UT2003 Software Renderer

The Unreal Technology Website (thanks Frans) now offers a zip file with the integration of RAD's Pixomatic software renderer into UT2003. They also suggest lowering screen resolution for software rendering should you have a slower CPU (which is probably most likely, since this is also targeted at users without adequate hardware 3D acceleration). This thread on the UT2003 is where to post questions or bug reports about this.
View : : :
35 Replies. 2 pages. Viewing page 1.
Newer [  1  2  ] Older
35.
 
Re: No subject
May 22, 2003, 08:58
35.
Re: No subject May 22, 2003, 08:58
May 22, 2003, 08:58
 
Yeah, but if you want to process sound to that degree you're looking at what Carmack is currently going through with light for Doom 3.

Games in the past have been lit at the point at which the map is built, increasingly now they are lit as they are rendered and Doom 3 will be done entirely like that. You're basically looking at real time ray tracing which is hardcore and hard work. The processing needed is immense.

In the past sound has either just happened, with vague concern to 3D and EAX. If you want it properly positioned, occluded, etc like A3D then in this day and age of extremely complex geometry, you are looking at much the same thing as with the light. You have to raytrace all the sound around the place.

It's gonna be hard work.

And my lunch is ready so I don't have time to say more now.

34.
 
Re: No subject
May 22, 2003, 02:19
34.
Re: No subject May 22, 2003, 02:19
May 22, 2003, 02:19
 
I don't know much about the specifics on this but would it be possible for the sound card to process geometric (and material if such is available) information from the CPU in a similar way video cards do?

Isn't this how A3D2.0 in Halflife worked?

Creative own the patents to A3D now so I suppose its possible that their next sound card will provide a similar api.

33.
 
Re: No subject
May 21, 2003, 18:19
Tom
33.
Re: No subject May 21, 2003, 18:19
May 21, 2003, 18:19
Tom
 
Ya, they haven't given too many details - we'll have to wait and see how it turns out in the game. The physics I was referring to was not really the physics of acoustics, but the physics of the game world - the physics and collision detection that govern object behavior and interaction with the world.

32.
 
Re: No subject
May 21, 2003, 15:50
32.
Re: No subject May 21, 2003, 15:50
May 21, 2003, 15:50
 
Deus Ex 2 is going to have a great physics-based sound system.

Fundamentally acoustics is all about physics. While the end effect has been mentioned, it's still unclear however just how in fact this Physics-based sound system is supposed to be implemented.

No doubt they'll still be using EAX for environmental acoustics, and I can only hope that they go full-bore on it.
"Nothing livens up a robotic hymn of doom more than an amazing pair of jugs." - Brak
31.
 
Re: 3D Accelerators Overfiltered?
May 21, 2003, 05:27
Zar
31.
Re: 3D Accelerators Overfiltered? May 21, 2003, 05:27
May 21, 2003, 05:27
Zar
 
The Sentinel,

Oh, no, there will always be a need for artists. Concept, modeling, animation and textures. In fact, the newer stuff just gives them more work to do in quantity and quality. No doubt about that. What I meant was that the limitations of traditional game art will be lifted and there will be new aproaches to creating new effects not seen before.

Texture art will always be needed to at least paint things like blemishes and imperfections as well as the coloration of a given surface. The effect that light has on the surface is what will be calculated. Also, some texturing will relate to supplying artistically realized "maps" of one kind or another that are used by the engine to "key" certain effects off of, such as specular maps. I'm sure other types of "maps" will apply to effects I haven't imagined here.

For sure, though, there will be no more drawing shadows directly into textures. Light and shadow is ususally drawn into static traditional art, but 3D game artists will need to think differently about this in the future. And many of the newer materials effects could never be achieved by an artist the way they need to behave in a game anyway.

30.
 
Re: No subject
May 21, 2003, 02:41
30.
Re: No subject May 21, 2003, 02:41
May 21, 2003, 02:41
 
> I read pretty much the same thing about Thief 3, but
> then they are both being developed by Ion Storm using
> the same engine.

Yep. I suspect the sound system was more necessary for Thief, but Deus Ex got the benefit as well (I imagine that Thief got a few goodies in the same manner...)

29.
 
Re: No subject
May 20, 2003, 23:45
29.
Re: No subject May 20, 2003, 23:45
May 20, 2003, 23:45
 
Tom: "And from what I've read in interviews, Deus Ex 2 is going to have a great physics-based sound system."

I read pretty much the same thing about Thief 3, but then they are both being developed by Ion Storm using the same engine.


28.
 
Re: No subject
May 20, 2003, 22:54
Tom
28.
Re: No subject May 20, 2003, 22:54
May 20, 2003, 22:54
Tom
 
Sensaura and/or A3D supported dynamic, geometry-dependent acoustic environments, but that stuff never really caught on. CPU usage, complexity, developer laziness, the fact that the difference was usually subtle in practice, the domination of Creative and EAX ... ah well.

