R420 IQ Comparison

[crushinator]

I'm detecting NV40 brilinear in 10 seconds, while it took me 5 minutes to choose R420 @ 16xAF as the sharpest shot, yam!

 

[Ruined]

TO ALL THOSE WHO MADE NICE PICTURES WITH RED CIRCLES:

YOU ARE BLIND.

Load up the crompresonator and actually COMPARE the images.
your circled areas are most assuredly not relevant. You are decieving yourselves.

Are we back to the mathematical thing again?

OMG I totally missed the grass!

If you are referring to this:
http://www.reflectonreality.com/nv17r420/

There are differences in the red circles. Subtle, but they are there. There are differences in other parts of the image too such as grass which are obvious, etc, but those differences in the red circles are the ones I actually noticed with my eyes. And 5 or 6 people saw them too, I don't think we are all blind.

But no, its not a big deal.

 

[Malfunction]

TO ALL THOSE WHO MADE NICE PICTURES WITH RED CIRCLES:

YOU ARE BLIND.

Load up the crompresonator and actually COMPARE the images.
your circled areas are most assuredly not relevant. You are decieving yourselves.

Are we back to the mathematical thing again?

OMG I totally missed the grass!

If you are referring to this:
http://www.reflectonreality.com/nv17r420/

There are differences in the red circles. Subtle, but they are there. There are differences in other parts of the image too such as grass which are obvious, etc, but those differences in the red circles are the ones I actually noticed with my eyes. And 5 or 6 people saw them too, I don't think we are all blind.

But no, its not a big deal.

lmao.... I agree, I don't believe I am blind yet I guess I am suppose to believe it now since so many can't see it. I noticed what you pointed out easily, mine are more subtle but there no less.

 

[Bjorn]

Couldn't find any obvious differences in those screenshots. But why are they all so dark ??

I hope Dave finds his CoD disc so that we can get this out of the way. And then it would be great with some videos/screenshots of Ati's brilinear, "fast trilinear", Nvidias "full trilinear",brilinear.

 

[ChrisW]

Man, you guys are so full of it it's insane. The difference between those circles are so slight you would almost never tell the difference. Why don't you guys tell us which card is producing the correct image and how you determined that. And let's just forget all about the differences in gamma/gamma correction/PS2.0 to PS1.1/etc. of the two cards while you are at it and just say it's obviously a difference in trilinear filtering.

 

[Malfunction]

Man, you guys are so full of it it's insane. The difference between those circles are so slight you would almost never tell the difference. Why don't you guys tell us which card is producing the correct image and how you determined that. And let's just forget all about the differences in gamma/gamma correction/PS2.0 to PS1.1/etc. of the two cards while you are at it and just say it's obviously a difference in trilinear filtering.

As long as we are gonna ask each other to do things, tell me... what is Trilinear AF suppose to do and which of the screen shots appears to be doing more of it?

If you can answer that one Chris, then I will be able to understand why you wouldn't be able to see any difference between the pics.

Fair enough?

 

[ChrisW]

Man, you guys are so full of it it's insane. The difference between those circles are so slight you would almost never tell the difference. Why don't you guys tell us which card is producing the correct image and how you determined that. And let's just forget all about the differences in gamma/gamma correction/PS2.0 to PS1.1/etc. of the two cards while you are at it and just say it's obviously a difference in trilinear filtering.

As long as we are gonna ask each other to do things, tell me... what is Trilinear AF suppose to do and which of the screen shots appears to be doing more of it?

If you can answer that one Chris, then I will be able to understand why you wouldn't be able to see any difference between the pics.

Fair enough?

I thought we were talking about trilinear filtering not anisotropic filtering.

 

[Malfunction]

Man, you guys are so full of it it's insane. The difference between those circles are so slight you would almost never tell the difference. Why don't you guys tell us which card is producing the correct image and how you determined that. And let's just forget all about the differences in gamma/gamma correction/PS2.0 to PS1.1/etc. of the two cards while you are at it and just say it's obviously a difference in trilinear filtering.

As long as we are gonna ask each other to do things, tell me... what is Trilinear AF suppose to do and which of the screen shots appears to be doing more of it?

If you can answer that one Chris, then I will be able to understand why you wouldn't be able to see any difference between the pics.

