StellaArtois
Newcomer
Ok - I'll have to go with the minority. Right picture looks best on my monitor, without using any fancy zoom programs or anything. Well that's my initial reaction to them.
AA: which one looks better?
- Thread starter Ante P
- Start date
I also think the right one looks better.
On my monitor I see almost no visible aliasing at all. On the left one I see a noticable staircase effect though.
But in the full pics the nVidia pics has some strange aliasing in the left part of the image, I dunno what's up with that.
personally I feel nVidias 8x look soooooo much better in all the games I've tried. It's almost impossible to spot any aliasing at all, but ATis 6x I see it pretty clearly, perhaps not jaggies per se, but the gradients are not as "smooth" as Nvs 8x.
though In those full screen pics I fail to see any SS effect at all, the textures on the ground looks exactly the same(?)
On my monitor I see almost no visible aliasing at all. On the left one I see a noticable staircase effect though.
But in the full pics the nVidia pics has some strange aliasing in the left part of the image, I dunno what's up with that.
personally I feel nVidias 8x look soooooo much better in all the games I've tried. It's almost impossible to spot any aliasing at all, but ATis 6x I see it pretty clearly, perhaps not jaggies per se, but the gradients are not as "smooth" as Nvs 8x.
though In those full screen pics I fail to see any SS effect at all, the textures on the ground looks exactly the same(?)
Doomtrooper
Veteran
Thats why full screen shots are a must (see image quality thread) as there may be areas where one card exhibits its AA better than others.
gokickrocks
Regular
same here
No option for "what the heck am I supposed to be able to tell with just a tiny portion of the screen?"
I think it's disengenuous to post such a small sample from a screenshot. Full screen shots should be used to compare aa. If I want to zoom in on a particular spot I can do that myself. Don't know of many photo viewing programs that don't include a zoom.
On a side note, I could only get the pictures on the left to come into view. Don't have the patience to wait for the pictures on the right to load. Waited 5 minutes and all I got was the top and bottom picture. There is supposed to be pictures on the left and right correct?
One thing I have noticed is that the aa in screenshots doesn't look nearly as good is it does in the game on my 9700. Jaggies that aren't visible in the game are clearly visible in screenshots. Suppose that has to do with the gamma correction though.
I think it's disengenuous to post such a small sample from a screenshot. Full screen shots should be used to compare aa. If I want to zoom in on a particular spot I can do that myself. Don't know of many photo viewing programs that don't include a zoom.
On a side note, I could only get the pictures on the left to come into view. Don't have the patience to wait for the pictures on the right to load. Waited 5 minutes and all I got was the top and bottom picture. There is supposed to be pictures on the left and right correct?
One thing I have noticed is that the aa in screenshots doesn't look nearly as good is it does in the game on my 9700. Jaggies that aren't visible in the game are clearly visible in screenshots. Suppose that has to do with the gamma correction though.
Went and had a look at the full screens. As far as the verticles go, in some places it looks better on the nv25 in other places it looks better on the 9700. It all depends on where you look. That is why I think full screenshots are important. As far as the horizontal lines goes though, the nv25 pictures look horrid. Don't know if that's due to the angle though. Overall I think the 9700 picture looks much better. Mainly because of the horizontal lines being so bad on the nv25. The verticles could go either way depending on where you look.
I opened these full screen in acdsee and used my mouse wheel to scroll back and forth between them. The difference was drastic enough that even my wife could clearly see the difference. (something that doesn't happen often with her.) She also thought the 9700 screen looked much better.
I opened these full screen in acdsee and used my mouse wheel to scroll back and forth between them. The difference was drastic enough that even my wife could clearly see the difference. (something that doesn't happen often with her.) She also thought the 9700 screen looked much better.
I'd Disagree.Ante P said:bandwidth ain't too shabby either
SS AA will eat more bandwidth, and the GFFX already has less bandwidth than the 9700.
I Highly doubt performance of 8xS will be reasonable to use, except in a few cases (very old games), where its ordered grid nature will reduce its possible effectiveness.
IMO, this is the failing of the GFFX - lack of quality AA modes.
I focused on the 3rd tube on the nearest face going up, because there was better anti-aliasing in the nv25 image (in contrast, all other cases I noticed the nv25 was no better and often worse, as noted in other's comments about horizontal edges) . I then looked more closely, and after observing the other tubes behind it, I concluded that the angle of the line is different and that is the reason the nv25 anti-aliased better.
As I result, I think using the pictures for detailed comparison is completely invalidated, because the line angles we'd use for comparison are arbitrarily different throughout the image, and that has drastic effects on how smooth the edge looks.
If there is no way for the program to guarantee equivalent images, it seems a pretty questionable tool for comparison to use under any circumstance.
[opinion based on our prior conversations]
By the way, your "SSAA + MSAA" threads about the R300 "just happening" to have occurred just prior to this flawed comparison where you "happen" to pass of your evaluation as a given ('well "in action" 8x does a better job anyways and that's what matters' ) is not a good idea in combination with having established you don't care what methods you use as long as they "get results". I'd recommend you take more care to assure the equivalancy of the comparison images when you've introduced such questions about your motivations.
[/opinion based on our prior conversations]
Maybe we could try again with another program that has equivalent images for comparison, as I think the idea of the thread is a good one. Some interesting discoveries possibly due to gamma settings with the split of the vote as well. I'm still waiting for Chalnoth to jump in about that.
For reference in case we move forward with the gamma idea, perhaps we could pick a reference method of determining our gamma settings and compare between people who prefer the left and who prefer the right.
As I result, I think using the pictures for detailed comparison is completely invalidated, because the line angles we'd use for comparison are arbitrarily different throughout the image, and that has drastic effects on how smooth the edge looks.
If there is no way for the program to guarantee equivalent images, it seems a pretty questionable tool for comparison to use under any circumstance.
[opinion based on our prior conversations]
By the way, your "SSAA + MSAA" threads about the R300 "just happening" to have occurred just prior to this flawed comparison where you "happen" to pass of your evaluation as a given ('well "in action" 8x does a better job anyways and that's what matters' ) is not a good idea in combination with having established you don't care what methods you use as long as they "get results". I'd recommend you take more care to assure the equivalancy of the comparison images when you've introduced such questions about your motivations.
[/opinion based on our prior conversations]
Maybe we could try again with another program that has equivalent images for comparison, as I think the idea of the thread is a good one. Some interesting discoveries possibly due to gamma settings with the split of the vote as well. I'm still waiting for Chalnoth to jump in about that.
For reference in case we move forward with the gamma idea, perhaps we could pick a reference method of determining our gamma settings and compare between people who prefer the left and who prefer the right.
