Why RE4's lighting may be the GCN's Best

[Li Mu Bai]

Actually it seems to me that EMBM can do everything that DOT3 can, plus additional effects that I have yet to see DOT3 do.

Could you give me a perceived example Tagrineth? I'm being sincere here. Also the major difference between the two is that emboss does per-pixel without adding the geometric data. Also emboss doesn't support the computation of specular highlights, but this can be circumvented ala RS3, by adding non-bumpped specularity.

 

[Bohdy]

Li Mu Bai: I made that mistake too originally;

EMBM = Environment Mapped Bump Mapping, ! Emboss Bump-Mapping (which is a hack). ;-)

 

[Fafalada]

Actually it seems to me that EMBM can do everything that DOT3 can, plus additional effects that I have yet to see DOT3 do.

They aren't comparable in that fashion.

DOT3 is a dot product between two vectors.

EMBM written out as shader math is a composition of several dot products and a dependant texture lookup.

In other words, the former is one of the elementary operations that compose the latter. Certain graphic chips implement the EMBM sequence as a hardwired formula where you only change input parameters. IIRC, that's how it's implemented in Flipper - however that isn't the main reason why it's faster then NV2a at it

 

[Tagrineth]

Actually it seems to me that EMBM can do everything that DOT3 can, plus additional effects that I have yet to see DOT3 do.

Could you give me a perceived example Tagrineth? I'm being sincere here. Also the major difference between the two is that emboss does per-pixel without adding the geometric data. Also emboss doesn't support the computation of specular highlights, but this can be circumvented ala RS3, by adding non-bumpped specularity.

I'm not talking about embossed bump mapping, hahaha... even the poor feature-starved GS can support that with ease. Emboss mapping has its uses, but it's crap for general use.

EMBM is Environment-mapped bump mapping, which was pioneered by Matrox. They have a bunch of demos and stuff at their site.

And as faf said, EMBM includes computing dot products, so logically it would have to be able to do anything DOT3 could do.

 

[Li Mu Bai]

Blast, the acronyms are so close to one another.

 

[Fox5]

Ah! I'm so confused...so what's the difference between the two? Does this mean gamecube has better bumpmapping than xbox, but it has never been used? BTW, does EMBM on gamecube come with a greater performance hit than dot 3 on xbox?

 

[cybamerc]

Tagrineth:

> EMBM is Environment-mapped bump mapping, which was pioneered by Matrox.

In a manner of speaking yes, but it was invented by Bitboys.



Fox5:

> Does this mean gamecube has better bumpmapping than xbox

Different. They each have their uses eventhough they can be used for similar effects.

> but it has never been used?

It has been used more than a few times (wouldn't say many though).

> BTW, does EMBM on gamecube come with a greater performance hit
> than dot 3 on xbox?

I dunno about real world performance (Xbox has higher clockspeed but there are other factors). EMBM can be done in as little as two cycles on Cube I believe.

 

[Li Mu Bai]

Marconelly wrote: Splinter Cell had a mixture of dynamic vertex lights and projected light maps, and also HDR effects, but it's shadows implementation was all over the place.

Dynamic vertex lights are accomplishable, as are projected light maps on the GC. (ED used them) But what of HDR effects? ICO & GT3 use versions, or possibly hacks of it because they are done in software & not on a hardware "shader" format. (although still looking very much impressive regardless) HL2 as well as SC used pixel shaders to produce the effect, the TEV can produce the same effects ala RS3. So was Ubi wrong initially when they said only the Box could reproduce such effects, or did system fanboys spin this statement? (as the GC got the PS2 port primarily)

 

[ShinHoshi]

Correct me if I am wrong :

Bump-Mapping is just used to simulate roughness on planar surface by just lighting it like if it was actually roughness.

Environmental Bump-Mapping is used mainly to calculate reflections over a surface or a volume. You simplify the environment to reflect by just using an environmental box/Sphere to do your calculations.

 

[Teasy]

ShinHoshi

I think what your refering to there is Cube Environment Mapping. The technique used in WaveRace: Blue Storm for the water reflections.

Environment Mapped Bump Mapping is used for simulating roughness/bumps. It can also be used for stuff like heat distortions and surface morphing ect.

 

[StefanS]

ShinHoshi
I think what your refering to there is Cube Environment Mapping. The technique used in WaveRace: Blue Storm for the water reflections.

Are u sure? I've never heard of WR:BS using Cube Environment Maps...

 

[ShinHoshi]

Yep, that's what I thought. Just that's the first time I heard about EMBM.

 

[cybamerc]

The term "Environment Mapped Bump Mapping" is actually a little misleading because the reflection map doesn't actually have to be a reflection map (nor of a particular type). It can be anything.

 

[Fox5]

ShinHoshi
I think what your refering to there is Cube Environment Mapping. The technique used in WaveRace: Blue Storm for the water reflections.

Are u sure? I've never heard of WR:BS using Cube Environment Maps...

I have. It's used for the reflections on the water.

 

[StefanS]

ShinHoshi
I think what your refering to there is Cube Environment Mapping. The technique used in WaveRace: Blue Storm for the water reflections.

Are u sure? I've never heard of WR:BS using Cube Environment Maps...

I have. It's used for the reflections on the water.

O.K. that explains why WR is only 30fps...

 

[Fox5]

But why is it 30 fps in 1 player or 4 player?(ok, in 4 player they do drop the detail)

 

[StefanS]

But why is it 30 fps in 1 player or 4 player?(ok, in 4 player they do drop the detail)

Well, Cube Environment Mapping has to be done by Gekko, so the limiting factor has to be there or between Gekko and Flipper... The physics and additional calculations stay more or less the same in multi and single player, while geometry increases. Thus geometry shouldn't be the limiting factor...

 

[Li Mu Bai]

Allright, now back to the question at hand. True HDR functionality, done by the TEV's pixel shader equivalent. (as pixel shader is mainly an N'vidia coined marketing term to describe programmable color combiners) After reading ERP'as TEV assessment, it seems entirely possible. Yes? No?

 

[Evil_Cloud]

But why is it 30 fps in 1 player or 4 player?(ok, in 4 player they do drop the detail)

Well, Cube Environment Mapping has to be done by Gekko, so the limiting factor has to be there or between Gekko and Flipper... The physics and additional calculations stay more or less the same in multi and single player, while geometry increases. Thus geometry shouldn't be the limiting factor...

Perhaps there wasn't time to get a constant 60fps in time... It was a launch game after all... Halo was also 30 fps...

 

[StefanS]

But why is it 30 fps in 1 player or 4 player?(ok, in 4 player they do drop the detail)

Well, Cube Environment Mapping has to be done by Gekko, so the limiting factor has to be there or between Gekko and Flipper... The physics and additional calculations stay more or less the same in multi and single player, while geometry increases. Thus geometry shouldn't be the limiting factor...

Perhaps there wasn't time to get a constant 60fps in time... It was a launch game after all... Halo was also 30 fps...

I remember an interview where they said they were deliberately aiming for 30fps to insure a solid frame rate...
i thought about it again...
Gekko and Flipper share a bandwith of 2.6 GB/s. Now Flipper is eating his usual part of that. But Gekko is calculation the AI, physics (that involves wave calculation and interference; a heavy load!) and he has to do CEM for nearly 50% of the polygon involved, which also a huge load. At 60fps that would eat twice the bandwith...