Sign in with
Sign up | Sign in
Your question

Nvidia cheating again? with 16 bit precision...

Last response: in Graphics & Displays
Share
July 18, 2006 3:43:33 PM

Here's a shot of two images from Rydermark:

ATI



Nvidia

July 18, 2006 3:47:57 PM

ATI's looks nicer, especially on the water. Nicer shadows, and quality.

Interesting.
July 18, 2006 4:05:42 PM

Quote:
ATI's looks nicer, especially on the water. Nicer shadows, and quality.

Interesting.




Really quite interesting, as I see the Nvidia pic being nicer, just look at the boat; its lines are smoother and less jagged than the ATI pic.

hball
Related resources
July 18, 2006 4:07:57 PM

I thought the same thing, the water up close may look different, but I wouldnt say the ATI pic is especially better. Things in the distance look better in the nvidia pic to me, especially the boat and the waves coming off it.
July 18, 2006 4:22:41 PM

and the ati card is leaving funny shadows all over the place, looks kind of silly.
July 18, 2006 4:33:41 PM

The ATI one looks photoshoped to me...



In the box I have drawn on each picture you can see where on the ATI photo there is a discrete line. On the left side is how the water looks on the rest of the ATI photo. On the right side, the water matches the water on the nvidia photo.

I'm not a graphics or photoshop pro by any means but that is what it looks like to me.
a b U Graphics card
July 18, 2006 4:51:06 PM

Its weird, in some parts the nvidia one looks better in others the ati, hard to tell which is best.
July 18, 2006 5:09:11 PM

The only difference I can see is the shade of blue on the water, neither looks actually lower in color depth but the ATI one appears to be much lighter. If anything I'd agree that the Nvidia one actually looks better due thanks to the water's darker color and it blending better with the edges.
July 18, 2006 5:20:44 PM

Quote:
The ATI one looks photoshoped to me...

I agree. Flipping back and forth between the images, the ATI one distinctly looks like it has been sharpened in Photoshop in the region around the boat and its wake.
Looks like the poster is cheating if you ask me.
I don't doubt that there may be a visible difference due to the 16bit precision, I just don't think that this pic is the proof.

-mcg
July 18, 2006 5:27:21 PM

Hehe, thats pretty funny, but its so horribly photoshopped, im not sure which side the inquirer has decided to take on that particular article, but theyre definately trying to push something.
You can see for yourself if you switch between the two pictures fairly quickly, you will notice that there is a section that looks like someone took the sharpen tool to it, its really quite easy to see and could have been far better done. -_-
July 18, 2006 5:41:06 PM

Quote:
Hehe, thats pretty funny, but its so horribly photoshopped, im not sure which side the inquirer has decided to take on that particular article, but theyre definately trying to push something.
You can see for yourself if you switch between the two pictures fairly quickly, you will notice that there is a section that looks like someone took the sharpen tool to it, its really quite easy to see and could have been far better done. -_-


Copycat! :D 
July 18, 2006 5:58:00 PM

I have to agree too. I've been using Photoshop for a while, and the waves in the ATI capture look like sharpened. It's really noticeable.
And there's that line to the left of the boat, on the right edge of the image...
Really suspicious... :?
July 18, 2006 6:27:30 PM

Quote:
I have to agree too. I've been using Photoshop for a while, and the waves in the ATI capture look like sharpened. It's really noticeable.
And there's that line to the left of the boat, on the right edge of the image...
Really suspicious... :?


I also noticed a 'line' that cuts UP out of the water.

Look off from the boat at approx. 10 o' clock.... part of the fence where a person is standing behind it is lightened and sharpened along with the water.

I call Photoshopped,.. and on top of that, must've used a mouse.
July 18, 2006 7:26:59 PM

That was my first impression when i saw this on the inquirer website.
WTF!?! nvidia's looks better.

Then if you switch between the two quickly in the same spot you'll realize the water just gets brighter in the ATI pick.

Then there are those obvious cropping lines some of you have highlighted.


I think we just caught the Inquirer spreading FUD once again.

I have work to do right now, but i wasn't busy i'd run a difference filter on those two pictures and see what comes up..
should be a big jagged box over the water if its a photoshop job. :wink:
a b U Graphics card
a b Î Nvidia
July 19, 2006 12:13:05 AM

Quote:
ATI's looks nicer, especially on the water. Nicer shadows, and quality.

Interesting.


Well at least you noticed what WAS different. BRAVO!

