Anyone able to talk to J. Camrack and get a response? I g...

[Socos]

I have not seen the question asked to the god of games, Mr. Camrack what he thinks of Image quality from the two top end cards? I know he must of been using a NV30 since it tapped out, and hopefully he has also tried the 9700. He must have an opinion.

I would like to know what he thinks of their AA methods, as well as overall IQ?

Someone please ask him!!!

 

[binmaze]

Well..

As for !AA IQ, they'd be almost the same, imo.
Even though it seems NV30 use FP16 and 9700p use FP24, virtually none gonna really be able to discern the differences...

And I don't think we need JC's opinion on AA qualities. It's too obvious.

 

[KimB]

Yes, AA quality is obvious. The 9700 Pro has better AA.

 

[JF_Aidan_Pryde]

Yes, AA quality is obvious. The 9700 Pro has better AA.

 

[Pete]

He spoke about IQ either in his last .plan or in Rev's Q&A session. Check 'em out.

 

[epicstruggle]

Im going to say something that might make me look very stupid. Isnt image quality getting to be more and more subjective. I looked at the comparison images of anandtechs fx review. and could only tell a difference when they blew up the image way up.(statement based on having AA 4x+, and AF on)

later,

 

[binmaze]

Im going to say something that might make me look very stupid. Isnt image quality getting to be more and more subjective. I looked at the comparison images of anandtechs fx review. and could only tell a difference when they blew up the image way up.(statement based on having AA 4x+, and AF on)

later,

You must be able to find those jaggies on GFFX screen shots.
But then again, it's not easy to discern the differences while you're in the heat of gaming...

Therefore, if you don't notice that while gaming, it's all ok for you. If you do, it'll annoy you.

It's nvidia hype about GFFX supreme IQ which turned out to be worse than 9700pro that really puts GFFX in awkward situation about these IQ things.

 

[THe_KELRaTH]

If you take a look at the two UT2003 pics on HardOCP where both AA/AF is on it's easy to notice that the GFFX lowers the AA level quite dramatically as to be quite noticeable at the horizontal offsets.
You will also notice the effect of having Gamma correction.

NV GFFX Ultra 8xS AA x8AF
http://www.hardocp.com/image.html?image=MTA0MzYyMDg1OTVjVVNkMzFISXhfNV8xM19sLmpwZw==
ATI9700Pro x6AA x16AF
http://www.hardocp.com/image.html?image=MTA0MzYyMDg1OTVjVVNkMzFISXhfNV8xNF9sLmpwZw==

Here's a closeup of whats happening with GFFX AA
http://www.hardwareaccess.co.uk/NV30-ATIAA.jpg

 

[surfhurleydude]

I dont know, but based off of those UT screen shots, quality wise, I think the GF FX's 8xAA might be slightly better than the Radeon's 6xAA... Only reason I say that is because the gun in the GF FX screenshot has a few less jaggies, and that's after analyzing it. Quality wise, I think 8xAA 8x Ansio GF FX is pretty much equal to Radeon 9700 Pro 6xAA 16xAnsio based off of these screen shots. You really have to be picky to notice much of a difference, but for practical use they seem about the same quality wise... Speed wise though.....well...... you can just look at the counter and see. GF FX's 8xAA doesn't seem very practical.

 

[binmaze]

...

I dont know, but based off of those UT screen shots, quality wise, I think the GF FX's 8xAA might be slightly better than the Radeon's 6xAA... Only reason I say that is because the gun in the GF FX screenshot has a few less jaggies, and that's after analyzing it. Quality wise, I think 8xAA 8x Ansio GF FX is pretty much equal to Radeon 9700 Pro 6xAA 16xAnsio based off of these screen shots. You really have to be picky to notice much of a difference, but for practical use they seem about the same quality wise... Speed wise though.....well...... you can just look at the counter and see. GF FX's 8xAA doesn't seem very practical.

Though i totally agree IQ is quite subjective, AA method of R300 is better than that of NV30. Period.
Even people like Chalnoth acknowledge that.
You'd better keep comparing screen shots.

 

[THe_KELRaTH]

surfhurleydude: look closer at the horizontal pipework at the upper right side of the 2 pics.

 

[surfhurleydude]

Though i totally agree IQ is quite subjective, AA method of R300 is better than that of NV30. Period.
Even people like Chalnoth acknowledge that.
You'd better keep comparing screen shots.

