View Full Version : nVidia vs ATI, Imagequality: the new battle

2003-03-13, 02:10 PM
I'm sure some of you forum and IRC regulars know that i'm addicted to AA, FA and ATI's texture filtering... Figured i'd make a post that shows the differences between the 2 makers when it comes to imagequality.

Both the GF4 and the Radeon 9500+ series pump out enough FPS, hell in most cases the FPS you get is overkill so whats next to look at... Imagequality.

So enough delays, on whit the pictures.


Note, performance setting is the worst image quality for ATI cards. That shot shows the difference in rendering... No AF in that shot.

Now, on to the AA..


The choice is yours, go whit the good ol nVidia or go whit ATI? nVidia does have a better track record whit drivers, but will ATI's CATALYST series win over the buyers? *shrug* not my call, just figured i would show you the differences.

AA - Antialising

In short: removes jagged edges.

-Explaining this completely would mean explaining the whole mathematical sampling theory. It boils down to this. To recover a signal , or image, you need a minimum of samples to be able to give a realistic representation of the image. The problems start with texture maps being either too close or too far away from the viewpoint. If the polygon is far away you only have a limited number of points to show the texture map, so logically you have to drop a lot of the real pixels of your texture map. This creates some sort of interlace effect : one line is shown and one is not. This can result in weird patterns (moiré patterns) appearing, and makes the texture map look completely different from the real one. A similar problem if the polygon is close to you. You need more info than there is resulting in the generation of random noise (meaningless values). Most of the time the last problem is solved by MIP-Mapping while the other is solved by the anti-aliasing. Another point where anti-aliasing is used is with straight lines, usually abrupt changes in color from pixel to pixel, the border of a polygon for example. If you draw a straight line (under an angle) using a paint program and you zoom in, you will discover that the line looks like a stairway. To remove this and make the line look like a line points in different colors are added to the side of the line to make it look more like a real line. Most of the time it is this kind of Anti-Aliasing they are talking about. All Next Generation hardware support some form of AntiAliasing but only Full Scene AntiAliasing as used by PowerVR Second Generation does not need the CPU for this. The Edge Anti Aliasing techniques used by Voodoo2 and others needs the CPU to prepare the data (sorting, etc...). Results of AntiAliasing are amazing.

AF - Anisotropic filtering:

In short: clear, not smeared textures on all surfaces and angels.

Conventional texture filtering techniques do not compensate for anisotropy--the elongation of the screen pixel when it is mapped into texture space. This results in either blurring or aliasing, depending on the choice of texture level-of-detail. To achieve sharp textures, a card can use anisotropic filtering, a process that involves an elliptical kernel whose shape and orientation depends on the projection of the destination pixel onto the texture map

Note: all pictures and rights are the property of their owners. The explenations are taken (i think those are good and i'm to lazy to write my own hehe) from www.beyond3d.com a site i suggest if your into the more technical sides of things.

2003-03-13, 02:28 PM
"Intel or AMD - who cares?! Nvidia or ATI - who cares?! Windows or Linux - who cares?! It's about the experience. The technology. It's the journey, not the destination. There is no end point. No perfect buy. No perfect company. No perfect chip. Stop looking for your identity in a piece of hardware! " Tom's Hardware

www.tomshardware.com (http://www.tomshardware.com)

2003-03-13, 02:32 PM
Who cares, the guys wanting the best hardware for the best experience.

Thats who.

Pretty pointless reply there hunter83, if you don't have annything better to say in a post then who cares then please don't bother saying it.

2003-03-13, 02:33 PM
No offense to you intended, I just like to post that whenever someone brings up these sorts of topics.


2003-03-13, 08:29 PM
im getting a Radeon 9700 PRO 128MB when the 9800 comes out, since of course the prices will drop :rolleyes:
Lots of ppl in my outfit have GF4 and boast about em, but ehy just cant understand the quality of the Radeons. Hell even PCGamer and GamePro put the Radeon 9700 Pro above the GeForce FX, which is their new card. Its just better

2003-03-13, 10:12 PM
You're numbers are very incomplete.

You don't say which GF4 you are using and you don't say on wht kind of system they are on. You don't say if they asre both 128 meg or if one is 64.

This is also in only one program. Different cards perform differently in different games.

Far be it from me to remind you that the Geforce ti 4400 and the 4600 outperform all of ATI's card when benchmarked in a game like Dungeon Siege.

2003-03-14, 05:52 AM
The real systems they are on and the FPS they use are irrelevant, those shots show rendering only. The first shot was taken whit a GF FX 5800 ULTRA whit the standard drivers that comes whit it on a high end system.