Developers can, have, and will innovate using their own sound systems. One of Half-Life's lesser-known innovations was its cool sound system which did its own mixing and reverb in software, I believe. And from what I've read in interviews, Deus Ex 2 is going to have a great physics-based sound system.

But developers usually have their hands full with the other parts of a game to put much programming effort into the sound system, it seems...

27.
 
Re: No subject
May 20, 2003, 21:27
27.
Re: No subject May 20, 2003, 21:27
May 20, 2003, 21:27
 
Sound cards...

I don't know much about the specifics on this but would it be possible for the sound card to process geometric (and material if such is available) information from the CPU in a similar way video cards do?

Currently under EAX, acoustic environments are mostly prerendered by the developers. Room dimensions, material information (glass, brick, cement), etc. are all entered into the card for each room where they would be relevant.

In addition to this being a laborious process (Most devs settle for a half-assed sound job), the information does not change dynamically. The parameters given to the sound card can only approximate the actual environment. The disadvantages of this are prominent in outdoor environments, where a large reflective surface such as a building can quickly switch positions relative to the player.

Of course, it would take a fairly powerful sound card to do acoustic modelling the way video cards do transformation and lighting. But with the advent of new 3D engines utilizing complex surface properties, it seems increasingly unfair that sound cards aren't able to share such a wealth of useful data.


This comment was edited on May 20, 21:30.
"Nothing livens up a robotic hymn of doom more than an amazing pair of jugs." - Brak
26.
 
No subject
May 20, 2003, 19:46
26.
No subject May 20, 2003, 19:46
May 20, 2003, 19:46
 
I would like to see a card and a game work in a manner that the card could proccess and store differnt AI routine personalities. Be it an actualy individual or how the AI controls a group of people, it would really take a load off the CPU, as could a physicas card do the same. The CPU could be completley in charge of the OS and managing system resources and data flow, or better yet, put this all in one chip...

Uuhhh. WTF?

25.
 
2 cents
May 20, 2003, 18:23
Tei
25.
2 cents May 20, 2003, 18:23
May 20, 2003, 18:23
Tei
 
Software Rendering for UT2003?

BWAH-HA-HA-HA!

That is all.

24.
 
Re: No subject
May 20, 2003, 18:14
24.
Re: No subject May 20, 2003, 18:14
May 20, 2003, 18:14
 
> Physics card? AI card?

I remember wishing for physics hardware a few years back, but a friend of mine pointed out that you run into potential I/O bottle-necks with that kind of thing. With graphics hardware, you're talking about the end-of-the-line for the process -- no information has to flow back to the game code; the video card just outputs to the display.

With a physics card you'd pass it a bunch of calculations to process, but it has to send the answers back again before the game engine can continue working on the problem at hand, so you have a two way flow of information complicating matters, and the game developers would have to support two significantly different methods of calculating their physics.

On top of that, at the rate that CPUs are advancing it wouldn't be long before your physics card would be outdated.

There are other issues as well, but these were enough for me to think that it was a nice idea, but maybe not a practical one.

23.
 
Re: Remember When
May 20, 2003, 17:35
23.
Re: Remember When May 20, 2003, 17:35
May 20, 2003, 17:35
 
A slight correction: Mike Sartain and I both wrote Pixomatic - Mike generally did the top part of the pipeline, like the drawing primitives, transforms, and clipping, and I mostly did the pixel pipeline. Mike did all the UT2003 integration work, though (thanks, Mike!).

22.
 
Re: No subject
May 20, 2003, 17:34
22.
Re: No subject May 20, 2003, 17:34
May 20, 2003, 17:34
 
Sound cards, graphics cards, I wonder whats next for the gamers budget plans.

Physics card? AI card?

Hhmm...

I have thought about that as well... I would buy 'em. Although, shouldn't the CPU be able to handle the physics and AI since the graphics and audio have their own hardware?

21.
 
Re: 3D Accelerators Overfiltered?
May 20, 2003, 17:33
21.
Re: 3D Accelerators Overfiltered? May 20, 2003, 17:33
May 20, 2003, 17:33
 
Zar, at first you scared me. I was thinking that there would be no use for the texture artist, then I realized how untrue that was. I don't doubt the ability of the post D3 era engines to mimic most materials to some degree, but I truely doubt it will spell the end of careful artistic manipulation. I doubt there will ever be an engine that can generate procedural lips, tree leaves, or bark.

Wish I had more time to finish this thought, but I have to run... quittin' time.


20.
 
Re: 3D Accelerators Overfiltered?
May 20, 2003, 16:54
Zar
20.
Re: 3D Accelerators Overfiltered? May 20, 2003, 16:54
May 20, 2003, 16:54
Zar
 
Creston,

The reason 3D acceleration tended to create "sameness" is that much of the graphics work was handed over to the chipmakers, not the programmers.

