article on wsj on nalu and ruby

[Geeforcer]

Correct me if I am wrong, but didn't R100 or R200 demos feature some chick as well? Does anyone remember what I am talking about?

 

[KimB]

Correct me if I am wrong, but didn't R100 or R200 demos feature some chick as well? Does anyone remember what I am talking about?

You can check it out yourself from ATI's demo page:
http://www.ati.com/developer/demos.html

It appears that the "Rachel" and "Ark" demos feature female characters.

 

[Scarlet]

Isn't there a very good reason to do Nalu in the water?

Her hair, which I assume is where more of those polygons are going, will not move very rapidly, thus any performance challenges in that area are less likely to be noticed. If it was hair being whipped about in a breeze, I would assume that is somewhat harder.

Any comments?

 

[Scarlet]

What area of the US has the worst male-to-female ratio?

You guessed it - Silicon Valley.

Don't you mean worst "female-to-male" ratio?

No - I meant what I said. Anytime the ratio of males to females is more than 1:1, it suks. The other way around is heaven. For males that is. Are there any female posters here? Seems like everyone here is male....

 

[jolle]

Isn't there a very good reason to do Nalu in the water?

Her hair, which I assume is where more of those polygons are going, will not move very rapidly, thus any performance challenges in that area are less likely to be noticed. If it was hair being whipped about in a breeze, I would assume that is somewhat harder.

Any comments?

Well, cant say for sure, but the speed things move in 3d, seems to me
as if it wont matter performancewise..
One of the reasons could be to not have to make some background setting
for it..
like that fairy on a singel branch with a texture behind it for FX5.. dawn..

Also the water gives another effect to render, the lightrays from the sun
and how they are cut of in her shadow...

 

[Scarlet]

Re: Girls, girls, girls

Sounds like he's saying Ruby is the sort of thing that looks like a game, whereas Nalu looks like an impressive, but sterile tech demo.

A single Dawn or Nalu may look great on it's own, but it's limited even for a non-interactive game animation, because it's only the one character on a very simple background.

That's just another way of saying that Ruby is less impressive visually.

If I wanted to look at what game characters could look like in upcoming games, I'd look at development screenshots/videos for upcoming games.

Nope. Not saying anything of the sort at all. Ruby is visually more pleasing (to me) than Nalu, but that is a matter of personal taste. What I was trying to convey was that Ruby was placed in a more complex environment than Nalu is, and as a consequence I was better able to see what is possible in future games. Nalu has a lot of polys, but I am not sure that in watching Nalu you get a lot of value out of those polys. IMO nV should have made a more complex scene and spread those polys around some. Three Nalus with fewer polys would (again IMO) have been more impressive than the one.

Just my opinion.

 

[Scarlet]

Isn't there a very good reason to do Nalu in the water?

Her hair, which I assume is where more of those polygons are going, will not move very rapidly, thus any performance challenges in that area are less likely to be noticed. If it was hair being whipped about in a breeze, I would assume that is somewhat harder.

Any comments?

Well, cant say for sure, but the speed things move in 3d, seems to me
as if it wont matter performancewise..
One of the reasons could be to not have to make some background setting
for it..
like that fairy on a singel branch with a texture behind it for FX5.. dawn..

Also the water gives another effect to render, the lightrays from the sun
and how they are cut of in her shadow...

The sunbeams and shadows things could just as easily been demonstrated in an early morning scene in a forest/meadow with light fog near the ground (Hey - they coulda done that with Dawn!). Besides, I thought the modeling of that was not quite so well done. As a RL diver, the sunbeams looked kind of "stiff" to me. Any other divers seen the demo and have an opinion?

 

[KimB]

Re: Girls, girls, girls

What I was trying to convey was that Ruby was placed in a more complex environment than Nalu is, and as a consequence I was better able to see what is possible in future games.

I don't really see how that matters. Either way, it's not current games. With a complex object in a (relatively) simple environment, you're just looking a little bit further into the future.

As for rendering entire environments, well, if you consider the Chimp demo (I haven't seen Ruby, of course), the view is limited. This is obviously done so that the developers can maximize poly counts while not worrying so much about the person showing off the demo pointing the view in the wrong direction and tanking performance. That really turned me off of the supposed "high detail environment" of the Chimp demo.

If Ruby's the same, color me unimpressed. Give me an Unreal Engine 3 demo video any day.

