Anandtech Fall 2003 Video Card Roundup - Part 2: High End Shootout **60 pages**

[NFS4]

http://www.anandtech.com/video/showdoc.html?i=1896

 

[josedawg]

Darn, beaten to the punch by 6min

 

[hominid skull]

I want animated slides for the iq comparisons with big images. There's no way i can tell the iq differences between the three images when they are that small.

Anand just don't get it.

Atleast have a link to a bigger image so we can check for ourselves, i just don't trust nvidia anymore.

 

[Dug]

:Q
I wish I could open the pages up from work. For some reason Anantech is the only site that we can't access. (Well we can but it takes 3-5 minutes for each page to load.)

 

[First]

Originally posted by: hominid skull
I want animated slides for the iq comparisons with big images. There's no way i can tell the iq differences between the three images when they are that small.

Anand just don't get it.

Atleast have a link to a bigger image so we can check for ourselves, i just don't trust nvidia anymore.

Um, it doesn't really matter, there were no noticable IQ differences.

 

[hominid skull]

ok, from what i can tell in aquamark from such small iq comparison images, is that with no aa the new nvidia drivers are a step forward compared to the 45.23's. But the images that are labled highest AA/8xAF (whatever highest AA is) the nvidia drivers don't look like they are using AA at all and a lower AF setting too.

 

[First]

Originally posted by: hominid skull
ok, from what i can tell in aquamark from such small iq comparison images, is that with no aa the new nvidia drivers are a step forward compared to the 45.23's. But the images that are labled highest AA/8xAF (whatever highest AA is) the nvidia drivers don't look like they are using AA at all and a lower AF setting too.

Hum, not sure I see that. Where exactly do you not see the AA being applied correctly?

 

[DaveSimmons]

<keanu> whoa! </keanu>

9800xt is still the best card, but fx5900 / 5950 are looking much better than they did a few weeks ago. People dumping their 5900 in FS/FT might want to cancel the sale

 

[hominid skull]

Originally posted by: Evan Lieb

Originally posted by: hominid skull
ok, from what i can tell in aquamark from such small iq comparison images, is that with no aa the new nvidia drivers are a step forward compared to the 45.23's. But the images that are labled highest AA/8xAF (whatever highest AA is) the nvidia drivers don't look like they are using AA at all and a lower AF setting too.

Hum, not sure I see that. Where exactly do you not see the AA being applied correctly?

i've made a nice animated gif for you: here.

take a look at the edges on the vehicle, see what i mean.

Would have been more visable if we had bigger images to play with.

edit..

If added another comparison here. This is with the F1 99-02 with the all on settings. Notice the aliasing of the yellow line on the 52.14 drivers, also the AF of the tarmac is a lot better with the ATI card/driver combo. I'm sorry, i wouldn't call Nvidias 4xAA8zAF that at all, it's much lower in quality - which is the point of the new drivers. Lower quality for more FPS.

 

[First]

Originally posted by: hominid skull

Originally posted by: Evan Lieb

Originally posted by: hominid skull
ok, from what i can tell in aquamark from such small iq comparison images, is that with no aa the new nvidia drivers are a step forward compared to the 45.23's. But the images that are labled highest AA/8xAF (whatever highest AA is) the nvidia drivers don't look like they are using AA at all and a lower AF setting too.

Hum, not sure I see that. Where exactly do you not see the AA being applied correctly?

i've made a nice animated gif for you: here.

take a look at the edges on the vehicle, see what i mean.

Would have been more visable if we had bigger images to play with.

Uh, I would have to say that is a ridiculously small difference. Basically no difference.

 

[hominid skull]

ok, not huge aliasing but if I a nobody computer user can spot defects like that, on images that small, then why couldn't the authors of the article spot them on images 4x that size?

also take a look at my edit

copy/paste.


I've added another comparison here. This is with the F1 99-02 with the all on settings. Notice the aliasing of the yellow line on the 52.14 drivers, also the AF of the tarmac is a lot better with the ATI card/driver combo. I'm sorry, i wouldn't call Nvidias 4xAA8zAF that at all, it's much lower in quality - which is the point of the new drivers. Lower quality for more FPS.

 

[oldfart]

I can see it. There is one horizontal part of the vehicle right about in the middle that has jaggies that are very apparent even in that small image. I would have never seen it, but is is there for sure.
hominid skull, you have very good eyes. to me, the foliage on the bottom also looks sharper on the Ati card, but I have to agree. The small images are hard to work with. A bigger image would be better.

Anyway, fantastic review. Kudos for the AT guys for burning the midnight oil on this one! I an pleased to see the Det 52.xx drivers improving performance without big IQ losses.

 

[First]

Originally posted by: hominid skull
ok, not huge aliasing but if I a nobody computer user can spot defects like that, on images that small, then why couldn't the authors of the article spot them on images 4x that size?

also take a look at my edit

copy/paste.


I've added another comparison here. This is with the F1 99-02 with the all on settings. Notice the aliasing of the yellow line on the 52.14 drivers, also the AF of the tarmac is a lot better with the ATI card/driver combo. I'm sorry, i wouldn't call Nvidias 4xAA8zAF that at all, it's much lower in quality - which is the point of the new drivers. Lower quality for more FPS.

