Reading all these Parhelia reviews...

I'm sorry Chalnoth the FSAA quality on Parhelia is unmatched...



shot0012.jpg
 
Doomtrooper said:
I'm sorry Chalnoth the FSAA quality on Parhelia is unmatched...

Yes, I agree, when it works.

With release drivers, it doesn't work properly in all games and all game situations. It is turned off in some cases where it shouldn't be. Hopefully later drivers will fix this.

In the meantime, this FSAA will be all but useless in game games in which it only works partially.

Again, I think that all this adds up to is a video card that would be rather pointless to purchase for most 3D game enthusiasts (Some who have the cash may wish to splurge for the 3-monitor view...). Right now its biggest selling point will be the 10-bit color, which should be important in professional markets.

As for games, it is very possible that the NV30 and R300 will be out before the Parhelia's drivers are up to par (It is also possible that FAA will never properly detect all edges...but we'll have to wait and see for that...).

One shining light for Matrox is that as models and environments gain more and more detail, FAA's shortcomings should become less and less apparent.
 
Doomtrooper> Look at the top of the gate in the Parhelia Shoot, comparing to the GF4 it looks awful, but when it works it's really great.
 
I'm sorry, but this is just nonsense. The Parhelia basically sucks.

#1) It's not fast, so if you want speed this isn't the card for you.

#2) 3D image quality is far inferior to the Radeon 8500. Anisotropic filtering (which is much more important than stupid AA) looks much better on the R8500 and has a smaller performance hit.

I think the bottomline is many people I just not looking at the facts. This card excells in no one area, and for the price it doesn't make sense of anyone to buy. If you want speed, get a Ti4600, if you want 3D image quality get an R8500, if you want 2D image quality and FOUR monitor support (instead of just 3) get a G200 4-head display (and if you want only 2, get a G200 dual head).

All these cards are cheaper and better at what they do than the Parhelia. So much better that even if you are a professional who works at home (the only people who would benefit from being able to game on a professional card), you'd be better off just putting together a second gaming system (which probably makes good sense anyway, since the whole thing won't be totally cluttered with games/professional apps).

I think what's wrong with online forums is that even if Parhelia had turned out to be a steaming pile of dog crap, there would still be a small minority of people who would support it. Parhelia might not be quite that bad, but it's pretty bad.
 
Chalnoth said:
Yes, I agree, when it works.

With release drivers, it doesn't work properly in all games and all game situations. It is turned off in some cases where it shouldn't be. Hopefully later drivers will fix this.

In the meantime, this FSAA will be all but useless in game games in which it only works partially.

I disagree - yes, FAA doesn't yet work completely or sometimes at all in some games, but for that there's 4x SS FSAA as an alternative and we can't be sure yet in which percentage of games it will work flawlessly, it might be like 90%, or 20%, only time will tell. The framerate when falling back to 4x SS suffers, but IQ apears to be similar if not better than the R8500's Smoothvision. With Parhelia there seem plenty of choices to make regarding IQ and speed, more than with any other card, and the drivers will certainly get better and solve many isues with every release. So if price does not matter and you like having a choice, the Parhelia is an attractive card!

What bothers me is the inconsistency of FAA, it AAs some areas of the screen marvelously, but ignores others, the UT2003 shots above demonstrate it perfectly -while the yellow pattern on the floor looks perfect, the roof over the gate is not AAed at all. On the GF4 the yellow pettern looks more aliased, but at least the roof gets AAed too... its "weird" to look at. Max Payne has similar problems, like several other games, I hope that's gonna be possible to fix without loosing much performance ...
 
Brent said:
EgonOlsen said:
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.

There's nothing to show that Parhelia has a higher "minimum fps". This is a perfect example of just grasping at straws to support this product. When it's running almost half as fast as a GF4, I guarantee you that it's minimum FPS is going to be lower. All the marketing nonsense about how this card "doesn't slow down" when in complex situations turned out to be just BS. At 1600*1200 it's too slow to be playable without AA/Aniso... So what if it only takes a 30% performance hit when they're on, it still will be running that much slower. Maybe this is a good card if all you play is the Tomb Raider series of games. :rolleyes:
 
Big thanks to Reverend bringing me back to ground level. I was making same mistake as most everyone else. just looking if it's fastest and not looking if it could be reasonable card for me.

I am with the Parhelia. Definately. Only thing slowing me down from buying a one is money. I don't have enough of it. And hopefully when I got enough for buying a one, we hopefully have better drivers and a bit lower prices. (maybe even the 64MB version which fits even better to my needs.)

Reverend, Thanks! I needed that one. ;)
 
The problems that I fell no one is being honest about.

-The General IQ of The parhelia in 3D is not better than R8500 or GF4.