MSI K7N2G nForce2 Motherboard
AMD Athlon XP 2400+
Maxtor 80GB ATA-133 HDD
ATi Radeon 9700 Pro (Catalyst 3.1)
GeForce FX 5800 Ultra (Detonator 42.68s (for 3DMark03) and 42.69s)
512MB (2x256MB) Corsair TwinX PC3200
WindowsXP Pro + SP1

The other 2 shots where taken whit an eqivalent system of the one above.

As for difference in performance in different games, thats because of the rendering differences and qualities of the pictures. btw, i'd like to see a link to that test lex.

2003-03-14, 06:06 AM
Why did you have the Nvidia "balanced" and ATI "Optimal Performance"? and then you have the Nvidia w/o AA and ATI with AA. seems to me like the tests aren't equal. I'm no expert so i may be wrong.... but what about showing the Nvidia with AA and Optimal performance. No offense is intended i'm just trying to learn.

2003-03-14, 06:16 AM
I have not taken the shots or the tests. My post was simply made to show the difference in rendering quality between the GF FX 5800 ULTRA and the Radeon 9700 pro, and ATI vs nVidia in general.

You're reading Optimal performance wrong, optimal performance means lowest posibile image quality. Best FPS, but worst image quality. Balanced for nVidia means a blend of performance and image quality.

The other shots are simply to show what AA really does and how much of an improvement AA is on images/games. GF4 cards can't do AA whitout loosing serious performance while if you look the radeon 9700 pro card does AA and still keeps a better FPS than the GF4 on the same system.

Just a post to compare the image quality of the 2 makers and to show people why some folks, like me bought a radeon 9700 pro over a GF4 card when the performance in most games are the same FPS wise but theres a difference in image quality.

Shots are taken from 3dveliocity and hardocp, both neutral and trustworthy sites.

2003-03-14, 07:02 AM
OK i got ya. Thanx for the indepth explanation :)

2003-03-17, 04:04 PM
This is the same dilemma I had a couple of weeks back. ATI or nVidia?

I was going to get an AMD board, that was a definite. So, naturally you gravitate toward the nForce chipset that is manufactured by ATI's rival nVidia. So, I decided to get the TI 4200 because the last thing I wanted to have was configuration issues. If I see shadows, great. If not, no big deal. Frames per second are better than detail IMHO. No need to flame on, just an opinion.

Z, you da man though :)

2003-03-22, 11:20 AM
well im on a asus nforce 2 board ( A7N8X delux ) with a radion 9700 pro (herculese) and i have no probs at all.

2003-03-22, 07:39 PM
If you want to post a real test use scores from 3Dmarks. If someone happens to have both cards and time, run the test on THE SAME COMPUTER with THE SAME SETTINGS(turn on everything, AF 8xAA the works) and then post the scores, or better yet screen shots of the scores.

Many magazines agree that the ATI outpreforms the GfFX, but id like to see for myself.

2003-03-22, 09:10 PM
Well, the comparisons are all well and good, but it all comes down to this, for me at least: When I'm playing a game/fighting, I'm not paying attention to whether or not I can see the checker design on the back of a plane. I see the plane and that's it. Woopdedoo. Of course, if I'm going to be taking screenshots, I will adjust my playing for that.

I mean, it doesn't really matter as long as you can see what's going on.

2003-03-22, 09:42 PM
cant agree more stryker but in some games it helps to have a card that can do that little extra be, makes the whole experiance of playing a game that much better.

hellsfire not sure about 3dmark as its a sore point with both camps and to be honest its poor as a gaming benchmark, if you want to test a gfx card use ut2k3 at least if you test with this theres less chance of special sets of drivers for the game to make it run that little bit quicker, as there is for 3dmark (both gfx card makers are guilty of this one).

but to be honest i have a radion 9700 pro and i was waiting for a fx but my current card a gef 3 ti500 died so i had to upgrade to what i could get my hands on at the time so it was the 9700 and im pleased with it.

2003-03-23, 03:40 AM
This isn't a raw performance example, cause honnstly todays CPU'd can't keep whit the vidoe cards. It's an example to see the difference between the rendering quality of nVidia vs ATI and an example of AA.

As for the 9700 outperforming the FX, it outperforms the FX at higher resolutions whit AA and FA enabeled. Don't really need tests for that, just need to consider the bandwith differences between the cards. FX has roughly 16 gb/sec while the 9700 has 19.2 gb/sec.

Annyways, if you really want numbers: http://www.tomshardware.com/graphic/20030311/index.html

note the first shot is taken on the same system, just swapped the cards.