Which BLEEDING EDGE Feature is Valve Likely to add first?

Which BLEEDING EDGE Feature is Valve Likely to add first? Post a Poll

  • 1. SM 3.0 (Shader Model 3.0)

    Votes: 60 26.3%
  • 2. Parallax Mapping (Aspect of SM 3.0)

    Votes: 7 3.1%
  • 3. 64-Bit Optimizations

    Votes: 56 24.6%
  • 4. ULTRA-TEXTURES (ala Doom 3)

    Votes: 18 7.9%
  • 5. True HDRI Support (High-Dynamic Range Imaging)

    Votes: 60 26.3%
  • 6. Side-choice: Beta Phys-Gun (Blue Beam in E3 vids)

    Votes: 11 4.8%
  • 7. 3DC

    Votes: 16 7.0%

  • Total voters
    228
mirageacg said:
Bump Maps and Normal Maps are two different things, similiar principle tho, a bump map consists of a texture composed of a height field which is designated by the range of shades from black to white, black being the lowest, and white being the highest, also bump maps do not affect geometry in any way shape or form.

Normal maps, also known as Displacement maps, take the idea of a bump map, but to a higher level, the height fields are no represent by the full range of colors (0,0,0 to 256,256,256) allowing for greater detail, and if so desired normal maps can "physically" alter model/terrian geometry, while bumps maps only simulate the said effect.

Actually... no.

A normal map is an RGB map that tells the engine how light falls on the surface. When a normal map is created, it requires an extremely high res version of a model. Then, the 'normal' of all those polygons is stored in the map. The normal is the angle the polygon is at, sort of anyway. The normal is stored in an RGB value (red for the x-axis, green y-axis and blue z-axis) and that tells the engine what the normal of that polygon is and how light affects the normalmap. It simulates a high res model without adding polygons to it. It's a completely different approach than bump mapping is. And bump mapping is closer to displacement mapping than normal mapping is.

And btw, I think you can store the heightmap/normalmap for parallax mapping in the alpha channel of the normalmap. Not sure though.
 
Alrighty, I know when i am beat, but not on all fronts, that is great that HL2 supports parallax mapping and that it can be enabled and viewed in the VST, that is a happy mistake on my part. (But out of curiosity what vid card are you using, because if it is anything lower then a geforce 6, then that is "simulated" Parallax mapping, meaning oyu don't get teh full mind blowing, oh-my-god-my-eyes-are-melting-and-staining-the-shag-carpet effect).

But concerning normal and bump maps, normal maps are closer to true displacement mapping due to the fact that the normal map is based solely on the geometry of the model, resulting in a perfect displacement. Bump mapping however relies upon using a light-source in order to determine what geometry is simulate, so kinda like a limited normal map, with only the green field enabled. And as far as i know Valve stated that they were using bump maps, because they didn't need any displacement due to their extremely high poly count models, only bump to provide texture/depth.
 
mirageacg said:
Alrighty, I know when i am beat, but not on all fronts, that is great that HL2 supports parallax mapping and that it can be enabled and viewed in the VST, that is a happy mistake on my part. (But out of curiosity what vid card are you using, because if it is anything lower then a geforce 6, then that is "simulated" Parallax mapping, meaning oyu don't get teh full mind blowing, oh-my-god-my-eyes-are-melting-and-staining-the-shag-carpet effect).

But concerning normal and bump maps, normal maps are closer to true displacement mapping due to the fact that the normal map is based solely on the geometry of the model, resulting in a perfect displacement. Bump mapping however relies upon using a light-source in order to determine what geometry is simulate, so kinda like a limited normal map, with only the green field enabled. And as far as i know Valve stated that they were using bump maps, because they didn't need any displacement due to their extremely high poly count models, only bump to provide texture/depth.

Sure it could be it looks better on a GF6 and that DX9c supports it better, but it's certainly possible on lower hardware. I've got a techdemo running on my 9800 Pro.

And normalmaps are not suitable for displacement mapping because displacement mapping happens only at one axis at a time (if the z-axis moves up and down, the x and y-axis stay at the same position) and normal mapping uses all three axis.
 
MrBongo said:
Either you're being smart or sarcastic, but yes, Doom III- despite rumors -does in fact have textures. :angel:

i want so se some damned evidence, man!! show me them!!! ive got the game damnit! i cant see anything to back it up!
 
