Reading all these Parhelia reviews...

[Reverend]

... has made me wonder why I try to fight the good fight.

I don't particularly mind the comparisons with other available cards in almost all the reviews. The problem has to do with the mentality that "You must be on top or near the top of the frames-per-second chart" is all that matters. It don't matter to these guys that being above completely playable framerates is what really metters, not being the fastest. That and all the IQ considerations that goes with this.

I will be taking leave for a while to get away from something I have a passion for but of which I feel that the "reviewing industry" has become sickeningly drugged by the fastest-is-always-all-that-matters mental state. I hope to come back refreshed, be able to read reviews all the way to the end instead of hitting the X button halfway through... and start, again, to fight the good fight.

Hopefully that time is around the NV30's availability.

 

[PVR_Extremist]

I think the main complaint has been "fps versus price". Personally I think that £330 GBP is too much for anything bar a complete PC system!!!

I hear ya though and hope you feel suitably refreshed to give us plenty decent reviews in the future

 

[Doomtrooper]

I totally AGREE ...well said.

 

[Typedef Enum]

Yep...That's my feeling as well.

The only knock on Parhelia, from my POV, is the pricetag. If Matrox could knock about $100 (or close) off the $399 sticker, then I think you would have a much more compelling product.

That aside, look @ the image quality side of things...Sure, it might not be John Carmack's ideal chip for Doom III, but the image quality is 2nd to none...Furthermore, the AA is, by far, the best looking thing available...

But once you turn everything up...even with the immature drivers, it _still_ will beat pretty much everything out there...AND with better IQ.

The only other thing that sorta' dissapointed me was the AF limitation...but they say that it's limited within the drivers, so we should see this addressed in the near future...Performance can only go up from here.

 

[EgonOlsen]

But speed is a key-feature and the card isn't a budget card. So i expect it to perform at least on par with other products in that price range. If you buy a high-end car (not card!), you would also expect it to go at least as fast as other high-end cars. A Dodge Viper limited to 130Km/h won't be accepted by anyone even if you can't drive much faster nowadays. The argument "hey, it drives fast enough for daily use" doesn't count for me.
The same is true for Parhelia.

 

[pascal]

If you have time read this review (got the link from nvnews): http://www.gamepc.com/labs/view_content.asp?id=parhelia128&page=1

Parhelia is a tremendous leap forward in chip design for Matrox, and for the entire industry in general. Working from scratch without a hint of G450/G550 logic inside, the Parhelia graphics processor features some incredible features not seen in any other graphics processors to date. Features like a 512-bit GPU, and 256-bit DDR SDRAM memory are industry firsts, long with new board features like tri-monitor configurations, 10-bit color rendering, and dual-DVI connectivity. Parhelia is a bold statement from Matrox, who is once again showing that they have what it takes to compete in this market, no matter what their history. On paper, Parhelia has every feature that the competition has, and more.

Maybe they dont have the fastest card on earth in all benchmarks but they are pioneers with the 256bits cards for consumers.

See this Benchmark: http://www.gamepc.com/labs/view_content.asp?id=parhelia128&page=7
or this: http://www.gamepc.com/labs/view_content.asp?id=parhelia128&page=8

Quality is really important. Have a relaxing time.

 

[KimB]

I think there's more to be disappointed about with the Parhelia than just performance. The anisotropic quality leaves much to be desired. There is no fallback multisampling FSAA mode, leaving FSAA largely useless unless FAA works, and it doesn't work all the time currently.

Granted, except for the anisotropic filtering, I see nothing that can't be resolved with better drivers, and I am pretty confident most, if not all, of the performance and FAA issues will be resolved before long. The problem is that by that time, we'll have the NV30 and R300. Somehow I doubt the Parhelia will do well against them in the least.

 

[Brent]

But speed is a key-feature and the card isn't a budget card. So i expect it to perform at least on par with other products in that price range. If you buy a high-end car (not card!), you would also expect it to go at least as fast as other high-end cars. A Dodge Viper limited to 130Km/h won't be accepted by anyone even if you can't drive much faster nowadays. The argument "hey, it drives fast enough for daily use" doesn't count for me.
The same is true for Parhelia.

I like this analogy myself...


I think the point most are dissapointed about is the speed you get (without AA and Aniso) at the price that this card costs. The performance of the Parhelia without AA and Aniso is on the level of GF3 Ti 500 performance, which is definitely not at the same price range as the Parhelia is at. For $400 you get the performance of a $150 card.

However, like you said, it's not just about high FPS, you have to look at all the aspects of the card. This card is being pushed for high end people looking for the best 2D IQ as well as a full feature set. With AA and Aniso this card is faster then a GF4 Ti 4600 and 8500. This card is no doubt all about IQ.