-FAA does not AA all edges it misses some as well as have compatability issues even in morrowind (trees have artifacts)

-FAA and aniso may be faster than 4XAA+8xAn on GF4 but the performance of both is unplayable for many many of todays games.

-FAA does nothing for textures

-Parhelia's Standard FSAA is to slow in comparrison to the GF4

-Parhelia's aniso is substandard in quality, and has a performance hit

-Quad based Displacement mapping will never be used

-its 10bit color does not have enough Alpha bits for a 4 pipe card

-Its 256bit memory interface is very "sloppy" does not prove out its performance

-It has no occlusion Culling


There is more than that. For anyone to say its simply about pure FPS... I just dont think that is doing a completely honest view of the sintuation. The truth is with this cards stats it shoulf be at *least* as fast as a Ti 4200. ITs not really even close in any form of playability in most of todays games. It just does not add up for a 399$ card at all. Even with the 8500's driver issues it was at least comparable to a standard GF3 at the time of its release. There are benchmarks where it is off by 30-40 FPS. No you dont have to be the fastest, but you do have to be respectable.

Most people are not going to use FAA at 40 FPS when you can use 2xAA+8xAn on a 200$ GF4 with a slightly adjusted LOD and get pretty nice results at 60+ FPS.

This is not even taking the soon to be released Detonators that make big impreovements in Aniso speed. The parhelia Probably wont even win any benchmarks one they are released.

The end result is a card that would have been great 6-8 months ago. Everyone also seems to have forgotten several of the statements that Matrox folks made a few months ago. Statements that lead pepople to think it was a gamers card, that it has top of the line performance and features, that it woulf put matrox back on top.

I think anyone would be hard pressed to say they have accomplished any of these goals. Otherwise S3 lead the industry with the Savage 2000's T&L, and Ati has lead the industry with the best Aniso solution.

Its not just about having *stuff* its about having stuff that works and works well.
 
Hellbender or whatever your name is, I am so darn sick of your depressing posts. An't you more at home at Rage3D? You are truely one of a kind.. :eek:
 
JF_Aidan_Pryde said:
Hellbender or whatever your name is, I am so darn sick of your depressing posts. An't you more at home at Rage3D? You are truely one of a kind.. :eek:
Hey.
Keep it in PM's ok?
 
woah, Hold on.. All my posts are not depressing. I did not rant or rave, I made some very valid poinsts. I just dont understand. I was trying to explain why so many reviewers are havinf troubble being posotive about this specific card. Notice I said posotive statements about ATi and Nvidia.

Please, try to take a step back and see the big picture. Even JC has said much the same thing recently.

I am sorry if i have upset you.
 
Hellbinder[CE said:
]
blah blah...
Maybe I can put a different spin on your opinions.

-The General IQ of The parhelia in 3D is not better than R8500 or GF4.

With the exception of the limited AF, I don't see how they could have made it any better. You can not add things to the scene that are not supposed to be there for the purpose of IQ. Why isn't just as good, good enough?

-FAA does not AA all edges it misses some as well as have compatability issues even in morrowind (trees have artifacts)

Smoothvision does not(supposedly) work like it was intended to, MSAA doesn't get alpha textures, as you can see every AA technique out there has it's problems.

-FAA and aniso may be faster than 4XAA+8xAn on GF4 but the performance of both is unplayable for many many of todays games.

So you are saying that using a GF4 with AA and AF is unplayable?

-FAA does nothing for textures

Again neither does SmoothVision, or MSAA. Anyhow, I thought thats what AF and LOD bias is for?

-Parhelia's Standard FSAA is to slow in comparrison to the GF4

The quality versus the speed hit is subjective. Dare I say blur filter?

-Parhelia's aniso is substandard in quality, and has a performance hit

Equal to GF4 at the level Matrox lets us use. As for the performance hit, nothing is free in 3d.

-Quad based Displacement mapping will never be used

In DX9 it will, but I never expected Carmack to use DX9 anyhow.

-its 10bit color does not have enough Alpha bits for a 4 pipe card

Again this is a limitation Carmack spoke of that I wouldn't mind hearing more about.

-Its 256bit memory interface is very "sloppy" does not prove out its performance

I haven't seen the actual plans for the hardware of Parhelia so I couldn't tell you this one, I am a bit curious as to what you think makes it sloppy.

-It has no occlusion Culling

I thought this was the point of them going for a 512 bit GPU. They knew they couldn't create OC algorithms to compete with ATI and NVIDIA who had been doing theirs for two generations, so they made a fat bus to compensate.

Most of what you have said is subjective, so I don't understand how you can make broad statements about the usefulness of this card. Opinions differ, take that into consideration before being rash with your own thoughts.
 
LittlePenny said:
...
-its 10bit color does not have enough Alpha bits for a 4 pipe card

Again this is a limitation Carmack spoke of that I wouldn't mind hearing more about.
...