Whereas alot of other people here seem to making up differences in order to convince themselves of something.
Seriously people, it's a benchmark rendered on a predetermined path with the image taken at the same point, do you not know that there's image apps out there that can analyze the diff? ML obviously doesn, and I bet the InQ does too. Which is why the accusations of Fanbois against the InQ is faQin' hillarious.

ML, I'll save you the time, if you haven't already done it yet to your own embarassment. Here's the pixel per pixel differences;


Now show me those areas of difference outside the water again, like the thing in the distance or was that distant water?

Now obviously thee's a difference, the question for the author's or Rydermark is what the refrast, and who's the one not measuring up.

So let me understand the counter argument that people are trying to make here. IT doesn't mater that someone may be rendering this incorrectly you prefer the partial precision (kinda like vasoline on a lense, eh? :roll: ) so therefore it's ok, or damn the InQ for brinding it to anyone's attention.

Maybe I'm mis-rendering you argument due to your lack of precision people, set me straight, explain your position cause I'm missing your points obviously. Sounds like the typical mentality of the ends justifying the means. I guess we've got a bunch of BarryBonds-BenJohnson suporters here, or maybe I need some of what you guys are taking.
July 19, 2006 3:14:12 PM

LOL.

Great post, Ape. :) 
a c 102 U Graphics card
a b Î Nvidia
July 19, 2006 3:48:30 PM

LOL alt tab from pic to pic.... Welcome photoshop :) 
July 19, 2006 4:43:56 PM

nvidia seems anti-aliased to me. It could be some difference on the AA algorithm for instance.
nvidia looks better to me. The sharp water area and blurred landscape just dont seem to fit in the ati scene.
a b U Graphics card
July 19, 2006 5:17:11 PM

Quote:
LOL.

Great post, Ape. :) 

My thoughts exactly. :lol: 
July 19, 2006 5:36:09 PM

Wow, that has got to be the worst lasso job I've ever seen in Photoshop. If they were going to cheat this blatantly, they should've at least taken the extra 5 minutes to use the polygonal lasso or quickmask and done it right.
July 19, 2006 9:56:45 PM

Hi!
I've tried "high quality shaders" and I've got this artifacts on screen:




How can I fix this problem, please?
July 19, 2006 10:16:47 PM

lol @ mordoreye

nvidias looks better to me. Its obviously photoshoped thou cause the waves near the boat look pixely.
July 19, 2006 10:22:17 PM

rofl - that is such a lovely hat!
July 19, 2006 10:45:09 PM

very nice photoshopping... much better than the inquirer's.
July 19, 2006 11:10:21 PM

If you brighten that picture of yours you can clearly see the cropping marks around the water.



Furthermore i went ahead and applied the sharpen more filter on the nvidia screen shot... and well anything look familiar?



It seems that the ATI card is equipped with a photoshop sharpen filter that improves quality.
You heard it first at the Inquirer. :lol: 
July 20, 2006 12:02:07 AM

I think the Ghostbusters is who you gotta call to get rid of those artifacts. Seems like too much danger for just that one little boat to combat all on his own...
a b U Graphics card
a b Î Nvidia
July 20, 2006 4:26:59 AM

Quote:
If you brighten that picture of yours you can clearly see the cropping marks around the water.


Could be but even with a negative with log filter and edge enhance and then overlaying it again over the original, it does match areas where the waves crest. The only area that causes an issue for me is the area at the water's edge where the girl is;



Quote:
Furthermore i went ahead and applied the sharpen more filter on the nvidia screen shot... and well anything look familiar?


Only the exact same thing as when I appy either the Gaussian blur or on the ATi image, look familiar to you too?;



Quote:
It seems that the ATI card is equipped with a photoshop sharpen filter that improves quality.



OR nV has added a blur filter to their render path. :tongue:
Personally I'd rather have sharpen than blur, but more importantly I'd want accurate rather than not, regardless of who does it.

Quote:
You heard it first at the Inquirer. :lol: 


Nah, I think everyone has said that ATi improve ImageQuality. :twisted:

The thing is, regardless of which it is Blur or Sharpen, like I said it's people's reactions in this thread that are hillarious. I reserved comment, like I usually do on these issues until more concrete info comes out or I see blatant BSing. The later is what drew my attention yesterday, because your post started out with the basics relaying what is a questionable story, but being somewhat even handed about it. Then things turned into a 'ends justify the means thread' and that's just ridiculous, especially forthe made up stuff.

I still stick with the idea that until we get the response from the developer, addressing the anomalies (there are many needless to say, but many span both examples), and to me most importantly, until we see an image using the refrast tool and then it should be obvious who isn't rendering the image correctly. Now one would assume with this whole stink that that's been done, but that is giving in to the idea that this is a real problem and not a hoax. But it's still the only way to find out the answer (other than admission of a hoax if that is the case).