Playable framerate is still relative. What one enjoys to play at another may not. The problem is the Minimum framerate in gameplay. While we may be cruising along at a nice 60FPS in one scene it may jump down to 30FPS in another scene or if a battle ensues in a first person shooter, it's this DIP in framerate that our brains notice and is that which annoys us.

Maybe instead of Average FPS we need to start focusing on the Minimum FPS in a game or benchmark. The higher the Minimum FPS the better. The game will be smooth throughout gameplay.

Also, it stands to reason that if we have an insanely high average FPS that as video cards become more powerfull, pushing up that average FPS then the Minimum FPS will also rise in turn. We have seen this with Quake3 for example, or any old games. When it was first released it was choppy on some vid cards, the minimum FPS would dip way down. But now, with faster video cards the Minimum FPS has come way up that now we can play the game in any resolution at a high frame rate which means smooth gameplay.

So that IMO is still important, because what we want is to be able to use a high level of AA with Anisotropic and all features in a game turned on and still maintain a playable framerate througout gameplay. These features are important, and I want to be able to use them, at their max settings, and still have playble frame rates at high resolutions

 

[Doomtrooper]

Firingsqad did the best review IMO, they did show the cards IQ which is overlooked so much. Anandtech and Tomshardware were the creators of FPS benchmarks overlooking IQ constantly.
This card matches/outperforms the fastest FSAA card presently with much better IQ and probably beta drivers, I don't see where all the poor comments are coming from.

I hope Ti4600 owners didn't buy a $400 US or $700 Canadian video card to run it without any eye candy like FSAA and AF. IMO that card shouldn't even have those features turned off by deafult.

To me the selling point on this card is FAA or another FSAA card, the price needed to be lower to be competetive though..$299-$350 US.

 

[maskrider]

At first, I am very worried about my wallet with its launch, but reading all those reviews made me wonder, really made me wonder on how it actually performs with FAA16x on (I am a FSAA fantic, but most reviewed games were not the ones I played), I think I will not buy it until its price is around US$300 (I have never wanted to buy an over US$350 card, except the super tempting 3dfx Voodoo5 6000 at that moment)

FAA16x is a good thing to have, but my major entertainment DVD (hardware accelerated - iDCT, which I was spoiled by ATI - software players just can't match its quality) is not as good nor available on any other hardware up to date (no GF4MX, please).

Hope the reviews later on will have more on the picture quality (that I most cared) than numbers, numbers and numbers with games that I don't play.

 

[Evildeus]

Well i disagree with Doomtrooper on this. I think the Firingsquad review is quite low. I was disapointed with it.
On FSAA, there's no hint on the drawbacks:
Parhelia

GF4

Sure the FSAA 16* looks tremendous, but when you see the flaws its disapointed. But time will tell us on this.

On AF, well 2* is not worth, even if there's a bug or if it's a future.

So right now, on quality, i agree and disagree with Reverend.
I think the point is: Quality at what cost?

If there's a standard, let say >50 FPS in minimal peaks, then the best quality is all i want. But if it's quality at the expense of playbility (?) then there's no way for me to buy the card.

On the Parhelia, i would say, yes the quality is their, but due to the price it does not worth it. I can't put 550€ in a card for quality at the expense of being playable (FPS) and a card that will not last long in my computer.

I've just buy a GF4 TI 4200 for 215€, sure, not the same quality, but i'm not sure i'm willing to put 300-350 bucks just for some little more quality and that will last as long as my GF4 TI 4200 in my computer.

Sorry for my english, hope you did understand my point, and i'm looking for the review by B3D .

 

[Doomtrooper]

My only issue with screen shots on the Geforce 4 is the screen shot is not ALWAYS capturing what is being seen on the screen. I refer to this link here:

http://www.beyond3d.com/forum/viewtopic.php?t=1306

 

[Randell]

From my point of view although it competes very well with the gf4 a lot of the time when AA is enabled, it is not AA that looks usable at 1024x768 from a perfromance point of view with upcoming FPS games (e.g. UT2003, either on the Gf4 or Parhelia. I know FPS's aren't the be all and end all, and RPG/RTS's/Flight Sims and racing sims can all benefit from either or both triple head & surround gaming, but the I play FPS's online as well, and I want my next upgrade to give me playable AA in UT2003 onwards with at least a 4xAA solution.

If reviews were not obsessed about FPS raw performance numbers and looked at what else it offers to a gamer who plays a gamut of games, works at home, IQ, artists, etc etc. It may appeal to someone who cant afford a Wildcat but wants multiple monitor support, colour precision and decent gaming performance, and then it looks cost effective (at the moment).

 

[Freon]

A $400 card should be close enough to the top of the heap to inspire the confidence that said money was not wasted 1 or 2 years after purchase. So far I have yet to see anything so confidence inspiring that I'd drop $400 on one.

