What in the world is up with Nvidia?!?!

[BenSkywalker]

Not according to the game developer Gearbox; nVidia cards are not rendering the camo how it was meant to be rendered.

Find a quote.

If you think so then try running in on a DirectX 7 board and look at the difference, and then do the same for Doom III.

OK, FC looks different- although it still looks far better then most games coming out today. D3 won't run on a Radeon or Kyro.

The bulk of D3's shaders are just doing simple multi-texturing

That is some pile of BS. D3 is using shaders(not the BS PR definition either), albeit very simplistic ones, for nearly every on screen pixel.

in terms of flashy effects you only get the odd haze or light effect here and there Not so in FC where the entire game is built around shaders.

Do what you are saying yourself. Go ahead and kill all of the shaders in FC and check it out. Tell me when you get D3 to run on either a Radeon or a Kyro. For a long time people were claiming the R100 was the basis for DX8 and that can't even launch the game. Doesn't make it a shading monster in any way, but the R100 significantly exceeds the DX7 specs.

Excuse me? Have you somehow missed the benchmark results where the NV3x is sometimes half the speed of the R3xx? This thread is filled with shader based titles that repeatedly demonstrate this fact so how long will you continue to deny the results?

Forget the NV3x and R3x0, the R420 and NV40 both are too slow to put a serious shader load on them. The R3x0 is a lot better then the Voodoo1 at shaders too, is that going to make you talk about how playable it is when it's pushing under 5FPS running 640x480 when we get some honest shader heavy games? Are you going to talk about how good its shader performance is then?

You seem to be on a troll mission to turn this in to a NV3X v R3x0 discussion- I'm running the R3x0 here myself and have been for some time as you know. The R3x0 on a relative basis has considerably faster shaders then the NV3x, find me saying otherwise anywhere. The problem is, relative is needed in there as a disclaimer as the parts are both so incredibly slow.

The benchmarks speak for themselves; nothing to do with hype.

WTF are you on about with this statement? What benches show what? Last I was aware Mafia ran fairly evenly on the R3x0 vs the NV40. That game has almost no shaders in it(and you can kill the couple it has easily enough). Two of the games you listed off as titles using shaders(JKII/JKIII) run significantly faster on the NV3x hardware then they do on R3x0 hardware. The benches show us how these parts perform, they do not in and of themselves say a d@mn thing about the shader load they are under.

 

[oldfart]

Cloaking devices are supposed to make you invisible, not make you stand out more.

Been through this before. No, its NOT supposed to make you invisible. Wow, rendering nothing. Now that is hard to do! Cloaking is supposed to look like the effect in the Predator movie and no, it does not make you stand out more than uncloaked. With an ATi R3xx card, it looks exactly the same as it does on the Bungie Web site. With an nVidia card, the effect is missing. If during multiplayer gaming you prefer to be invisible, that is another matter and is a personal preference.

 

[BenSkywalker]

A) My R9800Pro doesn't look quite like that 'screenshot'

B) That 'screenshot' is a render, not done by any video card(you know this of course)- we've been through this before

Been through this before.

Yes, and there were vid clips showing you that you are not completely invisible in the nV shots, just harder to see unlike the neon billboard effect you get with the way Gearbox has the shaders set up for ATi parts.

The nVidia boards are capable of running it just like ATi's parts with a few clicks(and 3DAnalyze) if GearBox honestly meant for it to be the same. All of the boards work that way just fine.

 

[oldfart]

Originally posted by: BenSkywalker
A) My R9800Pro doesn't look quite like that 'screenshot'

B) That 'screenshot' is a render, not done by any video card(you know this of course)- we've been through this before

Been through this before.

Yes, and there were vid clips showing you that you are not completely invisible in the nV shots, just harder to see unlike the neon billboard effect you get with the way Gearbox has the shaders set up for ATi parts.

The nVidia boards are capable of running it just like ATi's parts with a few clicks(and 3DAnalyze) if GearBox honestly meant for it to be the same. All of the boards work that way just fine.

My 9800P looks very much like that shot. A heck of a lot closer to it than the practically invisible way it looks on an nVidia card. As I said before, it comes down to personal preference. Which way do YOU like it to look? Obviously Bungie thinks it should look like the "predator camo" since that is the effect they show on the web site. I think its a pretty cool effect. I dont play Halo MP, but in SP, the baddies are harder to see cloaked than uncloaked. It does not make them stand out more by any stretch.