And personally I'd say it's unlikely Fuad/theInQ are perpetrating the hoax (they've got too much to lose IMO), but they might be perpetuating it and the unwitting but willing victims of it by being drawn in by it's originators.
I'd say the InQ is a 50/50 new site reporting whatever they hear on the grapevine fact/fiction. But they don't have much to gain from this, and a bunch to lose IMO.

None of the possible scenarios would be anything unusual for this industry really, and I guess the reaction of some people in this thread isn't anything new either. :roll:

Well as usual, only time will tell.
July 20, 2006 4:53:45 AM

Has anyone actually pointed out that these are the exact same picture.

EXCEPT, someone took a lasso marquee tool to the "ATI pic" and used a sharpen filter on it.

He wasn't editing his ATI screenshot, he was editing his "nVIDIA screenie" to create an "ATI screenie".

This is like so ridiculous lol.
July 20, 2006 5:17:41 AM

That's exactly what numerous people have said, including myself.
You'd known that if you had actually read through their responses
a b U Graphics card
a b Î Nvidia
July 20, 2006 5:26:21 AM

Quote:
Has anyone actually pointed out that these are the exact same picture.


Uh, yeah. Read the thread. :roll:

(Edit, LOL, ML beat me to it :tongue: )

This has been discussed on about 20 different sites all claiming they know for sure one way or the other, and the InQ replying in kind;

http://www.theinquirer.net/default.aspx?article=33131

And then them updating their original statement, etc.
http://www.dailytech.com/article.aspx?newsid=3400

B3D has been discussing this long before it hit here, but really people at B3D wouldn't be what I call an unbiased critic for the InQ. :twisted:

While it looks like something, like I said before, there's no way to know for sure without the original. Because despite the statement byt dailytech that the nV is the original, there's appears to be a similar outline with the nV picture (compare the image using all 3 (a / diff / nv) and the question if it were photochop is which way did they go blur or sharpen?

The girl I posted before is my only indicator it may be dailytech's version if anything.

Quote:
EXCEPT, someone took a lasso marquee tool to the "ATI pic" and used a sharpen filter on it.


Actually that would be use a sharpen filter on the "nV pic" or a Gaussian blur filter on the "ATi pic" to get the resultant image of the competitor.

It's worthy of note that Dailytech points out that if a fake it looks like it wasn't the author/Fuad.

Like I said, eventually all will be revealed. :wink:
July 20, 2006 5:38:31 AM

Quote:
That's exactly what numerous people have said, including myself.
You'd known that if you had actually read through their responses


I haven't actually seen anyone say directly that they are in fact the same picture, and one generated from the other. there are numerous references to "The ATI one has been edited..." blah blah and so forth. Maybe the rest of my post about the lasso and filter were redundant, but I believed at the time that no one pointed out what i said in my first statement.

And if it is in fact what everyone else has said, you are free to provide quotations.

peace nig.
July 20, 2006 5:48:36 AM

Ape i think its pretty obvious that the original is the NV screen.
If you zoom in on the bottom right of the pic the marquee or crop lines, whatever you want to call them are clrealy visible.

Also way in the back you can see where the crop lines cross from the water onto the boardwalk... (the picture of the girl by the green door you posted)

The gausian blur doesn't eliminate those crop lines.
In order to get the ATI pic to look similar to the NV would require a lot of work, mainly pinpoint accuracy.
Whoever photochoped that picture was a careless amatuer... and it shows.
a b U Graphics card
a b Î Nvidia
July 20, 2006 5:49:24 AM

Quote:

I haven't actually seen anyone say directly that they are in fact the same picture, and one generated from the other.


Actually it wasn't as relevant since the picture data would be dependant on post image creation and how it was saved.

That the images are the same except for that would be the same as if they were rendering the exact same scene and then adding something to different captures of the same scene.

So like so much else, its importance is limited without knowing for sure the source/original.
a b U Graphics card
a b Î Nvidia
July 20, 2006 6:17:12 AM

Quote:
Ape i think its pretty obvious that the original is the NV screen.
If you zoom in on the bottom right of the pic the marquee or crop lines, whatever you want to call them are clrealy visible.


And you see, I see it on both images especially enhancing with a sin or log filter. To me it's like BOTH were altered only ATi's seems more noticeable, and here't the thing depending on the way it was made, rendered or chopped it could be both. BTW, are you looking on a CRT or LCD, it's more noticeable for the nV (by more I mean as in more than zero) when I look on my CRT, on my LCD it's almost no diff. But.....

