Does anyone find this statement a Joke ??

[Doomtrooper]

I was reading Tom Pabsts review of the Parhelia when I came across this:

In Part II, we'll be looking at the Parhelia's image quality and its performance at maximum quality, with a comparison to an NVIDIA GeForce 4 Ti4600. Here, we won't be including ATI's Radeon 8500, because its anisotropic filtering, which involves various driver tricks, is somewhat controversial and would make a comparison unnecessarily complicated.

As proven on these very forums ATI's Anistropic filtering delivers excellent image quality but is left out due to driver 'TRICKS' ??

http://www17.tomshardware.com/graphic/02q2/020625/parhelia-06.html

 

[Nagorak]

Tom's Hardware suxx0rs.

 

[tamattack]

The statement (ex-8500 because of different aniso) and the intent (IQ) are obviously at odds with one another.

If you apply this type of thinking throughout, then Matrox FAA cannot possibly be compared to NV FSAA (whether performance or IQ-wise) because they are such different implentations.

What you are left with is an IQ review focusing on default rendering -- no aniso, no AA. Then why bother reviewing IQ at all?

 

[Moloch]

What do you expect from Tom's guys? they are damn fools and we all outta know that by now.

edited to be more broad(not just tom his staff)

 

[Doomtrooper]

Especially after countless threads on this forum dispelling ripmapping and other theories and Iq comparisons like this ONE.

Excellent analogy tamattack...

 

[Doomtrooper]

I just noticed this Parhelia review was done by Lars Weinand, isn't he the founder of RIVA Station ?? Whats a Nvidia fan site founder doing a review for a supposed non-fan site ??

 

[Seiko]

Jeez, Toms at it again

Ok, this isn't meant to be a flame nor try and put the R8500 on top of any perch but when a major web review site intentionaly leaves out one of ATIs strongest features I start to get worried? The lame excuse of controversial methods is absolute bull**** at best and highly biased at worst. I'm under no false illusion that the 8500 is a faster card than GF4 as it clearly isn't but that's not the point! Surely as consumers we all need to be told about all methods and approachs to image quality. I don't care about the numbers, i.e.) FSAA*4 or FSAA*2, 64tap, 32 tap etc I simply want to see some subjective reviews denoting the superior IQ and resulting FPS. Sure I'd also expect to see FPS results without any IQ improving features to see the raw speed of a card but either way I don't want to see any MISSING results just because it makes the review sites favourite card look poor!!!!!!!!!

Apologies for the moan but Toms really does need to get back on track if ppl with a little knowledge will ever be able to find there reviews useful!

Absolute trash!




At least Anands had the balls to post the 8500 strengths!

 

[nggalai]

What do you expect from Tom? he's a damn fool and we all outta know that by now.

Or rather, what do you expect from Borsti (Lars of Rivastation)?

Yet another useless review.

ta,
.rb

 

[Brent]

agreed, it's a pile of doo doo from tom

The Anisotropy of the 8500 SHOULD DEFINITELY be compared with Parhelia.

wtf is he smoking

which involves various driver tricks, is somewhat controversial and would make a comparison unnecessarily complicated.

somewhat controversial? I think not, i've read throug the anisotropic threads here and it has a definite answer. And I don't see how it could make the comparison complicated. Other site have shown some comparisons, like Anand, with anisotropic of GF4/8500 and Parhelia with noproblems and nothing 'complicated'.

This is just one of the many reasons I don't visit THG.

 

[Doomtrooper]

I vote for Brent as Hardocps new webmaster

 

[Brutal Deluxe]

So let me see, the "tricks" used with R8500's aniso makes it unfit for comparision? But the "tricks" used with Parhelias FAA technique is ok to use??

I was under the impession that the R8500's trick was to apply aniso only where deemned nessesary. And that Parhelias trick was to do the exact same thing with FAA.

 

[Simon F]

Does anyone find this statement a Joke?

Only if you consider Macbeth or Julius Ceasar to be among Shakespeare's comedies. I think 'tragedy' (or perhaps 'travesty' ?) is the term needed.

I haven't read the entire article (I'm far too busy for that) but 2 points came to mind:

First, it strikes me as odd that the review begins with performance measurements (which seems to automatically put the Matrox device in a poor light) when the Parhelia is targeted as providing higher quality.

The second point, not aimed specifically at Tom's, is a more general one to do with aniostropic filtering testing. For example, in Tom's review there are static pictures of anisotropically filtered scenes from two different cards and because one looks sharper than the other, the reviewer draws the conclusion that the chip associated with the sharper image must be superior.