Again, these are incredibly small IQ differences that you would never notice when actually playing the game. Saying either driver is better than the other in terms of IQ would be misleading.

 

[lordtyranus2]

Why was this done before the Catalyst 3.8 was released?

 

[First]

Originally posted by: lordtyranus2
Why was this done before the Catalyst 3.8 was released?

Because ATI couldn't get us the drivers on time. Besides, you'll have a separate review on the 3.8's, when ATI can actually get them to us.

 

[lordtyranus2]

Because ATI couldn't get us the drivers on time. Besides, you'll have a separate review on the 3.8's, when ATI can actually get them to us.

Ah. Yea, I suppose this must have taken a few days at least to do. Doing new 3.8 reviews just seems like a lot of extra work from my vantage point, depending on how much it changes things.

In any case, excellent job, though I confess I only read the final words.

 

[josedawg]

Originally posted by: Evan Lieb

Originally posted by: lordtyranus2
Why was this done before the Catalyst 3.8 was released?

Because ATI couldn't get us the drivers on time. Besides, you'll have a separate review on the 3.8's, when ATI can actually get them to us.

Which is due out tomorrow @ 3pm EST

 

[OddTSi]

Originally posted by: Evan Lieb
Again, these are incredibly small IQ differences that you would never notice when actually playing the game. Saying either driver is better than the other in terms of IQ would be misleading.

No offense intended to any of the AnandTech crew but, wouldn't it have been better to do as hominid skull said and post larger images with perhaps an animated gif so that these differences were more visible. What is and isn't "too small of a difference" is really subjective and should be left up to the user to decide.

Just my $1.50

 

[lordtyranus2]

Which is due out tomorrow @ 3pm EST

They have to work on it, this kind of project doesn't take 10 seconds. Though I would have pushed it back until the 3.8s came out.

 

[draggoon01]

Originally posted by: Evan Lieb

Originally posted by: hominid skull
ok, not huge aliasing but if I a nobody computer user can spot defects like that, on images that small, then why couldn't the authors of the article spot them on images 4x that size?

also take a look at my edit

copy/paste.


I've added another comparison here. This is with the F1 99-02 with the all on settings. Notice the aliasing of the yellow line on the 52.14 drivers, also the AF of the tarmac is a lot better with the ATI card/driver combo. I'm sorry, i wouldn't call Nvidias 4xAA8zAF that at all, it's much lower in quality - which is the point of the new drivers. Lower quality for more FPS.

Again, these are incredibly small IQ differences that you would never notice when actually playing the game. Saying either driver is better than the other in terms of IQ would be misleading.

do you do the same with small difference in fps or other benchmarks?

 

[First]

Originally posted by: OddTSi

Originally posted by: Evan Lieb
Again, these are incredibly small IQ differences that you would never notice when actually playing the game. Saying either driver is better than the other in terms of IQ would be misleading.

No offense intended to any of the AnandTech crew but, wouldn't it have been better to do as hominid skull said and post larger images with perhaps an animated gif so that these differences were more visible. What is and isn't "too small of a difference" is really subjective and should be left up to the user to decide.

Just my $1.50

Oh sure, I agree, larger shots would have been better. But again, most IQ differences can easily be seen with the images included in the review. If you really need a bigger picture to see the difference, then it's quite clear that the IQ difference is small, at worst.

For example, the 51.75 drivers clearly had IQ issues with AM3, as shown by THG here.

Originally posted by: draggoon01

Originally posted by: Evan Lieb

Originally posted by: hominid skull
ok, not huge aliasing but if I a nobody computer user can spot defects like that, on images that small, then why couldn't the authors of the article spot them on images 4x that size?

also take a look at my edit

copy/paste.


I've added another comparison here. This is with the F1 99-02 with the all on settings. Notice the aliasing of the yellow line on the 52.14 drivers, also the AF of the tarmac is a lot better with the ATI card/driver combo. I'm sorry, i wouldn't call Nvidias 4xAA8zAF that at all, it's much lower in quality - which is the point of the new drivers. Lower quality for more FPS.

Again, these are incredibly small IQ differences that you would never notice when actually playing the game. Saying either driver is better than the other in terms of IQ would be misleading.

do you do the same with small difference in fps or other benchmarks?

Exactly. Anand and Derek mentioned a couple games where the fps difference was very small.

 

[lordtyranus2]

How big will the followup be? As extensive as the current one?

 

[josedawg]

Originally posted by: lordtyranus2

Which is due out tomorrow @ 3pm EST

They have to work on it, this kind of project doesn't take 10 seconds. Though I would have pushed it back until the 3.8s came out.

I know its not easy to do, just stating that the 3.8's will be out tomorrow at 3pm EST, so expect the reviews out tomorrow. Not too long of a wait.

 

[First]

Originally posted by: lordtyranus2
How big will the followup be? As extensive as the current one?

All that really needs to be done is compare the current CAT 3.7 results to the CAT 3.8's Anand and Derek will benchmark soon (9800XT). One card isn't as hard as four.

 

[Luagsch]

hmm my first post in this thread and probably not the last.

first of great job to derek and anand.

about that f1 99-02 stuff:
derek wrote that nvidia prsumably has problems with it and that they're working on it. maybe the graf-bugs are also getting eliminated (clearly fsaa bugs (just take look at the michelin-thingy, no fsaa on nvidia))