My original point was that you claimed ATi was cheating and not rendering the camo correctly. That is not true. You may not prefer that effect, but that doesn't mean ATi is doing it wrong.

 

[DAPUNISHER]

To interject my personal experience, I've played Halo on a 9800p and 6800GT and noticed no substantial changes in IQ </subjective testimony> Of course I was playing and not siteseeing per say

 

[kylebisme]

Originally posted by: BenSkywalker

Not according to the game developer Gearbox; nVidia cards are not rendering the camo how it was meant to be rendered.

Find a quote.

How about active camo for us NVidia users?
Not in the inital release. It is high on the first patch list.

http://gbxforums.gearboxsoftwa...b=5&amp;o=&amp;fpart=1

Will you give up your charade now?

 

[BenSkywalker]

My 9800P looks very much like that shot. A heck of a lot closer to it than the practically invisible way it looks on an nVidia card.

Closer then it does on a nVidia card without using 3DAnalyze. You can pick and choose which way you want to play it, with little effort. The shot you keep posting is a render however, the R3x0 path may be closer to it then the default nV path, but it certainly isn't the same.

Which way do YOU like it to look? Obviously Bungie thinks it should look like the "predator camo" since that is the effect they show on the web site.

The wavy way looks cooler- it simply loses all effectiveness as a cloaking device.

I dont play Halo MP, but in SP, the baddies are harder to see cloaked than uncloaked. It does not make them stand out more by any stretch.

I don't think so, not at all. Particularly when dealing with indoor areas the level of refractiveness makes them stand out far more then the normal Elite. In terms of MP, that is where it really comes across as a cheat as nV users are at an enormous disadvantage as they can't see.

My original point was that you claimed ATi was cheating and not rendering the camo correctly.

I looked in to it some time ago and found out that it was GearBox who screwed it up(back when we first had this discussion). It was just as valid as all the 'cheating' nV was slammed with for the better part of a year.

Still waiting for all of those on the bash nV bandwagon a few months back to jump on the same boat now that we know ATi was pulling the same sh!t all along(not with Halo, but in general 'cheats').

Snowman-

Will you give up your charade now?

Give up what charade? That they are going to include a switch to enable something that already works anyway? Must of been processor limited.

 

[kylebisme]

I'm guessing that is a "no" in delusional asshat speak.

 

[kylebisme]

Originally posted by: BenSkywalker

My 9800P looks very much like that shot. A heck of a lot closer to it than the practically invisible way it looks on an nVidia card.

Closer then it does on a nVidia card without using 3DAnalyze. You can pick and choose which way you want to play it, with little effort. The shot you keep posting is a render however, the R3x0 path may be closer to it then the default nV path, but it certainly isn't the same.

Oh ya and 3DAnalyze makes the text flash out when you use it so you can't read your teammates messages or see who killed who; great way to play the game.


also, there is no evedence to prove the shot is a render no mater how many times you say it.

 

[BenSkywalker]

also, there is no evedence to prove the shot is a render no mater how many times you say it.

For ignorant people who know nothing about the game's rendering engine I would say you are right. It is actually plainly obvious to anyone remotely familiar with the game(I'll give you a hint because of your displayed comprehension in this thread so far- some rendering feature doesn't work in Halo in no small part due to the active camo/shader- that rendering feature is clearly in use in that screenshot).

 

[kylebisme]

I posted a virtualy identcal shot from my 9700pro last time we went though this, Ben; I'm not going to waist my time humoring your dillusions by doing it again.

 

[kylebisme]

Oh, I just realized your vague allusion was to the AA; which of course would be silly to argue directly as one can get AA in a screenshot by many means including simply downsizing the image.

 

[BenSkywalker]

I posted a virtualy identcal shot from my 9700pro last time we went though this, Ben; I'm not going to waist my time humoring your dillusions by doing it again.

I have a R9800Pro myself, I know exactly what it looks like and it isn't quite what they are showing.

Oh, I just realized your vague allusion was to the AA; which of course would be silly to argue directly as one can get AA in a screenshot by many means including simply downsizing the image.

Would you care to explain how it is they are achieving a MSAA downsample effect using a 2D filter? Also- where did you get your copy of Halo that plays in 3rd person while you are zooming in with the sniper rifle? The screen is a render, I doubt you will ever come to that realization but it is painfully obvious to anyone slightly familiar with the game or its rendering engine(they don't even have to be familiar with both in this instance).