Fair enough?

I thought we were talking about trilinear filtering not anisotropic filtering.

My bad, that is what I meant. I was reading something else... so drop the AF from my question and answer it. Thank you,

 

[mozmo]

One thing i've noticed with brilinear vs true trilinear is it appears ever so slightly that the textures are a little sharper in brilinear, this is probably due to the less agreesive blending of a higher res texture with a lower res blurrier mip below texture. True trilinear seems to give a smoother image across the whole depth , Although to some the sharper textures from the less agresive brilinear might appeal to them and look better, of course the side effect is possibly more shimmer when moving. One thing that most can agree is that brilinear seems to remove the nasty mip level change lines that bilinear has.

 

[noko]

Dave it would help if you included a third shot with straight bilinear to show where the mipmap lines are clearly and to show the increase in IQ going to Tri/opt Tri.

I am not convinced this is a good shot for this comparison but could be wrong. I would highly recommend Mafia road view because it is bright, flat and long. Any filtering problems will probably show up real quick in Mafia. That was the game I was having a cow about with my Raddy with AF on quality mode until rTool showed up to force full Trilinear. That is if you still have Mafia.

 

[Ostsol]

Dave it would help if you included a third shot with straight bilinear to show where the mipmap lines are clearly and to show the increase in IQ going to Tri/opt Tri.

There would have to be a bilinear shot for each card, since there's no guarantee that the LOD levels are the same in the drivers and that the cards do not handle LOD different internally.

 

[ChrisW]

Man, you guys are so full of it it's insane. The difference between those circles are so slight you would almost never tell the difference. Why don't you guys tell us which card is producing the correct image and how you determined that. And let's just forget all about the differences in gamma/gamma correction/PS2.0 to PS1.1/etc. of the two cards while you are at it and just say it's obviously a difference in trilinear filtering.

As long as we are gonna ask each other to do things, tell me... what is Trilinear AF suppose to do and which of the screen shots appears to be doing more of it?

If you can answer that one Chris, then I will be able to understand why you wouldn't be able to see any difference between the pics.

Fair enough?

I thought we were talking about trilinear filtering not anisotropic filtering.

My bad, that is what I meant. I was reading something else... so drop the AF from my question and answer it. Thank you,

Explaining trilinear filtering to you would be a waste of time. Assuming you think the NV17 image is correct and the X800 is wrong, I want to know how you came to that determination. It should be simple for someone that already has all the answers.

While you are at it, explain why ATI comes much closer to the reference raster:
http://www.tomshardware.com/graphic/20040504/ati-x800-35.html

And before you say something about the color mipmaps, look at the black and white mipmaps.

By the way, it looks like nVidia does not do correct trilinear filtering in their color mipmaps either. Where is the outrage over this? Where are the calls for nVidia to start doing correct trilinear filtering and re-benchmark everything?

 

[Bjorn]

While you are at it, explain why ATI comes much closer to the reference raster:
http://www.tomshardware.com/graphic/20040504/ati-x800-35.html

And before you say something about the color mipmaps, look at the black and white mipmaps.

By the way, it looks like nVidia does not do correct trilinear filtering in their color mipmaps either.

You missed a part from that review though:

The interesting thing is that X800 and Refrast look better on the screenshots in a head to head comparison. This is because of the higher level Mipmaps at 45-degree angles. Although this looks better on screenshots, it can cause sparkle in motion.

Here´s what Microsoft has to say about it:

Quote: "The DX9 reference rasterizer does not produce an ideal result for level-of-detail computation for isotropic filtering, the algorithm used in NV40 produces a higher quality result. Remember, our API is constantly evolving as is graphics hardware, we will continue to improve all aspects of our API including the reference rasterizer."

 

[Ostsol]

Quote: "The DX9 reference rasterizer does not produce an ideal result for level-of-detail computation for isotropic filtering, the algorithm used in NV40 produces a higher quality result. Remember, our API is constantly evolving as is graphics hardware, we will continue to improve all aspects of our API including the reference rasterizer."

In other words, the reference rasterizer is just an implementation of another interpretation of the spec, which apparently isn't all that strict on this subject.