Quote:
Also way in the back you can see where the crop lines cross from the water onto the boardwalk... (the picture of the girl by the green door you posted)


And yeah like I said that's the only one that truely bothers me. Both show what could aguably be artifacts, but that's the one that stands out (especially in the difference calc) as being truely out of the range of expected and seems like an artifact. And the only way to get that artifact in using our two techniques would be for the sharpening (which still doesn't preclude some blur in the other picture, only that the sharpening actually caused a terribly noticeable issue). It's the thing that makes it hard to consider a render error as much as a post processing error, but anything is possible. I also just wouldn't expect something so sloppy if you're going to put it infront of a wide audience like that. Heck I made more effort in my chopped 3Dmark06 pre-release gag image to ensure proper blends. :roll:

Quote:
The gausian blur doesn't eliminate those crop lines.
In order to get the ATI pic to look similar to the NV would require a lot of work, mainly pinpoint accuracy.


Well really not that much pinpoint accuracy, it's kinda easy to assist that, however if it is a messy chop why bother with doing that much. And while it wouldn't eliminate those crop lines it would mute them alot which is what I also see at the same time (maybe just tired :mrgreen: ). Admittedly it looks more dramatic on the ATi image. Now while the water areas could be easily explained as definited render regions or something, that area by the girl doesn't fit that argument. Now the thing that was weird from the start is the image textre stitching, but really it tough to know what's going on with something we can't get out hands on, are there static regions (even pre-rendered) in the benchmark that could explain some boarders?
The most annoying thing is not being able to get a pre-release copy of the benchmark which would help reseolve alot of this. Sofar it's alot of talk of what could only be called 'limited' evidence.

Quote:
Whoever photochoped that picture was a careless amatuer... and it shows.


Yep, although like I said, I'm still not going to point fingers yet, but I will comment on the 'what if , what about' aspect, because it is interesting.

I will say though I'd hate to be in Fuad's position right now, as he seems to be relying on the word or the makers or Rydermark, and yet he and the InQ are the ones risking the most (nV's likely looking at this alongside their lawyers 8O ).
July 20, 2006 7:30:49 AM

Sad that ATI supporters sink so low. Even if the pics had not be retouched there would be varying opinions on which was better. Some people like chocolate some people like vannila. Myself I prefer Nvidia, the hardware is so close in abilities that it comes down to drivers. ATI has always dropped the ball there, Nvidia's drivers have been more stable, easier to find and they work. Now from personal experience and reading in numerous forums it's very difficult for me to call anyone at ATI an engineer or programmer when it comes to drivers and applications. Careful about spraying Raid or anything like it to near your comp, it'll kill the buggy ATI software and you're hardware won't work :) 
a b U Graphics card
a b Î Nvidia
July 20, 2006 1:10:28 PM

Quote:
Myself I prefer Nvidia, the hardware is so close in abilities that it comes down to drivers.


Not really more comes down to apps than drivers.

Quote:
ATI has always dropped the ball there, Nvidia's drivers have been more stable,


Really? Guess you didn't download the 90 series when it first launched, and I guess you don't use their betas either (so 3-4 updates a year for you?)

Quote:
easier to find


Really? Yeah finding them at the same place every time seems difficult for both. :roll:

In fact the only difficulty I've read recently in finding either was an nV one, which seemed more of a regional issue;
http://forumz.tomshardware.com/hardware/Strange-happeni...

Quote:
and they work.


And so do ATi's. Both get equal praise. The variance in M$ is small and the variance in linux is getting smaller and I'd say each has an advantage in the other. And as for bugs, as I've posted before, they both have their share;
http://www.digit-life.com/articles2/digest3d/0606/itogi...

Quote:
Now from personal experience and reading in numerous forums it's very difficult for me to call anyone at ATI an engineer or programmer when it comes to drivers and applications.


Well really, who cares?

Quote:
Careful about spraying Raid or anything like it to near your comp, it'll kill the buggy ATI software and you're hardware won't work :) 


Yeah living in the past, wanna go to the past bud, remember the last cheating scandal with nV. So if you wanna argue nothing's changed, that's up to you, but it's a pretty ignorant position that's not a representation of the current reality where they are virtually indistinguishable nowadays, even the long standing linux differences are begining to fade. But hey if you're still installing your drivers off the CD that came with your ancient hardware, then I guess you wouldn't know the changes would you. :roll:
July 20, 2006 4:17:11 PM

Quote:
ATI has always dropped the ball there, Nvidia's drivers have been more stable, easier to find and they work.