 

[KimB]

The sunbeams and shadows things could just as easily been demonstrated in an early morning scene in a forest/meadow with light fog near the ground (Hey - they coulda done that with Dawn!). Besides, I thought the modeling of that was not quite so well done. As a RL diver, the sunbeams looked kind of "stiff" to me. Any other divers seen the demo and have an opinion?

But it was much easier to have realistic hair animation underwater. The equations are simply much less complex. Heck, I've never seen convincingly-rendered dry hair (as an example, consider the Final Fantasy movie).

 

[jolle]

another thing to remember is that the the demos are made with a purpouse.
Since Ruby has enviroment around her, its quite possible they are going to
demo Camera Focus, which requires a backgroundsetting to blur for it to make
any sense.. preferably with a few things at different distances from the Focal
point (Ruby) so you se how it blurs things in diffrent intensities down the
depth of the scene...

Or if you want the tech demo to showcase the polygonpushingpower of the
card, you can slap on a backgroundsetting with alot of stuff in it..

Heard alot of "Nalu is this and that many polys, and Ruby is "only" xxx polys"
crap, countered with "but ruby has a background yadayada"..
its all pretty uninteresting unless you compare 2 demos with the intent of
showcasing polycounts, which i dont think is the main feature of any of the 2.

Besides, the industry is moving to "smarter polys" instead of JUST more..
with Normal mapping, bump mapping and Displacement mapping which
are better alternatives then spending alot of polys on brickwalls, mountain
sides and whatnot...

 

[The549]

I think that it would be cool if both nvidia and ati had both single hi-poly characters and characters rendered in environments. And Chanloth you were saying that FF the movie had good hair rendering, right?

 

[KimB]

I think that it would be cool if both nvidia and ati had both single hi-poly characters and characters rendered in environments. And Chanloth you were saying that FF the movie had good hair rendering, right?

No, I'm saying it was unconvincing. Not that it was bad, compared to what anybody else has done, but I think it just goes to show how hard hair really is to render.

 

[SiliconAbyss]

Heck, I've never seen convincingly-rendered dry hair (as an example, consider the Final Fantasy movie).

I thought the hair was very good in FF, not perfect but almost. It's the last small percent that always gives an illusion away. I do agree, hair has got to be one of the hardest rendering challenges there is. The physics simulation of hair is like trying track the movements of a pigeon eating crumbs, only harder.

 

[KimB]

I think the main problem is that hair is typically pretty stiff, but not perfectly so. There's also the friction between each neighboring strand to consider: not a simple problem.

 

[The549]

Yeah, the hair looked good, but the physics was off; it looked a bit heavy i think. Like they forgot air resistance/hair had too much inertia.

 

[Heathen]

With a complex object in a (relatively) simple environment, you're just looking a little bit further into the future.

Nalu is a dead end which in no way represents the future. Environments are going to get far more coplex and the number of chaacters is going to increase.

 

[jolle]

well, future games will have higher poly counts, tho shifted more to enviroments then characters like you say..
Even so, surface effects such as Normal mapping frees up polygons from enviroments that could be better used on characters ingame, since those effects doesnt help the outline/contour of the models, while you dont se
many walls from the side to notice how the bricks arent really 3d..
Except displacement mapping, but it doesnt seem as if its going to be used
in a big way in a while.. some monsters will use it in HL2 its said tho..

But you do have a point with it, then again Techdemos usually dont compare
well to games, specially since they dont use always use other things games
do such as AI, physics, big maps and such..

 

[Scarlet]

I would like to see a good demo of normal mapping. Anyone have a link? I understand it is a good way to gain the effect of a lot of detail without a lot of work.

 

[jolle]

there is this HL2 video..
http://www.fileplanet.com/files/120000/129654.shtml

they show how the generate the map from source geometry in Softimage
and add it to lowpoly surface..

what it does it save the direction of each poly from the source into a texture
and use that information to cast shadows into itself, even if the surface is
flat, sortof like bumpmap, but BM only allows for height to be specified, not
which way the "fake" polygons are facing..

For a nice realtime demo, you got ATIs Car demo.. for 9700Pro..
looks like this:
http://www.tomshardware.com/graphic/20020718/images/car2.jpg
the "indentation" there on the hood is physicly flat, but using normalmaps
it gives a good illusion of being.. eh.. not flat.. hehe

 

[Ylandro]

Yeah, the hair looked good, but the physics was off; it looked a bit heavy i think. Like they forgot air resistance/hair had too much inertia.

That might have been intentional. Makes it a little more moving, more dramatic. Model it correctly, and you wouldn't notice it anymore.

It's like adding a ventilator when shooting photo's of a model...