View Full Version : 3Dmark05 High and better IQ...
jbirney
10-04-04, 05:50 PM
Remember.. the PRo version is the one FUturemark wants people to buy for professional benchmarking ,for more freedom and more settings for comparison.so this is a moot point already.
How many people are using the ORB using the free version vrs the pro version with the switch turn off? Which do you think a majority of the scores will be? Its far from moot.
here ... (http://www.beyond3d.com/forum/viewtopic.php?t=16710&postdays=0&postorder=asc&start=0&sid=f681acbfede7846f607fb083375006a5) . there is a great misconception about that , thanks to the apples vs oranges comparison of ATI marketing papers. 3Dc and DXt5 have the same advantages .is just that 3dc have slightly more precision. its not an all around solution and if it is not wisely used the qualiy can be lower.think about that before asking that feature to be used in your games..is not something that every game Developer will want to use for many reasons explained there.
Your missing the point. Its not if/should/why 3Dc was left out. Or if its quality is ok or hard to use (in fact it can be shown that DST is producing a lower IQ shadow). Its just a comparison to say you allowed a non-DX9 standard for one vendor, why not do it for the other? After all isn't that better than to start out all IHVs of only one company with a small head start?
ChrisRay
10-04-04, 06:24 PM
Alright People, I dont want to see any more flaming, I am carefully pruning this thread, If I see anymore flaming content from people. I am just going to remove it, We're reaching the point where we're gonna stop deleting posts and take more further action.
You can make a post with a clear concise point w/o flaming someone else.
Chris
Chris, if ATi is rendering thier alphas different, since fog is done with layers of textures, would we be seeing at least some fog in GT3, since it would have an additive feature, with the screenshot of the captin, over at B3d ogl guy stated the fog will be less visiable because of the layers, that just isn't logiacl, maybe I read it wrong.
Since the opengl guy said it wasn't a bug and that it was because they were calculating things slightly differently than the refrast, is this like the Humus tweak for Doom3? He substituted an alternative method at the expense of a slight difference in IQ (some things just didn't seem quite a bright, at least initially, since I'm not sure how it all turned out at the very end). Is the fog in 3DMark05 just bringing to light the same sort of thing?
I guess I just don't understand why they don't just calculate it the same way as the refrast or at least use a method that gets them a little closer to the reference image (especially since they seem to know about it).
Edit: Let me add a note just to be a little more clear. I realize no one matches the reference image. What I'm getting at is that ATI seems to know they don't match it because they're calculating things differently. So my point is, why don't they choose an alternative method of calculating it that gets a little closer to the reference image (right now the difference is too large IMO).
jbirney
10-05-04, 09:36 AM
OWA,
last I hear it could be hardware issue...
Your missing the point. Its not if/should/why 3Dc was left out. Or if its quality is ok or hard to use (in fact it can be shown that DST is producing a lower IQ shadow). Its just a comparison to say you allowed a non-DX9 standard for one vendor, why not do it for the other?
The funny thing is that 3Dc is supported in Directx 9.0 using a Four CC code, Nv could implement it as well (but Nv40 doesn't support it). Ati cannot use DST as this (non Dx9) feature has a Nv patent.
Can someone provides Nv screenshots of the same frames as the Ati screens I'm showing?
It seems that lightning appears on the boat/ropes as soon as the flame comes out of the border of the canon. The Nv shot don't show this, but maybe later?
http://members.chello.nl/h.appel1/11.jpg
http://members.chello.nl/h.appel1/22.jpg
http://members.chello.nl/h.appel1/33.jpg
Nv:
http://members.chello.nl/h.appel1/Nv.jpg
ChrisRay
10-05-04, 04:55 PM
My most apparent question is, Have you tried it without DST? Testing now.
Its rendering fine with the 66.70 drivers as far as I can tell. I mean seeing the lighting difference between the two, The light shows up fine on the back.
Having the a .PNG uploaded right now to nvnews server, will host a link a sec.
ChrisRay
10-05-04, 05:10 PM
Ok 66.70 Drivers, With Geforce 6800,
http://www.nvnews.net/screenshots/3dmark05/6800canyon2.PNG
The Lighting is clearly being renderered correctly, and these are with the fastest Nvidia drivers.