"You must be on top or near the top of the frames-per-second chart" is all that matters

A card must compete well in its class, which means being "in the running." It's not all that matters, but if a card is getting smeared in the performance arena by cards costing $75-200 less, you're going to have to look damned hard to make that up elsewhere. With the lack of decent anisotropic (an effect which causes the *entire screen* to look objectively better), reviewers are left scratching dirt. Performance is qualifed by price.

Reverend, have you ever thought that maybe its not the mentality, or the industry, or the reviewers--maybe the Parhelia just ain't that good?

*note that this post is coming from the perspective of an enthusiast gamer (the only market share I can imagine buying one). If its supposed to be a workstation card, knock yourself out.

 

[Sharkfood]

I've come to the conclusion that the internet portion of society and 3D hardware will always be predominantly the agenda laiden, uneducated, cookie cutter, dry benchmarks with no analysis ninnies... and the droves of people that like to go around just simply subscribing to their mindset and repeating it almost verbatim without ever even using/having experience with the product. (examples can be found here, actually. hehe).

It's relatively unimportant how this plays out. Some sites do it for reasons of ignorance and lack of understanding, others do it because they are either pro-IHV of a competitor, or anti-IHV of the IHV being reviewed.. and even more do it for no reason at all. But like I said, it's unimportant in the grand scheme of things. The fact that some 16-year old with a web host, FrontPage and some graphs is going to emphasize fps and not delve into "true" performance, nor have the facilities to try and explain *why* nor care is indicator enough of how valid such "data" is. And the hordes of 16-year olds that then go amuck on forums and selectively express "extrapolated" opinions from this "data" when they have never even used the product just raise more red flags.

The bottom line is- while there not be enough valid, reasoned, cross-checked information to make any kind of buying decision in today's number-numbness laiden, cookie cutter website reviews, and what's not emphasized in reviews is usually the most important part, consumers have many options to help them squelch out the noise and make a better decision- which has always been to simply try and use the hardware.

 

[LittlePenny]

Playable framerate is still relative. What one enjoys to play at another may not. The problem is the Minimum framerate in gameplay. While we may be cruising along at a nice 60FPS in one scene it may jump down to 30FPS in another scene or if a battle ensues in a first person shooter, it's this DIP in framerate that our brains notice and is that which annoys us.

Maybe instead of Average FPS we need to start focusing on the Minimum FPS in a game or benchmark. The higher the Minimum FPS the better. The game will be smooth throughout gameplay.

I can't help but think how this was a feature not hyped by many of the online reviews for the Kyro. Although I do wonder what the Kyro2SE's software TnL would have done to the steady frames.

 

[KimB]

My only issue with screen shots on the Geforce 4 is the screen shot is not ALWAYS capturing what is being seen on the screen. I refer to this link here:

http://www.beyond3d.com/forum/viewtopic.php?t=1306

That's only with Quincunx or 4x9 FSAA enabled. It's not like you can just throw out all GF4 screens because some of them don't capture what's displayed.

 

[ben6]

Reverend, while we may not see eye to eye on 100% of the things in reviews or hints we give , it would disappoint me greatly if you decide to take a leave of absence.

I share your opinion on Parhelia reviews . In fact I've held that reviews have focused too much on fps in particular games for a long time now. While I'm not as jaded as I was 2 years ago when I wrote my first review (Voodoo5 5500) , and think benchmarks have a place in reviews (which is why I've included them when I didn't before) , they're not the end all to be all of what a review should be.

This is the reason, though I have an opinion on the Parhelia, I will be going the extra mile with my review and taking my time instead of rushing it out the door in 1 day as at least one person has suggested (shame on you typedef enum!!)

If you do decide to take a leave, I hope you do it for the right reasons and not because of what others do with their reviews or the opinions formed because of those reviews.

In any event, you will be sorely missed by many if you take a leave, including myself.

Ben

 

[Althornin]

That's only with Quincunx or 4x9 FSAA enabled. It's not like you can just throw out all GF4 screens because some of them don't capture what's displayed.

That you know of, its only with Quincunx or 4x9 FSAA (which is a total misnomer, BTW, cause it implies 36 sample AA).

Rev: Who cares what the cookie cutter review sites post? I guess some people do, but dont bail out - show em how its done, instead!

I think that lots of performance issues will be ironed out in the next 6 months - just like they were for Radeon 8500 and GF3 (took about 6 months after each release to get up to par).
However, no reveiw site can say "buy parhelia! it MIGHT be better in 6 months"...

 

[KimB]

There is no realistic possibility for any screenshots to display anything but what the screen is showing when FSAA is not in use.

Additionally, the 2x, 4x, and 4xs screens have all been shown to accurately depict what is seen in-game.

Therefore, the GF4 screens are accurate, unless they are taken with Quincunx or 4x9.