D
But water is still cool, fire > water.
 
Lamont said:
Water is **really** difficult because it's such a dynamic object. It keeps it's volume while it can break appart, split, wrap around an object, ect...

This is how it's handled in a non-realtime environment. The feedback is realtime after the simulations are computed. Calculating millions of particles interactions with other particles and object surfaces is a little difficult ;)

http://www.nextlimit.com/realflow/index.html

Realistic water is the Holy Grail for CG artist/programmers. On the film side, it looks REALLY nice, but it's still a ways off. On the game side, surface deformations are being handled via shaders and surface deformations...

T2 fluids were very hard to do at the time, but pretty easy now with the current generation of 3D software.
Click to expand...

Thats very nice looking water, if we could only wait for our quantum computers to make good use of it. =)
 
QHalo said:
Wish they could be in better quality. 4 was impressive I must say, but in three the water on the wheel just looks.....eh. Unimpressed by that. Water that's displayed like that in general though, I've yet to see an engine that really does something of that nature well. Flowing water or falling water to clarify.

All in all, cool stuff. Sadly though unless Unreal 3 does something innovative and fresh, it will fall to the wayside for me. I haven't even played HL2. I'm just not driven enough to even go buy it and play it.
Click to expand...

Well with that new Physics Processing Unit (PPU) that i've heard a little bit about, we might get the power to make flowing water well in the next few years.
 
IceyMyst said:
Well with that new Physics Processing Unit (PPU) that i've heard a little bit about, we might get the power to make flowing water well in the next few years.
Click to expand...

Why would you need a PPU when you're going to be able to just have a dedicated CPU thread to it in the near future. Dual core + dual threaded = 4 threads.

All in the CPU with full memory bandwidth direct access to everything else, no special drivers, etc...

a PPU just doesnt make sense when your big bandwidth PCIe x16 is already dedicated to graphics and a CPU will have 2 threads minimum from now on. And probably 4-8 in the next few years.
 
the detail in the city and the lighting is great. the physics are also nice and seem finely tuned, but just not really impressive because its just becoming more common
 
IceyMyst said:
Well with that new Physics Processing Unit (PPU) that i've heard a little bit about, we might get the power to make flowing water well in the next few years.
Click to expand...
I guess I just never realized how much the pinnacle of programming would be water. When you think about the dynamics of water and even fire, you get just a small glimpse on just how hard it is to mimic it.

Any more info on the PPU? Sounds interesting.
 
Looks like good stuff to me. I am looking forward to seeing this engine when it is complete and in use in a few games.
 
chrisf6969 said:
Dual core + dual threaded = 4 threads...a CPU will have 2 threads minimum from now on. And probably 4-8 in the next few years.
Click to expand...

What the heck are you talking about?
 
R!P13y said:
I can not wait, can you preorder it yet?....... :D
Click to expand...

It's an engine, not a game

read the gamespot preview out today about epic's next game coming early next year.
 
chrisf6969 said:
Why would you need a PPU when you're going to be able to just have a dedicated CPU thread to it in the near future. Dual core + dual threaded = 4 threads.
Click to expand...

you're aware that the first gen smithfield 'pentium d' dual core chips are going to have hyperthreading disabled - right?

for the time being, hyperthreading on dual core is going to be an extreme edition only feature.
 
chrisf6969 said:
Why would you need a PPU when you're going to be able to just have a dedicated CPU thread to it in the near future. Dual core + dual threaded = 4 threads.

All in the CPU with full memory bandwidth direct access to everything else, no special drivers, etc...

a PPU just doesnt make sense when your big bandwidth PCIe x16 is already dedicated to graphics and a CPU will have 2 threads minimum from now on. And probably 4-8 in the next few years.
Click to expand...

By that logic, why have a graphics card if we can just use a CPU to render graphics? The reason we have graphics cards, sound cards, etc. is because they are specialized to deal with certain calculations and can therefore do them much more quickly.
 
Simply amazing when we never thought that gaming could get any better.
What kind of minimal specs do you think we are looking at here. I am hoping that my systems will run it well.
 
Lord of Shadows said:
What the heck are you talking about?
Click to expand...

Intel has been having Hyperthreading on their CPU's for a long time. Which is like having dual processors. A PPU is just another fancy processor for unloading some of the work load from the CPU.

But all new processors in the the future roadmaps are planning to have at least dual cores (2 processors) You'll be able to dedicated one of those processors to physics in a game. And a lot more likely to be used by game developers than a specialized add-in card, that requires drivers support, etc.

Both AMD and Intel are going to have dual core CPU's out by the end of year.

Intel has the capability of turning on Hyperthreading on each of the cores. (supposedly will only be on the Extreme Edition at 1st)

So Intel will have the equivalent of 4 processors in their highend systems.

