HardOCP needs to be fair

[Ramfart]

A picture is worth a thousand words! The picture tells me what I need to know despite the one half micron degree angle difference. Talk about splitting hairs people!. The guy has both cards to evaluate. Can anyone else talk from that kind of experience here?

Great Job as always guys!

 

[iZero]

it is impossible to get the screenshots at EXACTLY the same place in this game, this isn't like an FPS game where you can make a saved game point

how am I supposed to do the impossible?

all I can do is the best that I can do, and that is what I have done, the angles are close enough to make a valid comparison, end of story

If you can't ensure accuracy, then stop using the test. That's how it usually works.

 

[nst6563]

If you can't ensure accuracy, then stop using the test. That's how it usually works.

OR you can be like some corporate engineers and just fudge the numbers a bit here and there and then when someone tries to call you on it...simply deny the fact and tell them they are obviously mistaken and have gotten their data from an invalid source.

 

[TheGameguru]

It has to be explained repeatedly because it doesn't make sense.

I've said it before, and I'll say it again:

"It's not that we don't trust you, it's just that we don't trust you."

yet you still post here... how bizzare.. If I didnt trust a sites judgement or honesty or whatever I simply would no longer visit or participate in the sites operation. weird.

The problem with your method is that we have to trust your judgement on what is playable. Personally, I think that you're probably a bit optimistic. I am motion sensitive and I get headaches if I play a FPS with framerate dipping to the low 30s. That's why I play UT2K4 at 0xaa/0xaf on my 5900XT - I simply cannot tolerate the framerate loss. That's why we question your method. 30fps may be fine for one player but horrible for another.

Based on what your saying it would then be impossible to do a review the way [H] is doing. Since one could never factor in everyone in the worlds personal preference on what is considered playable. Just like some people can game on LCD's and others find them a ghosting mess.. [H] is doing all they can basically do in judging what is playable and thats to take a baseline average of 30fps (which is considered usually a standard most developers strive for) and reviewing around that as to what is playable. If they are showing the highest resolution/AA/AF setting that they can achieve in a given game and still achieve decent playable settings what is there to complain about?

 

[primea]

We've been screwed over too much by GPU vendors and manufacturers to 100% trust a reviewer. I don't think that [H] has ever been guilty of falsely skewing results, and I don't think that they will ever do so. However, there is *a lot* of money on the table here.

No offense to the [H], but I agree with you on this, you have to be objective. The review is a subjective review, and beauty is in the eye of the beholder. I think Brent tried to be as objective as he can on his subjective review. It is up to you now, to digest what he said, and what other sites said about these card, and hope that they helped you in choosing what card you want.

There is no sense in being emotional about a video card. It's just a tool that we will use, and discard in a couple of years ...

 

[burningrave101]

You also questioned the performance of it, which is why I brought it up. They did to 4x, despite you saying they didnt.

The simple fact is, they usewhat settings are what they think are the highest while being playable. Which looks like 2xAA most of the time in this review. If you dont like their system, there are plenty of others that dont use it.

Back to being ontopic. I thought Madden 2004 didnt work with AA? I know ATi's 4.7's allow you to force AA in Madden now (and other games that dont support AA), when before I believe it just wouldnt work. Does Madden have a problem with AA? If so, perhaps thats why ATi's looks better.

I questioned the performance of it for the IN-GAME TESTING. Not the Apples to Apples. The Apples to Apples comparisons seemed just fine. If your going to compare IQ then you need to use the best IQ settings for each card. 2xAA is nowhere near the best for the 6800.

4xAA + 8xAF is the settings nearly EVERYONE is running with the 6800's includeing me and thats why i know its playable. There is very little performance hit going from 2xAA to 4xAA for the 6800's. There is a much greater performance hit on the X800's though. But since Hard is looking for maximum playable settings then they need to run 4xAA on the 6800 and they can leave the X800 at 2xAA. Its no different then them running an entirely different resolution on the X800XT PE.

http://www.nvnews.net/vbulletin/showthread.php?t=30641

Unlike the X800 the 6800 supports alot of different modes of AA. 4x RGMS is one of the best. 2xAA is an unchanged legacy mode from the NV30's. Hardly anyone runs just 2xAA on these new 6800's.

http://www.hothardware.com/viewarticle.cfm?articleid=550

Here you can see the performance hit from running 4xAA on the 6800GT compared to the X800Pro. The X800Pro takes a much bigger hit.

2xAA/16xAF: Marginal performance Hit from moving from 2xAA/16x Quality AF, Large increase in Image Quality.

2xAA IQ Shot:

http://www.nvnews.net/articles/chrisray_iq/2xAA.png

4xAA/16xAF(RGMS): With Nvidias New 4x RGMS pattern the jump from 2x to 4x is about the same as the jump from from no anti aliasing to 2x. Bringing significant quality improvements at a very marginal frame rate hit from 2x. Highly recommended mode for 6800 users. Please note horizontal edges and how 4x AA effectively removes aliasing.

