A promise is a promise :) there you go!

Hmm, I wonder what kind of cooling solution the card has.

Vegetto-ex, you didn't actually see the card, did you?

 

Huh? The Radeon 8500 didn't support FP24, or any FP for that matter.

 

Well, I have no idea what it does for speed, but on my Radeon 9700 I turn the LOD bias one notch back from standard because it reduces the horrible aliasing alot. Maximum quality in the ATi drivers is nice for screenshots, but not in actual gameplay where the aliasing is horribly visible when you move. But that is my opinion.

 

No but R200 had better precision than GeForce 3. And it was you who went on endlessly how FP24 wasn't enough even though you see no reason to consider an 8500 over a GeForce 3. I guess it's only a big deal when it's ATI

 

not much actually, depends on the game of course but there's a lot of titles where even a lod bias of (+) 15 on nvidia boards doesn't affect performance more than single digit percentages

try running with +15 in 3dmark for an example
doesn't affect the socres all that much

 

Try UT 2003 or RTCW.

 

I don't think it's as simple as 4xOGMS 9tap either

It is ordered grid though, it does have a blur and on low angle edges it's worse than a 4xRGAA.



the pattern isn't the same between pixels but it does repeat on a regular basis.

You'd probably get a result like that if you used quincux then shrunk the result by 75% or something (bilinear filter in photoshop or whatever)

It's not a particularly good pattern and (/sticks neck out) it's probably not real

Well that's my opinion anyway.

 

Interesting as usual Bambers 8)

 

I pretty much agree with most of that, however, on the subject of IQ, remember GF-FX is perfectly capable of proper Trilinear and Anisotropic filtering - just load up early drivers for evidence of that, it's simply that in later drivers nVidia decided to use brilinear filtering instead.

 

The R200 had better color precision support than the GF3. The GeForce3 did all texture calculations in FP32 (this is why texture ops were separated from ALU ops in PS 1.x). The reason I didn't think the added color precision was that great is because it didn't appear anybody was going to use it.

 

Why, exactly, do you think it has a blur filter?

 

Gee, I don't know...

Picture removed to save bandwidth, sorry.

 

Okay, that's a blur filter. What are those images from?

 

On the left, Vegetto's mystery card. On the right, Ssp45's R9700.

 

omg, I just read through this whole thread and now its 4 am...

I have trouble believing the 3dmark score. 8)

I believe it when its published in the orb or something.

How well do you know this tester of the card? I mean... for all we know he could have been nvidia marketing dude or something

 

...unless you own a card that isn't supported by said earlier drivers, like the NV36, and NV38.

In any case, even for owners of cards who have the choice of reverting to older drivers, shader stability and speed is horrible with older drivers.

 

Thanks for that info. I'll check on the newer 24-bit for sure.
I did not know it was only 16.
I really like my colors.

Anyone know the model numbers on the newer LCD's with full 24bit?
I also want DVI inputs.

 

now take those two images and return them to the original size. the so claimed blur filter looks the same when you take it to the original size. woop de freaking do if you think it looks bad. it's zoomed. it looks like crap no matter what.

 

It looks like the Geforce 3 AA...or looking through a beer glass

 

What exactly are you trying to say? That there is no "so-called blur filter"?