Screenshots suggest stingy pixel precision

[Stoo]

The only difference is as I outlined above, the rest of the scene is *identical*, in fact, I'll show you..

area sharpend:


nvtweaked:
Linky (as it's a bit big)

The tweaked one isn't pixel perfect, but that was only because I was being a bit lazy with the brush, but you get the general idea..

[Mike Fishcake]

Ooop... it gets a bit more interesting : http://www.theinquirer.net/default.aspx?article=33131

[Spud1]

more interesting is right...I still think that it's faked, as even ignoring the screenshots, it makes no logical sense does it? Surely it would have been noticed before if it were true..

[cptwhite_uk]

I'd put money on it being a fake - as all other photo/images editors here has noted - the rest of the image is identical - pixel for pixel, all AA and AF applied exactly the same...this just doesn't happen. The area that has been edited is a defined chuck of the picture and is quite clearly something resembling the Unsharpen mask in paint shop pro (which I use). To even suggest this is a 'real' like for like test is just making a mockery of the whole debate. Nvidia may be guilty - but this isn't proof of it.

Also this is coming from someone who prefers ATI - but above all else prefers impartial and accurate information.

[Stoo]

Yup, it's a crock, the rest of the picture is pixel-perfect, with a *clearly* defined area where the image has been sharpened and lightened, sure, a shader *could* to a sharpen and lighten procedure, but why would it do it on such a strange area, and why would it being doing a sharpen and lighten when it's supposed to be showing the difference in the bit depth of the driver?

I've lost any respect that I once had for The Inq, sure, they have some valid information, but It's really not worth having to wade through all the crap for..

If I want a technical explaination for a driver/hardware behaviour, I'll ask our Rys, no BS, no spin, just pure fact.

[chuckskull]

I've just noticed, 'rydermark' is never referred to anywhere online except in this story and others pointing to the inquirer. No site, no download mirrors nothing.

I was going to offer to run the benchmark with my 7800GT and see if anyone was willing to do the same with a X1800/1900.

I have all the proof I need. Then again this is from the place that bought us reverse hyperthreading

[Spud1]

'rydermark' is in development apparently which explains why you cant run it.. It's not a public application yet.

[Stoo]

Even so, I'd exect them to at least have a website to tell us about themselves, even if there currently is no product available to download..

[chuckskull]

Yeah even the most early pre-alpha software has a web page. Also it raises a point whether an application so early in it's development should be used to decide that nVidia is suddenly cheating the GPU market (and all the class action suits, and fines associated with that).

Also Spud1 I challenge you to find an article about rydermark that doesn't link back or reference The inquirer. The only ones I found from before this debacle, are alluding to a new benchmark in development also linking to The Inq.

[Spud1]

Also Spud1 I challenge you to find an article about rydermark that doesn't link back or reference The inquirer. The only ones I found from before this debacle, are alluding to a new benchmark in development also linking to The Inq.

Hey I think this is a fake, don't get me wrong here I don't however, go with the assumption that every piece of software will have a web page while it is in development..Rydermark could be an internal name, an alpha name, a dev name, anything - who knows..

Of course it could just be an elaborate hoax but I see no reason why, and just because we can't find the information doesn't mean that rydermark doesn't exist

[chuckskull]

Of course it could just be an elaborate hoax but I see no reason why, and just because we can't find the information doesn't mean that rydermark doesn't exist

Of course, but no-one else but the inquirer has seen the software either, which is the really interesting point.

I'm sure nvidia would like a copy