I think the limitation is a current DoomIII design issue and concerns the use of stencil buffers. The same issue (but not the same reason) that meant Truform couldn't be used.
 
I really dont want to get into a tit for tat discussion. My reputation is on shaky ground as it is.. I dont want to tic even more people off...

However, i will point out a coupple things about your response.

Yes I am saying pretty much no one plays games with 4x FSAA and 8x Aniso even on a gf4. There are exceptions but they are few.

Smoothvision, indeed has issues. However as I pointed out your features can have issues if your product meets the average performance of its competitors at the same price, or with similar stats.

Benchmarks and comments by those in the know show that the 512bit cpu deos not deliver performance as it should.

The rest of my comments i will leave upo to the judgement of others. Reading them again I stand by them, and think that many others will back me up.
 
That you know of, its only with Quincunx or 4x9 FSAA (which is a total misnomer, BTW, cause it implies 36 sample AA).

Althornin,

It´s 4xOGSS + 9-tap blur filter (gaussian type of filter?), while Quincunx is 2xRGMS +5-tap blur filter. Where did you get that funky multiplication from anyway?


Reverend,

Giving up the good fight is easy. Carrying on is much harder. Despite of what you may feel or guestimate, your voice gets heard.

But if you think that withdrawing is the best choice for you right now, then none of us can really stop you. Good luck and I´ll hope to see you back soon.

Hellbinder,

The only valid point I grant you concerning your whole "across the boards" crusade against Parhelia is that if I personally sit back and think if I would pay 399$ for it, the answer would be negative. 250$ ? Debatable.

I haven´t the slightest idea though where you get all that passion, energy and almost often sounding like hatred for specific boards or vendors at times. Since Parhelia doesn´t sound like much competition to a R200 and even more the R300 I wonder why you actually bother in the end?

How long did it take ATI to reach though, what we see today in R100 and R200?

My reputation is on shaky ground as it is..

You should know why yourself.
 
Here's the bottom line on the Parhelia, in my opinion...

- It will give you pretty darn awesome signal quality...In all likelihood, second to none.

- Although the Anisotropic filtering capability is limited right _now_, it is capable of achieving GF3/4 level. It will surely be addressed in future drivers.

- There is no single Antialiasing method in consumer boards that isn't without some sort of 'con.' The closest thing to perfect I've seen is B3D's overview of 3DLabs older Wildcat board...and you just have a feeling that it would suck major wind in an actual game.

- On the topic of FAA...Although there are some technical issues to consider, I feel that some of the issues are probably driver related...though we will see how it pans out in the near future.

- On the topic of performance, these are about as close to 'raw' as you will probably find. Without a doubt, we will see the performance increase. the question is more a 'when' then an 'if.'

- Cost: Matrox needs to pull a GeForce3 maneuver in order to bring the cost down...Whatever it takes, they need to shave a good $100 off.

- Performance: Right now, on very raw drivers, Parhelia is generally capable of GF4-level performance when you crank everything up. The downside, obviously, is that you do NOT get GF4 level performance when you turn everything off. So the question remains...Which is better?

From where I stand....I would be lying to you if I didn't say that I really _wish_ this weren't the case...

But at the same time, I think about the GF4. True, you can achieve some insanely high framerates when you disable everything. But I simply never do that. I turn things on! Therefore, it is not realistic (my POV) to look at performance from the standpoint of highest-achievable; rather, performance with AA/AF enabled. In that regard, you're back to ~ GF4 level performance. There is great hope that, over the course of the next few months, that will turn into _greater_ than GF4 performance.
 
First, let's pick teams and flip for shirts or skins.

Sharky's benchmarks show it performing favourably with FAA and AF. This is probably my next card. I'm a little worried about the Doom III comments made by JC, so that may change my opinion. But that game is still a year away.
 
Ailuros said:
That you know of, its only with Quincunx or 4x9 FSAA (which is a total misnomer, BTW, cause it implies 36 sample AA).

Althornin,

It´s 4xOGSS + 9-tap blur filter (gaussian type of filter?), while Quincunx is 2xRGMS +5-tap blur filter. Where did you get that funky multiplication from anyway?
I know what it is, Ailuros.
Notice the "it implies"?
If i said hey, this card does 4x4 FSAA what would you think? no funky multiplication here, just how it was stated. 4x4FSAA would imply 16 sample - and 4x9 FSAA implies 36. There was even a thread about FSAA nomenclature here a while back.
And its not 4xOGSS either, is it? Its 4xOGMS + 9 tap filter. not SS, MS.
anyways, i was just nitpicking his nomenclature.

LittlePenny said:
-FAA does nothing for textures

Again neither does SmoothVision, or MSAA. Anyhow, I thought thats what AF and LOD bias is for?
Smoothvision DOES help texture aliasing. Just like any Supersampling form of FSAA.
 
Back
Top