 

[kylebisme]

Originally posted by: BenSkywalker

I have a R9800Pro myself, I know exactly what it looks like and it isn't quite what they are showing.

I think it looks the same, others have said the same, if you don't think it is the same then post us a shot to show us what you think is different.

Originally posted by: BenSkywalker
Would you care to explain how it is they are achieving a MSAA downsample effect using a 2D filter?

Why would I? I never made such a claim. I didn't even bother to blow up the shot to look for signs of multisampleing. Doing so wouldn't prove anything; obviously developers have more access to their programs so it isn't unreasonable to assume they might have the ablity to use multisampling with an internal build of the game.

Originally posted by: BenSkywalkerAlso- where did you get your copy of Halo that plays in 3rd person while you are zooming in with the sniper rifle?

I'd probably have to be working for the developer to do that, like the people who took the those screenshots.

Originally posted by: BenSkywalkerThe screen is a render..

So is that why it is in the screenshots gallery with other obviously in game shots with AA instead of in the renders gallery

Originally posted by: BenSkywalker
..I doubt you will ever come to that realization but it is painfully obvious to anyone slightly familiar with the game or its rendering engine(they don't even have to be familiar with both in this instance).

"Anyone" being you, and... well... just you so far, eh? And where are your credentials; in the bunk arguments above? :roll:

 

[BFG10K]

That is some pile of BS.

Is it? So just what exactly are those shaders doing apart from the odd light and haze effect? Where is the equivalent to individual swinging blades of grass, rippling water and rolling waves that Far Cry has?

D3 is using shaders(not the BS PR definition either), albeit very simplistic ones, for nearly every on screen pixel.

To do what exactly, apart from multitexture?

Tell me when you get D3 to run on either a Radeon or a Kyro.

Tell me when you get DEIW or Thief 3 to run on a GeForce; does that make it more shader intensive than Doom III? Yet you continually dismiss DEIW &amp; T3. Why? Why the double standard in your reasoning?

Forget the NV3x and R3x0,

I most certainly will not. You've been denying and twisting the issue for almost a year and now you want to sweep it under a rug? No thank you. Why don't you just accept the evidence and move on?

The R3x0 is a lot better then the Voodoo1 at shaders too, is that going to make you talk about how playable it is when it's pushing under 5FPS running 640x480 when we get some honest shader heavy games?

Now you're just being obtuse and you know it.

You seem to be on a troll mission to turn this in to a NV3X v R3x0 discussion-

Oh pu-lease.

The R3x0 on a relative basis has considerably faster shaders then the NV3x, find me saying otherwise anywhere.

Start with this very thread. If you weren't making excuses for every scenario that disproves your claims none of us would be here and this thread would have died a long time ago.

Last I was aware Mafia ran fairly evenly on the R3x0 vs the NV40. That game has almost no shaders in it(and you can kill the couple it has easily enough).

So now you're using games that have little shader effects to disprove the NV3x's inferiority? Have you momentarily forgotten what side of the fence you're on or something?

Two of the games you listed off as titles using shaders(JKII/JKIII) run significantly faster on the NV3x hardware then they do on R3x0 hardware.

That's quite true. Now what about the rest where the NV3x at times gets a proper pasting? Oh that's right, "those aren't really shader games" or somesuch, so they don't count. Right?

 

[BenSkywalker]

Snowman-

So is that why it is in the screenshots gallery with other obviously in game shots with AA instead of in the renders gallery

A good deal of those shots are renders too(just by quickly glancing at them). It isn't really worth having this discussion with you as you don't know what you are talking about(check out exactly how they pull off active camo compared to how MSAA works for starters- this has been explained by numerous sites so you don't need to listen to me about it).

BFG-

Is it? So just what exactly are those shaders doing apart from the odd light and haze effect?

You mean besides the entire lighting engine, shadows and normal maps which between them cover pretty much every pixel on screen in the game? I suppose not much, since it doesn't have water and all.

Where is the equivalent to individual swinging blades of grass, rippling water and rolling waves that Far Cry has?

Don't forget shiny pipes too.

Tell me when you get DEIW or Thief 3 to run on a GeForce; does that make it more shader intensive than Doom III? Yet you continually dismiss DEIW &amp; T3. Why? Why the double standard in your reasoning?

