Why did Matrox bother?

[typedef enum]

No, I'm not talking like "Why did Matrox bother selling this thing."

What I am talking about is the following review:

http://www.gamersdepot.com/hardware/video_cards/matrox/Parhelia512/001.htm

Why did Matrox even bother to send this guy a review unit? For the love of God...and I have no problem saying this in any forum, under any circumstances...this 'guy' has to be one of the worst people to send a review unit. If you want a guaranteed unthorough review, then I guess you might want to send them a unit for review.

Let's put this into perspective...

Starting with 3DMark 2001 SE from Madonion, we see that the Parhelia turns in ultra-lame results from such an expensive card. The ATI 8500 beats it pretty good, while the Ti4600 doesn’t even break a sweat.

OK, let me refresh an old memory. Check this out here...

http://www.gamersdepot.com/hardware/video_cards/nvidia/gf3/008.htm

This is the guys GeForce3 preview. OK, so the dude is comparing the GF3 to the GTSUltra. Look what he claims...

As we see here, under both 16 and 32bit color, the GeForce 3 leaves the GeForce 2 Ultra in the dust. In fact, there isn't even one test here where the Ultra comes close.

This was a joke then, and it's a joke now. I actually emailed him about this very preview when it first came out, and told me how I felt. Compare these numbers...

16-bit
16x12: 3308 vs. 3753
12x10: 3865 vs. 4353
10x7: 4363 vs. 4895

32-bit
16x12: 2279 vs. 3125
12x10: 3016 vs. 3949
10x7: 3832vs. 4677

The differences are almost negligible in 16-bit, and IMHO, aren't _that_ far off in 32-bit...and yet, the guy said that the GF3 "leaves the Ultra in the dust" and that the Ultra doesn't come close. What? What the hell is he talking about?

http://www.gamersdepot.com/hardware/video_cards/nvidia/gf3/010.htm

Same thing goes for this part...Look @ the actual numbers, and compare them to his own words.

And then his final conclusion. Let me state that he originally stated in this review that there were no "minuses." I emailed him about this, and he finally changed this review because of the cost. I couldn't believe he didn't find at least one...

Even though the majority of this cards features remain to be exploited by yet unreleased DirectX 8.0 titles, the ability to run FSAA at great resolutions with minimal performance loss is almost worth it alone. But wait, there's more. If you do buy one now, you'll have a card that will be ready to carry you well into next year's game titles. So if you're the kind of person that hates upgrading every 3-6 months, and have the money, then grab one as soon as they're on the streets.

OK, so let's revisit his Parhelia review. Quite honestly, even looking @ the numbers in the graphs (never mind the total lack of in-depth analysis), are the differences _that_ great? In some, yes...in others, no.

Here are some tasty quotes...

As with the previous two benchmarks, Matrox’s poor Parhelia gets a beat-down from the competition.

Advancing on the already powerful game engine, Jedi Knight II takes advantage of more advanced GPU features thus putting more of a strain on the graphics chip. Same old song and dance, Parhelia gets its ass handed to it.

Once again, the Parhelia’s low clock speed just can’t keep pace with the others.

And then read the guys conclusion. Just pathetic.

Did anybody see any in-depth screenshots? Nope, none.

Did anybody see a comparison in Image Quality between Parhelia and Ti4600/8500? Nope, not a single one.

Did anybody see a series of MAX IQ benchmarks, comparing the results to the competition? Nope, you sure didn't.

Honestly Matrox, why did you bother? Don't even waste your time/energy sending review units to places like this that have a history of _bad_ reviews. Again, I have no problem "telling like it is" because it simply is the truth.

[TiG]

I concur. A few natty benchmark graphs and a healthy dose of l33t does not a review make.
Especially when the "results" are spoken as gospel, with no mention of game settings/testbed specs/averaged runs to be seen.
Actually the specs may be in there somewhere, I just can't stand reading it through again to make sure. <shudder>

As for the conclusion ... sorry. I can't figure it out either.

[ben2207]

Scott, concur. Unfortunately, therre's way too much benchmarking Quake3 and 3DMark 2001 SE (add JK2, Comanche4, Serious Sam 2 and you have that review) going around in reviews. And that is just disappointing.

They used 2 screenshots from the marketing material to show off Surround Gaming and not a single shot showing FAA , Gigacolor, DVD playback the Reef demo displacement mapping , nothing . Just one shot from Matrox's marketing stuff of JK2 and one of Quake3

Maybe I should wrap up my review , use marketing screenshots of $5000 3 LCD setups and call it a review.

Naw I'd never do that.