 

[ChrisW]

While you are at it, explain why ATI comes much closer to the reference raster:
http://www.tomshardware.com/graphic/20040504/ati-x800-35.html

And before you say something about the color mipmaps, look at the black and white mipmaps.

By the way, it looks like nVidia does not do correct trilinear filtering in their color mipmaps either.

You missed a part from that review though:

The interesting thing is that X800 and Refrast look better on the screenshots in a head to head comparison. This is because of the higher level Mipmaps at 45-degree angles. Although this looks better on screenshots, it can cause sparkle in motion.

Here´s what Microsoft has to say about it:

Quote: "The DX9 reference rasterizer does not produce an ideal result for level-of-detail computation for isotropic filtering, the algorithm used in NV40 produces a higher quality result. Remember, our API is constantly evolving as is graphics hardware, we will continue to improve all aspects of our API including the reference rasterizer."

That's my point. Just what do you use to determine which image is correct? The only thing we have is the reference raster which it seems ATI is very close to. How can it be said that ATI is cheating when the image is so close to the reference?

 

[Bjorn]

That's my point. Just what do you use to determine which image is correct? The only thing we have is the reference raster which it seems ATI is very close to. How can it be said that ATI is cheating when the image is so close to the reference?

The problem is that Nvidias "high quality" that is using brilinear has also passed MS WHQL tests which means that they also must have ok quality compared to the reference rasterizer. And Ati now (all of a sudden) seems to think that MS tests are very good at determining filtering quality.

And it's pretty obvious that the WHQL tests doesn't say that much and as you can see, MS says that Nvidia has better quality even though it differs more from the reference rasterizer in this case.

 

[Malfunction]

Explaining trilinear filtering to you would be a waste of time. Assuming you think the NV17 image is correct and the X800 is wrong, I want to know how you came to that determination. It should be simple for someone that already has all the answers.

Don't assume anything with me, I merely pointed out what I recognize as a noticible difference. If you can't see the difference in the screen shots, you're either blind, tired or lying to yourself.

 

[Ruined]

Don't assume anything with me, I merely pointed out what I recognize as a noticible difference. If you can't see the difference in the screen shots, you're either blind, tired or lying to yourself.

You'd probably also need a decent monitor, too.

 

[ChrisW]

Explaining trilinear filtering to you would be a waste of time. Assuming you think the NV17 image is correct and the X800 is wrong, I want to know how you came to that determination. It should be simple for someone that already has all the answers.

Don't assume anything with me, I merely pointed out what I recognize as a noticible difference. If you can't see the difference in the screen shots, you're either blind, tired or lying to yourself.

I assumed you pointed this out as evidence that ATI's trilinear filtering method is lowering image quality somehow. If that was not your intention then I apologize.

 

[Althornin]

TO ALL THOSE WHO MADE NICE PICTURES WITH RED CIRCLES:

YOU ARE BLIND.

Load up the crompresonator and actually COMPARE the images.
your circled areas are most assuredly not relevant. You are decieving yourselves.

Are we back to the mathematical thing again?

OMG I totally missed the grass!

If you are referring to this:
http://www.reflectonreality.com/nv17r420/

There are differences in the red circles. Subtle, but they are there. There are differences in other parts of the image too such as grass which are obvious, etc, but those differences in the red circles are the ones I actually noticed with my eyes. And 5 or 6 people saw them too, I don't think we are all blind.

But no, its not a big deal.

lmao.... I agree, I don't believe I am blind yet I guess I am suppose to believe it now since so many can't see it. I noticed what you pointed out easily, mine are more subtle but there no less.

No, the differences you speak of DO NOT EXIST. The areas you have circled show no greater concentration of differences (ie, almost zero) than the rest of the image.
Actually compare the images in an editor, and you will see that your eyes are tricking you, sorry - you are both WRONG.

I also love how one of you says "others see the differences, they must be there" and the other says "exactly! who cares that others can't see it,t hey must be wrong".
Tell me you see the problem with saying that evidence for you is valid, while evidence against you is invalid.....

and whats witht he "mathematical thing" comment? surely even you can comprehend that the easiest way to see image differences is to compare them mathematically, not with your eye. If there is no mathematical difference, and you insist you "see" one, then you are wrong, period.