You have been saying D3's shaders are just for multi-texturing, the V2 can do that with no problem so you shouldn't have any issue gettting the game to run on a R100 at all.

I most certainly will not. You've been denying and twisting the issue for almost a year and now you want to sweep it under a rug? No thank you. Why don't you just accept the evidence and move on?

What evidence? I'm still waiting.

Now you're just being obtuse and you know it.

Not in the least. Shaders have been around for a long time in off line rendering, they were back when I used to work with it and some time before that too. I know what constitutes a heavy shader load.

Start with this very thread. If you weren't making excuses for every scenario that disproves your claims none of us would be here and this thread would have died a long time ago.

First off, quote me saying otherwise in terms of the nV3x and R3x0 in terms of shader speed. Quote me from any thread saying otherwise. Either stop running off at the mouth about what I've been saying or prove it. The archives are here as is this thread. I'm assuming you have twisted your mind to read me commenting on the nV3x parts running as fast as R3x0 in 'shader' games to read that the nV3x is comparable- the reality is that those numbers should show you the games are NOT as shader intensive as you may want to believe.

Next- Excuses for what scenarios? The R3x0 kills the NV3x in FC, does in TR:AoD too. You mention JKII/JKIII as shader games where the NV3x kills the R3x0- someone else brings up DF:BHD where the NV3x and R3x0 are swapping the lead back and forth- I brought up Halo as another 'shader' game where the NV3x and R3X0 are neck and neck. I didn't bring up three of the games that are 'shader' titles that the nV3x is as fast or faster then the R3x0- I have been saying that the games are not terribly shader intensive(which if you would get off your mission you may realize props up the discussion about the R3x0's faster shaders).

So now you're using games that have little shader effects to disprove the NV3x's inferiority?

I have not been blasting shaders because ATi is faster at them if you would for once read what I'm saying instead of dreaming things up in your head- I'm pointing out how useless shaders have been. Hell BFG, out of all the shader fanatics we have around here you've been the only one that has gone so far as to simply list off the shader games you have played. Two years we have been listening to the PR BS and you have been the only one to try and come up with a list.

Have you momentarily forgotten what side of the fence you're on or something?

I'm on the reality side of the fence. Sometimes you are there with me, other times you are not. In the past two years I've spent $500 on ATi hardware and $0 on nV- because I'm so biased and all.

That's quite true. Now what about the rest where the NV3x at times gets a proper pasting? Oh that's right, "those aren't really shader games" or somesuch, so they don't count. Right?

No, I'm pointing out that JKII/JKIII are not shader intensive at all. If D3 wasn't so heavy on the stencil fill and if its' shaders were altered a bit the R3x0 would be killing the NV3x0 there too. Why can't you understand what I've been saying about shaders for the last couple of years? There is roughly parity right now in terms of shader performance(NV4x v R420)- it has nothing to do with being partisan. SM 3.0 is nigh useless(which I've pointed out in the past) putting the NV4x and R420 on close to level ground- that level in terms of shaders just happens to still be far too low to put a real shader load on them.

 

[Regs]

52 FPS I think is a very good thresh hold. Most games are limited now anyway. The day's of Unreal 2003 Tournament are over. lol.

Im kind of siding with Ben however on this topic. There are simply no games out to test the theory of shader performance on the 9700/9800 cores. If I could find it, there is a quote on Anandtech.com that states just this fact.

But you cannot argue, absolutely cannot argue that the Nvidia FX failed to live up to expectations of the DX9 standard when it was announced by Microsoft. We are not talking about a exclusive feature made up by one GPU company, we are talking about a standard.

In the end it really didn't matter, since the only game that really did anything of any good for testing shaders was maybe Far Cry 2 . Far Cry however only came out recently in a relative sense compared to when the x800's and 6800's were released. Then soon to be discovered it only had limited shaders.

The real question is, did the FX line hinder progression of game development? Maybe. But that would be highly pessimistic since it always takes a long time for game developers to adopt any new form of technology. Mostly because Publishers take all the risks all ready, and they likely wouldn't let the developers release a game that would only work on a limited number of customers computers. That might of not been the case, but publishers would of not of taken the risk so early. This is why today we are seeing patches upon patches for new shaders and performance enhancements for the new GPU engines.

 

[Gamingphreek]

Nice post regs!

