PC Magazine's IDIOTS screw Matrox in a preview

[[Ch]amsalot]

For those that haven't already seen the August 2002 "First Looks" of the Matrox Parhelia in PC Magazine, be advised that the morons at Ziff-Davis compared Parhelia's 4xFSAA to GeForce4 Ti 4600's 4xFSAA. Of course, they should have used Parhelia's 16xFAA. F--king idiots.

Before you write your hate mail, be advised that I already sent them the following letter to the editor. I wonder if they'll have the balls to print it:

Dear Ed.:

It's pretty amazing to me that a publication with your history and "expert" staff can neither do a fair nor intelligent preview of a cutting-edge 3D card. I'm talking specifically about Mr. Salvator's "Matrox Aims for ATI and nVidia in 3D Graphics" article that appeared in the August 2002 "First Looks."

Mr. Salvator compared Parhelia's 4xFSAA with that of the 4xFSAA of the GeForce4 Ti 4600. As anyone interested in Parhelia already knows, the architecture is not optimized for FSAA (and hence the poor benchmark results). Rather, Matrox has implemented a vastly superior antialiasing technique known as 16x Fragment Antialising (FAA-16x). This technique antialiases only the edge pixels of triangles and applies 16x super-sampling to them. The result is a performance hit similar to that of GeForce4 Ti's 4600 4xFSAA, yet greatly improved image quality.

Incredibly, Mr. Salvator didn't even mention this extremely important feature of Parhelia -- let alone run a fair comparison benchmark -- doing a disservice both to Matrox and to your readers. Please re-test the Parhelia with FAA-16x in a future issue.

Sincerely,
Mark H. Miller

For everyone's info, they did THE EXACT SAME THING with ATI'S Radeon. They didn't print the "love letter" I sent them a while back about that either.

-[Ch]amsalot

[Stringy]

An unfortunate example of how "slow" and out of touch the print media can be when it comes to the fast paced PC market, esp. the Graphics market...

I'm surprised none-the-less that it came from ZD....
ohwell, even Gaints trip and fall sometimes...



Craig

[[Ch]amsalot]

I'm not surprised. This is the second moronic preview I've seen from this Salvatore guy. He shouldn't be allowed to preview 3D cards anymore.

FYI, show me one other print magazine that didn't report the 16xFAA feature of Parhelia.

Maybe reviewing input devices is more his speed.

-[Ch]ams

[xortam]

Its perfectly legitimate for them to be comparing both cards utilizing the 4xFSAA feature: This is an apples to apples comparison. I would be disappointed if I didn't see a review compare cards with the exact same feature setting. They should have compared the results to the FAA feature as well (I haven't verified if they didn't or not). The FAA feature is a strong selling point and many users seem to be able to take advantage of it quite often.

[[Ch]amsalot]

Xortam,

I respectfully disagree. Matrox's support of 4xFSAA is a "fall back" feature, not really meant to be used. Matrox is providing support for it as a courtesy to the consumer. It was patently unfair, indeed it was outright misleading, to use Parhelia's 4xFSAA in the benchmark. The fact that no mention was even made of 16xFAA shows either that the writer was deliberately attempting to paint Parhelia in a negative light or was an imbecile (or both).

I strongly condemn PC Magazine for this apparently NVIDIA-biased preview.

-[Ch]amsalot

[DuRaNgO]

You wanna see nV bias, go to the HardOCP.com front page and count the words: nvidia, nforce and gefart

I did not put a link directly to them for "obvious" reasons

[xortam]

If a user can't use the FAA feature due to problems with a game then they'll be very interested in comparing the performance with FSAA enabled. Perhaps the user is only interested in those troublesome games. Its a valid and valuable test to compare FSAA performance.

[[Ch]amsalot]

Xortam,

Let's agree to disagree

-[Ch]ams

[xortam]

Ill agree that you're wrong. Don't worry ... nobody's perfect. I thought I wasn't once, but I was wrong.

[P5ycho]

are you serious about the texture quality? since the textures are not touched at all with FAA 16 they should be a lot sharper?

[103er-Fan]

Originally posted by xortam
Its perfectly legitimate for them to be comparing both cards utilizing the 4xFSAA feature: This is an apples to apples comparison. I would be disappointed if I didn't see a review compare cards with the exact same feature setting. They should have compared the results to the FAA feature as well (I haven't verified if they didn't or not). The FAA feature is a strong selling point and many users seem to be able to take advantage of it quite often.

No, even that would be unfair:
Matrox 4xFSAA method is totally different to nVidias, also ATIs and 3dfx' methods are totally different to the nVidia method. So you can't even compare them, some are multsampling, some are supersampling. And both are offering different image and AA quality, especially the 2x settings.

I can somehow understand your statement though, but it is also unfair to compare nVidias Aniso (set to full) and compare it with the P.s Ansio which is only set to level 2.



Edit:

are you serious about the texture quality? since the textures are not touched at all with FAA 16 they should be a lot sharper?

Well, at least on my V5 5500 Grand Prix 3's textures looked MUCH better, especially on the distance, than without FSAA, and I don't think that this would be able with pure edge-AA.

[UberLord]

In a way the reviewers correct.

He's comparing x4 FSAA to x4 FSAA.
The fact that the P. does x16 FAA isn't here or there.

They treat aniso and the 8500 with similar ignorance in reviews.

But then its always been a case of comparing the same features of the competition + speed with the current market leader whilst ignoring the extra features the compeition provides. Was 3dFX now nVidia. I don't see that the reviewers are going to change anytime soon.

[xortam]

Originally posted by 103er-Fan
... methods are totally different to the nVidia method. So you can't even compare them ...

So maybe we're talking about a pippin apple to a Washington apple comparison. Its the reviewers' responsibility to make the reader aware of the differences in the feature comparison. FAA, FSAA super, and FSAA multi are all AA features and they should be compared to each other.

[103er-Fan]

Originally posted by xortam
So maybe we're talking about a pippin apple to a Washington apple comparison. Its the reviewers' responsibility to make the reader aware of the differences in the feature comparison. FAA, FSAA super, and FSAA multi are all AA features and they should be compared to each other.

Sure, but they should have also postet some FAA comparisions, because it is one of the most important features of the P. and I think the customers will like the feature and would like to know more about it, especially how it would compete against the usual FSAA methods.
I don't know which review was it, I think it was Firingsquad who compared both 16xFAA and 4xFSAA of the P with the Ti's 4xFSAA, and Quincunx.

[UberLord]

Originally posted by xortam
Its the reviewers' responsibility to make the reader aware of the differences in the feature comparison. FAA, FSAA super, and FSAA multi are all AA features and they should be compared to each other.

Fine.

You make average Joe User understand the differences

I'll bet they have a hard enough time understanding the basic conecpt of AA in the first place!

Remember, 99% of reviews are for the intellectually challenged. And 75% of statistics are plucked out of thin air