You mean, sad that fanboys have to live in the past to keep the brand preference justification alive. :roll:

Ive said this before and I'll say it again; I review graphics cards professionally for Tom's, and for WindowsXP gaming ATi & Nvidia's drivers are equally great. Both have a few quirks of their own, but they're all excellent gaming drivers.

If you want to talk professional 3d apps or linux, Nvidia has a definite advantage there, but that's hardly relevant to 99.9% of people who want a graphics card for gaming...

...but that still doesn't stop it from being regurgitated by every nvidiot who wants to diss Ati.

Have you even used Ati drivers since the 8500? It's not 1999 anymore fellas, try to get with the program. :p 
July 22, 2006 12:31:05 PM

I've seen some evidences in the water:




Red halos: Evidences of clone stamp.

Yellow halos: References of repeated waves.

White lines: The end of the images.

Wacom at the boat: Evidence of handpainted lines, similar that the ones that I make with my wacom tablet.

(Sorry for my bad english).


This is not a 3D image, this is a 2D colage!!

Evidences that I took from other forums:

http://www.cse.iitb.ac.in/%7Echinmay/VeniceTrip/IMG_092...
http://img485.imageshack.us/img485/1332/tower3webxv4.jp...

http://www.cse.iitb.ac.in/~chinmay/VeniceTrip/IMG_0919....
http://home.no.net/larsbw/images/fake_small.jpg

(Beyond3D)
July 22, 2006 4:07:47 PM

OMFG! 8O 8O
That's astonishing!

I should have noticed the blue outline of the tower before... Now I see it, and it's SO obvious... 8O

Shame on The Inquirer! (I think they're used to it :lol:  )
SOOOO LAMEEEE.... :evil: 
July 22, 2006 10:36:27 PM

Ok, lets assume nothing was altered.

The ATI pick looks sharper, but the edges look jagged.

The Nvidia looks like the water blends into the rest of the picture better.

So, regardless of whether or not the photos were altered I like the Nvidia pic better. In more contrasting terms, who cares if it is sharper if it looks like crap?
a b U Graphics card
a b Î Nvidia
July 23, 2006 2:48:24 PM

Still waiting for more from the developers or the InQ or even nV for that matter (considering their level of lawyer brawn that they've never feared to flex before with even partial cause), the level of silence is getting a little suspicious all around. Also the fact that there is so little supporting evidence for the 'forced 16 bit unless enabled' from other developers is going to require something a little more. Pretty much from the start I'd say it required something a little more, as does a true debunking. So far there' alot of stuff that looks out of the ordinary, but really it's hard to debunk something no one has seen other than these two static images of a dubious nature. But just one thing to consider...

Quote:

This is not a 3D image, this is a 2D colage!!


Those are for sure, but the question is whether that was a rendered segment , if you think about static backgrouds you might only have a segment of a scene that actually involves 3D, so it's still possible to have those background and side images be 2D composites if the 'camera' is in a fixed position, and the only thing that's rendering realt-time in 3D is the boat and waves as it proceeds through the scene.

Anywhoo, still waiting for something more than what amounts to a rumour.
Not looking good for Fuad and the InQ at this point.

The interesting thing from EB is the potential that these guys are behind Rydermark;

http://www.candellasoftware.com/services_demo.htm
"Using our own in-house DirectX 9-based Sahara game engine"

Based on the boat picutre in this link (check out the silver car for some video)
http://www.candellasoftware.com/artwork.htm

Also notice the engine in the original introduction of the benchmark and the one they use ;
It will use heavily multithreaded Sahara game engine with multiprocessor support.
a b U Graphics card
a b Î Nvidia
July 23, 2006 7:33:46 PM

Yeah, those have been there since day one of the thread, look at the front page.

They aren't really in dispute, it's just a question of where they come from, and why they are there, other than the obvious possible answer of hoax.

Still waiting on something more concrete than our post processing analysis.
July 25, 2006 2:00:50 AM

i reckon ur all f****ng mad

i switched between both pictures constantly and couldnt detect any difference with my human eye...

if there IS a difference well fark man... is it really that visible?

i think higher FPS is a lot more important
to think people would look at these pictures and think the ATi one is better in quality and t hen go out and buy an ATI card that produces lesser FPS is ridiculous... image quality? pfft, there is no difference in my opinion

---------
edit
---------
okay i looked closer yeh i see the difference...
I think the ATi one was photoshopped cos it looks bogus... besdies the nvidia one looks nicer anyway beacuse the ATi one looks 'artificial' but the nvidia one looks more 'real'...
!