HOwever we all knew that the FX series suxks at PSing. But the question is WHY! The reason for them sucking at AF is because it used an ALU or PS but what is the problem with PSing in general. Just bad architecture or just bad or hard coding?

-Kevin

 

[Regs]

Originally posted by: Gamingphreek
Nice post regs!

HOwever we all knew that the FX series suxks at PSing. But the question is WHY! The reason for them sucking at AF is because it used an ALU or PS but what is the problem with PSing in general. Just bad architecture or just bad or hard coding?

-Kevin

http://www.anandtech.com/video/showdoc.aspx?i=2031

 

[kylebisme]

Originally posted by: BenSkywalker

A good deal of those shots are renders too(just by quickly glancing at them).

... only if you are dillusional.

Originally posted by: BenSkywalkerIt isn't really worth having this discussion with you as you don't know what you are talking about...

... it is like you are talking to a mirror here.

Originally posted by: BenSkywalker(check out exactly how they pull off active camo compared to how MSAA works for starters- this has been explained by numerous sites so you don't need to listen to me about it).

Yet it doesn't prove anything in regards to this conversation no matter how many times you try and make it seem so.

 

[BenSkywalker]

... only if you are dillusional.

OK then, explain how it is they are utilizing a radiosity engine in real time on any current hardware(only in some of the screenshots though)- also explain why it is on some of the screenshots they are showing dynamic soft shadows calculated on a per pixel basis while in other screenshots they are using the old DX7 era poly constructs for shadowing. This should be no problem for you of course.

... it is like you are talking to a mirror here.

Really? So then all that talk that GearBox has had about how they are implementing the active camo by reading back data from the frame buffer(which of course wouldn't work with MSAA due to the oversized framebuffer) is wrong- why don't you explain to GearBox how it is they are actually doing it, since they don't know. I'm sure they would like to hear how there game is working since you know it so much better then they do.

Yet it doesn't prove anything in regards to this conversation no matter how many times you try and make it seem so.

Again, you will have to explain how it is that they are using framebuffer data from a MSAA back buffer to implement a feature being resized back to the native res on a read back. I bet the SC team would be interested in how it is you know how to work around these issues also as they are unable to utilize MSAA in there titles due to comparable technical limitations. I'm sure you can clear this up for them no problem.

 

[kylebisme]

You haven't even proven that MSAA is in use; if you want to blow up a pic and point out the artifacts then feel free to do so, but again that doesn't prove anything. There is no reason Gearbox would have been able to produce MSAA screenshots with the game engine, they could easly have an internal build of the game with a software based MSAA screenshot function similar to the RGSS screenshot function of Doom3. As for real-time radiosity and dynamic soft shadows, which pictures are you reffering to exactly?

 

[Gamingphreek]

Originally posted by: Regs

Originally posted by: Gamingphreek
Nice post regs!

HOwever we all knew that the FX series suxks at PSing. But the question is WHY! The reason for them sucking at AF is because it used an ALU or PS but what is the problem with PSing in general. Just bad architecture or just bad or hard coding?

-Kevin

http://www.anandtech.com/video/showdoc.aspx?i=2031

Well how many math units does the R300 have?

Also what makes ATI so good at all of this? Seems to me ATI uses 24bit over 32bit approximates at AF instead of actual... is it just a bunch of shortcuts? What makes there architecture so good.

-Kevin

 

[BenSkywalker]

You haven't even proven that MSAA is in use;

Anyone who can't tell for themselves I'm not going to waste my time on. You should be able to tell with a fairly quick glance(the smoothed over edges without the lack of clarity on textured/shaded surfaces that is the norm with SSAA).

There is no reason Gearbox would have been able to produce MSAA screenshots with the game engine, they could easly have an internal build of the game with a software based MSAA screenshot function similar to the RGSS screenshot function of Doom3.

Yes, there is. They can't do it as the active camo shader prohibits it- it is utilizing data that needs to be in the correct location in order for the active camo shader to work. If they are reading from an oversized MSAA buffer then the active camo will not work. This is very basic and shouldn't require explenation. If you are saying they are using some sort of software to render the screenshot then you are agreeing with what I said in the first place.

As for real-time radiosity and dynamic soft shadows, which pictures are you reffering to exactly?

It is quite clear which they are(actually, the soft shadows they have may not be dynamic- can't be entirely sure looking at a screenshot but they certainly aren't in game).

 

[kylebisme]

Show me a pic different than the shadows in the game.