IE: Why buy an add-in PPU card thats going to be on a PCIe 1x slot with limited bandwidth, probably no game support, etc....
 
I would like to see games drop that shitty ass fire *cough*halflife2*cough* and get some real good stuff, but water that abides by the laws of physics would be teh sexayyy
 
I'm not as worried about the water as I am about the morphing buildings. Is it simply some inheirant flaw in 3d processing that makes buildings "flatten out" on the extremities of the screen and then vertically elongate when moved to the center of the screen? Answers, please.

B
 
Interesting For Sure.

I do have some thoughts on UE3 that some of you may find interesting.

I used to think that the Unreal Platform would end up being the most robust platform for game development. I do a little unreal mapping for fun and I have spent a lot of time learining the 3rd party applications that are necessary to take full advantage of the engine. With that being said...UE3 will face some SIGNIFICANT hurdles going forward.

1) The cost of the engine.. The cost of this new engine is $$$$$ to put it lightly. I would REALLY expect that only a select few flagship titles will use it to begin with and sadly, I think that UE3 is almost being coded specifically for next gen consoles. That is not to say that the PC wont have the same if not better feature set and performance..it's just that epic will most likely court console developers to begin with as the margins on PC games are too slim.

2) Scalability. This like UE2.5 when it was released will carry some hefty system requirements. All of those who compalin about 2k4 taking up 6+ gigs will not get any repreive here. For consoles its fine, 9 gigs...and maybe 2 dvds. For computer users who install however....there goes another chunk of the HDD.

3) Content Creation. Modeling assets in Maya or 3ds max at the level of detail shown in the video is a CRAZY time consuming process. These demos look like they reflect a stunning level of VARIABILITY, aka, you dont see the same HL2 barrel all over the place. If they hope to keep that level and density for an entire title..they are in for some serious work. <Op-Ed> I really think that the engine needs to have a variability generator that can randomly change the dimensions and characteristics of certain objects based off of a template of a static mesh. The constraints could be set so that say a tree for instance, will vary in height by X% and width by y%. The textures can be lightened or darkened, or have a radom near transparant alpha applied at rendering to change the look slightly. This will go MILES in generating the realizm we will come to expect in a game.

4) Materials. Most people think that a wall has a texture, and maybe a shader but let me tell you how unreal handles surfaces.

-diffuse map = Basic Texture
-Specularity Map = what is shiny (grayscale image) or matte
- Opacity map = weather or not the texture is see through (grayscale image)
-Modifier = is the texture moving...growing shrinking?
-Mip Maps= several smaller images of the diffuse map---eats up vid mem for better performance.
-Normal Map= illusion of bumpyness on tex map (grayscale--I think)
-Detail Map (for an added level of detail- tilable small grayscale)
-Cube map= does it reflect its surroundings? A cubemap is 6 images in itself.

After all of those have been created, then you have to calculate lights, projectors, self illumination, weather it is solid, can block bullets or players etc.

Most every texture in those demos probably had most or all of these properties at some significant value..and I can tell that it must have been a real pain in the bum to creat them.

The bottom line is that epic really needs to create a drag and drop solution for creating and saving materials to eliminate redundant work. This may have been done in UE3 and I hope it does.

Lighting - Most of the assests in UE2.5 are static meshes, these allow us 100+ fps with detailed environements...who would have thunk that poly count would actually be useful in modern gfx cards. Anyway, SMs cannot cast shadows due to the dynamic lighting limitations of the engine and the calcs involved, this will be interesting to see how the new rendering pipeline handles MULTIPLE movers and static meshes in highly shadowed environements.

Finally, I think that the terrain system needs to be bumped up to a NURBS system. It is still too hard to get fine detail with terrains in UE2.5.


Thnx
 
^^ When did they change the name of a "bump" map to a "normal" map. I know bump maps are now called normal maps ... but why, and when did this happen ?

I was talking to a guy about getting some bumpy looking surfaces and I used the word bump map, and he was like ... what the hell are you talking about. And he promptly told me it was a normal map ?

I know bump maps are done with a grey scale. The value of grey determines how far the surface raises.
 
yea really. bump map actually made sense to me. its nice when these names make sense so an average gamer can get a general idea without having to read a long explanation. but normal map?
 
not sure when - I think it primarily had to do with the way the alpha tex is handled.

127 is the gray value that is normal...aka...no affect on the 'bumpieness' of the tex, where as values of1 00 and 150 will add bumps because they deviate from the NORMAL.

Who knows

M
 
physics looked neato, but things don't look like they still have any kind of solid weight to them. there was a dangling something or other that passed right through a wall, thought that was sort of a downer. otherwise looked neato. lotsa detail.
 
You must log in or register to reply here.
Back
Top