For some time the technological understanding of 3D acceleration made it necessary for everyone to do things pretty much the same way, from art to scene rendering. There wasn't much room to deviate from standard, simplistic multi-texturing because that was the only thing that benefitted from acceleration. Doing much more than that would cause an unacceptably slow framerate or would simply not be accelerated. So, for some time, people were forced to try to wrangle the most distinction they could by using the same old techniques.

Now, with the power to do more than simple multitexturing, to do more complex calculations, and to add dynamic elements to a scene on a pixel and vertex level, developers are not constrained to do everything the same anymore. The artistic vision can be more fully realized with fewer constraints. This power is the result of finally having enough power to do as many passes on a pixlel in a scene as need for the effect, plus special features like DOT3 blending and cube maps becoming a viable real-time capability.

But still, as great as the results obviously are because of these few key advances, this only lifts us half-way out of the old school mire. The engines that come after DOOM 3 and HL2 will probably make use of the full programmable pipeline in modern 3D cards. There is really no practical limit to the types of calculations you can do. We will see a full range of surface effects that make leather look like leather, aluminum look like aluminum and skin look like skin.

Already the DOOM 3 engine, for instance, goes a long way toward this with its specular maps. Talented artists can use the DOOM 3 engine to "mimick" a number of materials by carefully calibrating and mixing texture data with specular and bump data. But although this is very effective and flexible, I think the next generation engines will even BLOW DOOM 3 OUT OF THE WATER. Instead of having the capability to "mimick" leather, rubber, milk, blood or brushed metal indirectly with talented artistic "fiddling", you can have a precise calculation that actually mimicks the physical properties of the material with light equations designed for that material. Then its more a matter of turning some knobs and pushing some buttons and letting the engine depict the material, rather than carefully manipulating it artistically.

The possibilities will be endless, and it will again come down to the richness or poverty of artistic imagination and production design, not the limitations of technology.

This comment was edited on May 20, 17:03.
19.
 
No subject
May 20, 2003, 16:46
19.
No subject May 20, 2003, 16:46
May 20, 2003, 16:46
 
Sound cards, graphics cards, I wonder whats next for the gamers budget plans.

Physics card? AI card?

Hhmm...

18.
 
Re: Remember When
May 20, 2003, 16:30
18.
Re: Remember When May 20, 2003, 16:30
May 20, 2003, 16:30
 
Software will never compete with the high end graphics cards. And that wasn’t our goal. We wrote the Pixomatic software renderer for three reasons:

1. Driver issues / bugs / whatever. Basically a fallback for high end 3D games such as Unreal.
2. As Daniel mentions, cheap new PCs with 2 gig plus processors and no real 3D card.
3. For lower end games which don't push graphics nearly as hard as Epic, Id, etc. Of which there are several. A majority of the top selling games, actually.

I'm guessing that none of the folks here care about anything other than possibly #1 and even then they're savvy enough to fix that.

As far as how it looks, we support full bilinear blending so we can run Dungeon Siege or UT2003 without any of the washed out effects. At GDC people couldn’t tell the difference between software and hardware. Other than the framerate, of course. And obviously turning on bilinear at full res means you'd better have a decently fast cpu. (See #2)

-Mike

17.
 
Re: Remember When
May 20, 2003, 16:21
17.
Re: Remember When May 20, 2003, 16:21
May 20, 2003, 16:21
 
Last time I looked at software rendering must've been Quake 2. It looked really gritty and detailed, but ran super-chopola. Might try it again and see what a really high res looks like.
I can definitely see Daniel's point about average Joes buying really meaty machines with ropey onboard graphics. Its an example of pulling the wool over the eyes of the buyer imo. '...but its got a 120 gig hard drive, sir...'

For the victims of this, its good to see there is support from game developers. Jolly good show!


16.
 
Re: Remember When
May 20, 2003, 15:46
16.
Re: Remember When May 20, 2003, 15:46
May 20, 2003, 15:46
 
FWIW, Mike Abrash wrote Pixomatic and Mike Sartain (both now at RAD) did all the integration work into UT2003. I just polished some of it and changed our high level rendering code a bit to better facilitate software rendering.

Most people seem to forget that cheap new PCs (e.g. BestBuy/ DELL) come with horrible integrated graphics and no AGP slot though have a really powerful CPU which is a prime candidate for software rendering. Instead of having to return the game customers can now at least see that it's worth upgrading... our at least we hope so

We recently rented three "tricked out" machines for E3 and imagine our surprise when we realized that they didn't come with an AGP slot to put our shiny new graphics cards in

-- Daniel, Epic Games Inc.

35 Replies. 2 pages. Viewing page 1.
Newer [  1  2  ] Older