How many people actually think about the method being used to take away the aliasing in games while they are shooting up each other in a game of UT2003? While the AA method of the R300 might be better than the NV30s, as long as they look good when you are playing games, what does it matter? Since the GF FX needs 8xAA to compare (even though not totally match) to Radeon's 6xAA, yeah, it looks as if the Radeon obviously has a much stronger AA method. But it doesn't mean they don't look the same. I think it's the critics' job to compare and analyze screen shots for AA quality. As a consumer and gamer, it's my job to compare the AA quality in actual gameplay. I don't play screenshots, do I? Now I'm not sticking up for the GF FX's AA methods at all, because you can just see that the speed is absolutely horrible, but to say that the Quality isn't there is just something I don't agree with.

 

[surfhurleydude]

surfhurleydude: look closer at the horizontal pipework at the upper right side of the 2 pics.

I see... Yeah the GF FX does them slightly worse. However, the screenshots are at different angles... On a slightly different note, why are the bars black in the upper left in the GF FX screen and grey in the Radeon 9700 screen?

 

[binmaze]

But then again, it's not easy to discern the differences while you're in the heat of gaming...

Therefore, if you don't notice that while gaming, it's all ok for you. If you do, it'll annoy you.

It's nvidia hype about GFFX supreme IQ which turned out to be worse than 9700pro that really puts GFFX in awkward situation about these IQ things.

I agree with you on most points.
What i saying is AA IQ of R300 is better than that of NV30. It's a simple fact.

And even though you cannot notice the difference most of the time, you can sometimes.
If not you, someone might.

 

[binmaze]

On a slightly different note, why are the bars black in the upper left in the GF FX screen and grey in the Radeon 9700 screen?

Because GFFX doesn't do fog effect correctly.

 

[BRiT]

On a slightly different note, why are the bars black in the upper left in the GF FX screen and grey in the Radeon 9700 screen?

This is a case of the Nvidia GF-FX not having gamma-corrected FSAA. The ATI R300 has gamma-corrected FSAA. Having gamma-corrected FSAA allows for the pipes in this scene ending up being significantly closer to the intended color. Another strong reason why the ATI R300 has significantly better AA IQ.

It could also be related to Nvidia having poor drivers, ones which do not handle fog correctly.

 

[THe_KELRaTH]

I suppose I should have added a bit more detail in my 1st post re:You will also notice the effect of having Gamma correction.

surfhurleydude: Yup having slightly different angles doesn't help when trying to make a comparison and that's why I added the 3rd link as it better shows why the horizontal pipework looks worse. It seems that at offsets of 80 - 100 and 260 - 280 degrees the NV30 AA implementation drops from 8xS to x2AA, (this seems to be the same for any of NV30's AA levels), whereas the ATI implementations are the same at any angle.

 

[Pete]

ATi's pic looks better.

1. In the R300's pic, the gun is displayed against a light backgraound, thus you perceive the jaggies on top more easily. But it looks to me as if the R300 and NV30 are about equal there.

2. The R300 shows less jaggies on the building's far-left roof/overhang, and on the walkway's edge on the left side of the shot (under the overhang).

3. The pipes are obviously very dramatic examples, and not very good ones, as NV30's lack of fog will no doubt be fixed in later drivers. As I said regarding the gun, dark jaggies stand out more against a light background.

4. The overhang in front supported by those four angular "columns" is odd: the underside looks smoother in the NV30 screenie, but the top looks the same in both.

5. The NV30 does a better job with the forward edges of the columns, particularly visible with the far right column.

I do believe you can objectively determine which AA method is better, but I imagine in motion, once you reach a certain level, you'll be hard-pressed to spot differences (particularly in FPS's).

I look forward to 8x AA in R350.

 

[BRiT]

Quite right, it should be very easy to notice the difference between the Nvidia GF-FX top IQ mode and the ATI R300 top IQ mode. The Nvidia one will seem like a slide-show while the ATI one will be nice and fluid. With the latest Catalyst drivers, 3.1, ATI has delivered significant speed improvements in FSAA modes, so the difference is even more pronounced.

 

[Nagorak]

I do believe you can objectively determine which AA method is better, but I imagine in motion, once you reach a certain level, you'll be hard-pressed to spot differences (particularly in FPS's).

In reality you're not going to be using 6XAA when running UT2k3 anyway.