Neutrino said:
...
And the source engine supports realtime lighting? I don't think it would without a major rewrite and update. It's still using lightmaps after all.
http://www.valvesoftware.com/sourcelicense/enginefeatures.htm said:
- Dynamic lights, vertex lighting and light maps, many light types including flickering, pulsing etc.
- Real-time radiosity lighting

I am not sure if radiosity lighting is the same as just realtime lighting hmm..
 
The only thing missing is dynamic shadows... Would've been just short of the ultimate gaming engine (for now.. :p )...
 
that real-time radiosity thing must be a mistake,, it takes around 9 hours to render 1 "frame" of radiosity information for one of my halflife 1 maps

radiosity is a solution usually used to get really nice lighting that is totally static then apply it as an irradiance map (or light map). It's essentially painting the shadow information onto the texture...


The day real-time radiosity is a reality, we'll see some truelly insane levels (think moving clouds projecting shadows on the ground)
 
I put 3dc only because thats prolly what valve will do first since ati doesnt support sm 3 yet. HDR is already supported by the engine. HDRI is somthign totally different which has to do with static rendering, so no i dont think hdri is even needed lol.

In a perfect world i would like to see sm3 specific stuff as well as optimized sm3 pre-existing stuff. At the same time as a 3dc patch. To keep it fair between ati and nvidia.

However i highly doubt this will happen =(

shinobi real time radiosity isn't that far away heh.

mirage hl2 does support normal maps.
 
I have a question for yall is
mat_parallaxmap enabled to 1 by default for those of you with dx9 cards?
 
killahsin-[CE] said:
I have a question for yall is
mat_parallaxmap enabled to 1 by default for those of you with dx9 cards?

yes parallax is already supported by the source engine and all dx9 cards but it's not used in anyway yet.

I don't think it will be used in CS:S in anyway as it can cost quite a bit of performance. They may use parrallax in HL2 sooner or later though , at least i hope so
 
killahsin-[CE] said:
I have a question for yall is
mat_parallaxmap enabled to 1 by default for those of you with dx9 cards?

Yes.......
 
killahsin-[CE] said:
HDR is already supported by the engine. HDRI is somthign totally different which has to do with static rendering, so no i dont think hdri is even needed lol.

Actually HDRI is a specific application of the HDR concept, ie a skybox is an HDRI, and when HDR support is enabled, then voila!

killahsin-[CE] said:
mirage hl2 does support normal maps.

Could someone please show me where Valve has stated that HL2, (not Source), USES normal maps.
 
I bet 3Dc will get in there first if not out of the box.
HDR, and SM3.0 probably will be added in later.

Even from interviews on Beyond3D, Tim has said the UE3 doesn't use true displacement mapping that is in the 6800. It uses high poly models along with virtual displacement mapping for large scenes like walls etc.
 
search for D3DTutorial10_Half-Life2_Shading.pdf
in google, and you will find out a whole lot...
 
HL2 environments are VAST.

One of my friend who playtested HL2 at G-phoria said that after the place where you fight combines + rollermines you reach a town on a cliff, near a large bridge - it's just massive. (won't say more since it could mean spoilers)

Here, PHL has a screen of that area (this is 100% in-game as witnessed by my friend):

http://www.planethalflife.com/screenshot.asp?src=/features/articles/sigg04/phl_c17_Slide15_hi.jpg

It's extremely stupid to use dynamic shadowing in such massive environments - or otherwise you'll have massive framerate drops. The hell areas in Doom 3 were proof enough - and they were much smaller in comparison.
 
mirageacg said:
Could someone please show me where Valve has stated that HL2, (not Source), USES normal maps.

I can show you a picture:

http://games.telenet.be/images/reviews/screenshots/175s_35.highres-concept.jpg

This is a photograph of Zbrush, which is used to create high res models to use them as normal maps on low res ones.

Furthermore, there's a video of Siggraph or GDC where a Valve developer shows how to make normalmaps of a high res wall in XSI and shows the low res wall with normalmap in game.
 
No no what i am asking is on my 6800gt card mat_parallaxmap is set to ON by default. I am asking if for someone whoi has a 9600-x800xt pe is mat_parallaxmap set to 1 by default for those specific cards. Or is it set to 0?
 
Back
Top