Ok 66.70 Drivers, With Geforce 6800,
http://www.nvnews.net/screenshots/3dmark05/6800canyon2.PNG
The Lighting is clearly being renderered correctly, and these are with the fastest Nvidia drivers.
thanks for the shot chris, i guess the dst is what's making it not rendering right
66.70 is a non approved Futuremark driver...
And what when using an approved driver/no DST? :rolleyes:
(and please can you provide an 4xAA/8xAF shot with best quality settings in your Nv CP? Thnx)
ChrisRay
10-05-04, 06:02 PM
thanks for the shot chris, i guess the dst is what's making it not rendering right
Nope, It renders correctly with DST and without, That shots look the same regardless of DST being enabled or disabled.
66.70 is a non approved Futuremark driver...
And what when using an approved driver/no DST? :rolleyes:
(and please can you provide an 4xAA/8xAF shot with best quality settings in your Nv CP? Thnx)
And what when using an approved driver/no DST? :rolleyes:
I already explained it renders correctly with DST or Without,
(and please can you provide an 4xAA/8xAF shot with best quality settings in your Nv CP? Thnx)
No, I dont have the time. And frankly with your attitude, I have no reason to want to oblige you either.
jimmyjames123
10-05-04, 06:32 PM
The funny thing is that 3Dc is supported in Directx 9.0 using a Four CC code, Nv could implement it as well (but Nv40 doesn't support it).
ATI is the only current hardware that supports 3dc, period.
Ati cannot use DST as this (non Dx9) feature has a Nv patent.
NV does not hold the patent for DST, SGI holds the patent. Also, DST is not an NV-exclusive feature. On top of this, it has been said that DST should have theoretically better image quality at times than without DST.
66.70 is a non approved Futuremark driver...
And what when using an approved driver/no DST? :rolleyes:
(and please can you provide an 4xAA/8xAF shot with best quality settings in your Nv CP? Thnx)
Futuremark released a faqs today that might answer some of your questions in regards to DST since you have issue's with nvidia instead of knowing where and why DST is being used in 3dmark05 :rolleyes: , here's the link http://www.futuremark.com/companyinfo/Technical_FAQ_3DMark05.pdf
and quit embarrassing yourself :lame:
Diamond_G
10-05-04, 07:56 PM
I'm using 65.73's and the lighting isn't rendered in those shot on my system either... hmm wonder if it is a driver issue, quality settings used.
I just left everything at the defaults (driver-wise and 3DMark05). These are with the 66.29s which are supposed to be FM approved (I believe).
1829..........................1830................ .......1831
http://img80.exs.cx/img80/8724/nv_6629_gt3_1829.th.jpg (http://img80.exs.cx/my.php?loc=img80&image=nv_6629_gt3_1829.jpg) http://img43.exs.cx/img43/4528/nv_6629_gt3_1830.th.jpg (http://img43.exs.cx/my.php?loc=img43&image=nv_6629_gt3_1830.jpg) http://img43.exs.cx/img43/19/nv_6629_gt3_1831.th.jpg (http://img43.exs.cx/my.php?loc=img43&image=nv_6629_gt3_1831.jpg)
Here is 1831 with no DST. It just looks like the shadows are a little softer to me but pretty much the same otherwise (I think).
nvidia no dst..................ref image.................ati 8.07
http://img81.exs.cx/img81/4352/nv_6629_no_DST_gt3_1831.th.jpg (http://img81.exs.cx/my.php?loc=img81&image=nv_6629_no_DST_gt3_1831.jpg) http://img81.exs.cx/img81/3205/refimage_gt3_1831.th.jpg (http://img81.exs.cx/my.php?loc=img81&image=refimage_gt3_1831.jpg) http://img56.exs.cx/img56/2043/ati_807_gt3_1831.th.jpg (http://img56.exs.cx/my.php?loc=img56&image=ati_807_gt3_1831.jpg)
ChrisRay
10-05-04, 08:26 PM
I'm using 65.73's and the lighting isn't rendered in those shot on my system either... hmm wonder if it is a driver issue, quality settings used.