This assumption could be completely false because the sharper image may simply appear that way because of aliasing. I'm not saying that the particular chips used are or are not guilty of this, but that the testing method is flawed.

Unfortunately, I'm not sure what the correct approach would be. Perhaps the use of a test scene in which several frames are rendered and the output frequency analysed for noise may work. <shrug>

 

[JF_Aidan_Pryde]

It's pretty much like saying: "The Geforce won't be tested as it uses tricks like HSR to remove workload and hence the test would be unfair."

 

[Fuz]

Isn't it a good thing that ATI can produce better image quality and suffer less of a performance hit in aniso filtering? How some one can consider that as a negative is beyond me

On another note, you guys realise we are doing nothing but helping Tom's pockets get fatter by visiting his site?

I made a mistake by checking out his Parhelia review, it was the first time I had been there in years... and will be another few years till I visit his site again.

 

[BoardBonobo]

It's quite sad that over the last couple of months people have being openly sidelining ATi, The worst of it is that a few respectable magazines are doing the same thing, and these are magazines that have plenty of big readership.

In one recent head to head review the R200 got squeezed into a tiny side box and was basically written off as a non contender. But what was even worse with that is that they mentioned the R300 as an up and coming chip that would shine only until the NV30 came out and then it would be history.

Now it's nice to think that informed people will ignore such bias, but unfortuantely most people aren't that informed. And there is a psychological effect called priming. Make people think about something before you show it to them and they are more likely to recognise it. So if the general readship of sites and magazines are being primed to be negative (and even if you recognise the fact it still works at a sub-conscious level) about Matrox and ATi, then when the reviews are released people are going to percieve a more negative slant to them than is really there.

What makes it worse for Matrox is the fact that they put such a positive spin on the PR for the Parhelia inthe first place making (and I refuse to call it a failure) the disappointment even sharper. The only way I see Matrox shining here is if the Parhelia is taken as a whole and not decomposed in functional units that can be measured independantly. It was never meant to work like that.

Still I expected more from THG, but there seems to be a hidden agenda both in printed material and in the online community to spread FUD about certain compaines and there products.

 

[K.I.L.E.R]

I won't deny that even I considered to buy Ati's cards. They look good. The only thing stopping me are the drivers and FSAA implementation. I have options to use MS AA and SS AA in Ogl on my G3 which is why I like it so much. I will wait to see if Ati, Nvidia and other competitors offer what I want then I will decide very carefully if I should buy it. What TOM is doing makes it harder to decide what I want. Mainly because if you take away the competition because of "driver tricks" I will never know what it looks like so I cannot decide if that's what I would like. How about 3dfx's driver "trick" with HSR? That worked beutifully with vsync on. I was able to run Quake 3 at max frame rates with it on with no error. Even htough it was a "trick" it helped alot for me and I loved it. Weather it was fair or not to compete against is a silly discussion. If you want to compare graphics cards than throw all there is otherwise you are not only limiting options but you are also proving a bias against that card or company.

 

[John Reynolds]

I doubt the author of the review reached that conclusion on his own. Most likely it was fed to him from a competitor's PR dept.

 

[Doomtrooper]

Well I emailed Borsti and asked him why they made that statement and this was my reply:

Hi,

there are some investigations on what ATi is doing in their drivers. ATI
does not filter all textures. Many textures are not filtered at all. They do
not have Anisotropic filtering in the "classic" way. R8500 is very fast with
Anisotropic, but it´s no "real" Anisotroppic. That´s why I did not use it in
the Parhelia review... it´s worth an own article and comparsion and I have
to be very carefully! You can mind that ATI and NVIDIA do only tell us what
they want us to know... so it´s very hard to get those informations. And
it´s very complicated to explain (and compare)

If I used it... I would have start a discussion about ATI... but I wanted to
show what Parhelia can do... I hope you see my point.

Lars

Lars "Borsti" Weinand

- Editor In Chief & Founder - RIVA Station
http://www.rivastation.com
http://www.rivastation.com/index_e.htm (English)

- Senior Editor, Graphics - Tom's Hardware Guide
http://www.tomshardware.com

 

[nggalai]

Hi Doom,

wow, that's funny. That's almost the same thing Borsti said when I asked him about the weird bits in the R8500 drivers review a couple of months ago. I'll see whether I can find that e-mail in my archives.

Seems to take a lot of time to get to grips with different implementations of anisotropic filtering . . .

ok, enough borsti-bashing for today.

ta,
.rb

 

[ram]

http://www.digit-life.com/articles/digest3d/0202/itogi-video-q3ani.html explains it all, but that's no reason to not include comparisons.