Though unfortunately, one of my monitors is on the fritz, my review should be up this week (hrm screenshots showing FAA, anisotropic filtering, Gigacolor, DVD playback , Reef ,and displacement mapping hrm ... )

[Pace]

Once again, the Parhelia’s low clock speed just can’t keep pace with the others.

CORRECT! Parhelia's low clock speed cannot keep me with ATi/NVIDIA - I'm going Matrox all the way baby

PS: Pace&trade;

[Indiana]

Originally posted by TiG
A few natty benchmark graphs and a healthy dose of l33t does not a review make.

Talking in Master Yodas language??

[[GDI]Raptor]

That was a very bad review! They should have tested FAA-16x more. That is a very useful feature for a gamer.

[TiG]

Originally posted by Indiana


Talking in Master Yodas language??

Review, or review not. There is no try.

[Reckless]

His choice of words is not great but he's honest.

With regards to 16FAA, well it may just be the saving grace for this card. Unlike conventional Antialiasing, Matrox’s Fragment Antialiasing works by only smoothing over the edges in a given rendered scene rather than covering the entire image (like NVIDIA’s Quincunx does) which ends up blurring the entire scene. As you’ll see in our benchmarks, Fragment Antialiasing has an extremely low impact on performance and is quite honestly the best looking Antialiasing we’ve seen to date from any manufacturer.

So what's wrong with pointing out that quote as well as what you have already? He didn't provide any screenshots but haven't we already seen hundreds already? He said it was the best looking AA ever seen from anyone - what more do u want. FFS

Surely THIS is why he's so hard on the Parhelia:

Instead, what we ended up with is a pseudo DirectX 9 (Mostly DirectX 8.1) complaint GPU that offers some cool never-before-seen features like a Triple-head gaming display and 16X Antialiasing that looks great, but at performance levels that don’t equate into spending 400 bucks.

The card in his (and many others) opinion doesn't cut it for $400.

I don't care for the review, it's style or content. However, having waited quite a while with my G400Max limping along in games for the most anticipated ever, I was sorely dissapointed to be looking at spending £337 on card that gets beaten by a £125 GF4/4200. Yes the AA is stunning but NOT worth spending an extra £212. As you see, I'm not after outright speed but simply a much better cost/performance ratio. IMO, the card is not worth anymore than £250 as is.

TH is interesting but will ultimately never go anywhere. Until we all have flat panels for the price of CRTs, not many are going to even try it. Besides, TH is a 100% gaming feature (stretched desktop only) and this is a joke. 100% Gaming = 100% speed, err not with the Parhelia obviously When/IF they decide to provide three monitor (individual) support then that'd fix that 'slight' ommision!

Anyways, I've resigned myself to waiting to see what price the 64MB unit checks in at. Maybe by then the 128MB price would have dropped as well (very unlikely) or at the very least teething troubles (drivers, h/w compatability) will have been ironed out.

[rugger]

Shudder, (over the review)

And I though all that L33t talk was limited to IRC channels and slashdot.org. Whoever gives this guy bandwidth should be fired

Oi, gamersdepot.com, if you don't like the Parhellia, send it here, I could find a good home for it.

[Pace]

Reckless: Have you seen any SurroundGaming benchmarks? Wait on them before you declare it a slideshow.

P.

[jkakaley]

One thing that I have seen on every review is that this card is overpriced for it's performance. The thing that they fail to grasp is that you're not paying for speed. If you want raw speed get a GeForce4. What you're paying for is the ability to run games at high resolutions(I'd say 1024x768 and above is high res) with high quality Anti-Aliasing at playable framerates - something that the GeForce4 can't quite do yet.

[Jmkbiker]

Couldn´t agree with you more.

[typedef enum]

Personally, it's a bit of a double-edged sword...

Specifically, I believe it would be difficult, as a hardware reviewer, to wholeheartedly recommend this board. That might seem contradictory, given what I've stated previously.

However, my biggest problem is that although I tend to agree with what the bottom-line conclusion on this product might be, I absolutely disagree with their testing methodology, or lack thereof. To date, I have yet to see a single review, with the possible exception of xbit-labs/Digit-Life, that really gave this thing the full lookover. Virtually every other review has simply showed the card in its worst possible light: bottom line performance compared to a Ti4600 without so much as exploring the Parhelia features and/or contrasting the quality differences of such things as FAA vs. 4x.

Ultimately, I tend to agree that this board is slower than what I expected, and that the likes of NV30/R300 would be more viable...however, if you're looking for the whole package right _now_...don't want to wait...and think that surround gaming/2D quality/FAA is all-that-and-a-bag-of-chips, then this is obviously the card for you.