It would seem those drivers have an issue, But newer drivers it appears to be gone,
MustangSVT
10-05-04, 08:57 PM
I think this whole "IQ Issue" is getting pretty heated up. I think the DST offers better IQ, and that there's the less filtering in that one image (can't remember which one), but as the ATI ppl said, it's fixed through filtering from the CP (or whatever it's called), but then some ATI ppl say the way ATI does it is "correct", and then we the NV ppl say the DST looks better (which it does imho, but i probably WOULDNT notice it). Overall, I think everyone should just forget bout 3dmark2005. it's a stupid benchmark. it's not a game. When I see a stock x800pro beating stock 6800gt in like 90% of the games, i'll believe it, but until then, it's just a benchmark, get over it.. :wong:
gokickrocks
10-05-04, 10:05 PM
dont know if its been stated yet, but going back to the issue regarding the screenshot with the moire pattern...ati is rendering it the same as the reference rasterizer (i do have the screenshot, but no place to host it), which also has the moire pattern and the other box that was highlighted by ruined...you guessed it, ati also closely resembles it...many, if not all will probably say that they prefer it without the moire, and i agree, but then that wasnt the point of this discussion which was whether or not ati was degrading IQ in order to get higher scores
the fog situation seems to be the same as the aquamark incident: ati does "blending math" differently and when you start layering it you can see a visible difference...but its still completely rendered in a software sense
at this point we have learned that ...
- Nvidia use DST shadows as default in the free version ,but that can be disable in the professional benchmarking version. (which one looks better is just a matter of preference)
-ATI shows diferent IQ/ performance behaviour .. whether is AA or AF depending if you use in game settings or control panel. something similar to the Doom3 bug.
- Is pretty much clear that ATI is not rendering all the Dx9 effects.the missing FOg is pretty much confirmed and its very obvious the IQ diferences with or without the FOg .seems like an ATI hardware issue. and looks to be an effect heavily used in more than one game test. the only thing we need to know is in what way this has an impact in performance for the hardware that do it correctly. it can be minimal but also it can be a significant one. specially if they use many samples of high quality mist there ,all depends of what the developers wanted to do there.
gordon151
10-05-04, 11:00 PM
at this point we have learned that ...
- Nvidia use DST shadows as default in the free version ,but that can be disable in the professional benchmarking version. (which one looks better is just a matter of preference)
-ATI shows diferent IQ/ performance behaviour .. whether is AA or AF depending if you use in game settings or control panel. something similar to the Doom3 bug.
- is clear that ATI is not rendering all the effects .the missing FOg is pretty much confirmed. seems like an ATI hardware issue. and it is and effect heavily used in more than one game test. the only thing we need to know is in what way this has an impact in performance for the hardware that do it correctly. it can be minimal but also can be a significant one. specially if they use many samples of high quality mist there ,it depends of what the developers wanted to do there.
Loving friend NV40, the cards is rendering the fog. The issue is that it is barely visible. This is the same thing that raised flags (probably by you and a lot of other people on this forum) with Aquamark.
No Name calling Gordon. Be civil to each other please .:)
I said be civil Gordon not stupid.
Nubcakes, the cards is rendering the fog. The issue is that it is barely visible. This is the same thing that raised a flags (probably by you and a lot of other people on this forum) with Aquamark.
there is no point for Nvidia to render the FOg in the benchmark,if for ATI is "barely visible" to the point to be not there. screenshots comparisons shows that the theory of the "invisible FOg" " is another excuse of the many we already have heard in this generation from them. they should simply say that it is a hardware limitation period. that they can't do it properly. and not say that they dont have problems with it ,that just because people can't see it doesnt means is not there . :rolleyes: fortunately this is only a benchmark ,not a game ..but i wonder if the same issues will be common for ATI hardware with next generation DIrectx9 games ,if those use similar eye candy effects that the ones used in 3dmark.
cthellis
10-06-04, 12:36 AM
I'm a little late to this one, so I may be asking something already answered before, but...
...has it already been established that it's not a gamma issue? I remember a similar thing happening with Aquamark 3 a while back (and had Tom and others jumping up and down on it), even though with a simple Photoshop tweak you could see that all the fog was THERE... it was just too transparent.
vBulletin® v3.7.1, Copyright ©2000-2010, Jelsoft Enterprises Ltd.