4xAA IQ Shot:

http://www.nvnews.net/articles/chrisray_iq/4xAA.png

Also, check out the new 16xS mode which is also only available on the 6800's.

16xS The very first 16x mode ever available under a Direct3d Enviroment! It consists of 4x RGMS + 2x2 OGSS. Notice the 16 Grey Geometry samples and the 4 pink texture samples.

16xS/16xAF: The new 16x is exclusive to people with NV40 variant graphic cards. This mode is overkill. It does however provide the absolute best Image Quality. It effectively removes aliasing better than any other mode nvidia offers and also serves to sharpen textures. It does however have dire effects upon performance.

16xS IQ Shot:

http://www.nvnews.net/articles/chrisray_iq/16xS.png

Now of course, i wouldn't run that setting in the benchmarks. But its a great option for games your already scoring really high frame rates in.

Conclusion: Nvidias new 4x mode results in the best balance of image quality and performance. However there is lot of tweakability within Nvidia's drivers, Finding a mode that suits you shouldnt be hard.

And since unlike nVidia, ATI has given NO option to disable ANY of their Trilinear (Brilinear) or AF optimizations. There is no way the IQ is better on the X800's.

http://www.skenegroup.net/tertsi/ut2003/ati_optimizations/

There is an interesting pic.

 

[fallguy]

I dont care about NV news, stop spamming their links here. Its in almost everyone of your posts.

 

[IceWindus]

I come here for the forums and nothing more. I trust other sites for my computer hardware reviews as I don't believe in [H]s way of doing things. Simple as that.

 

[burningrave101]

I dont care about NV news, stop spamming their links here. Its in almost everyone of your posts.

You just dont care about the truth is what you dont care about. Screenshots and facts dont lie. Or do they?

I run 1600x1200 w/ 4xAA + 8xAF in every game i play and the performance is really good while the IQ is superb.

And the reason i post links in all my posts is in case any of you need to read it and see screenshots for yourselves before you can actually get it through your thick heads its true.

And my posts are in no way meant to appeal to you so whatever i post is done without any concern as to what you want to read lol.

The reason i use nV News so much is because there is a ton of people there that have alot of knowledge on the technology of the 6800's and other nVidia cards. There are actual coders and programmers that work with these cards. The majority of review sites just post numbers and pass along info they were told. The graphics cards forums out there are the grestest source of information. Its the actual users performance and opinion that i care about.

 

[@trapine]

Burning(NVIDIA)GRAVE strikes again fella
As already asked could you please stop spamming every thread with your NVnews links? this is [H] forums people come here to see there reviews not NVnews

 

[burningrave101]

Burning(NVIDIA)GRAVE strikes again fella
As already asked could you please stop spamming every thread with your NVnews links? this is [H] forums people come here to see there reviews not NVnews

They aren't reviews. They are links to screenshots and technical overviews of the 6800's AA options.

 

[@trapine]

I run 1600x1200 w/ 4xAA + 8xAF in every game i play and the performance is really good while the IQ is superb

BurningNVIDIAGrave Do you own an X800Pro or an X800XT or have you even Owned a 9800Pro or XT?
Have you benchmarked any of these cards.I and many others on [H] are much more likely to listen to brent and kyles reviews on these cards.Than your constant one sided babbles about Why Nvidia is so much better architecture than ATi.

 

[LiquidX]

Everyone should be forced to agree on everything and anything, I wish this would become law so none of this would happen, no one can dispute or critique anyone.. And fallguy did you get banned from Nvnews or something? That link got your riled up with emotion.

...back on topic.

 

[burningrave101]

BurningNVIDIAGrave Do you own an X800Pro or an X800XT or have you even Owned a 9800Pro or XT?
Have you benchmarked any of these cards.I and many others on [H] are much more likely to listen to brent and kyles reviews on these cards.Than your constant one sided babbles about Why Nvidia is so much better architecture than ATi.

Your free to think any way you want. But you've done nothing to prove your side other then support a review. I have at least gone to the trouble to back up my claims and explaining the reasoning. Something the rest of you are obviously incapable of doing.

There have been a number of users here on Hard trade in their X800Pro's for GT's. There have been a number of end users at nV News with both X800's and 6800's that run the cards head to head in the games they play.

My babble at least has a point and at least has something to do with the technical aspect of the cards. Your constant "Because [H]ard says so" really doesn't prove anything.

 

[Brent_Justice]

concerning 4XAA

I did try 1600x1200 4xaa/8xaf in MADDEN 2004 on both the 6800GT OC and X800Pro, performance was to slow at times to be playable. I found 1600x1200 2xaa/8xaf to provide the best performance in the game.

I use 4XAA, when it is playable. I love 4XAA too, I wish all the games would play smoothly at 1600x1200 with 4XAA, but they don't.

 

[sativa]

yeah the reason i posted earlier about the 2xAA vs 4xAA stuff is because there is a very small performance loss for the 6800, while the visual quality increases greatly.

I find it odd that the very small performance loss associated with going to 4xAA would cross the line to 'unplayable' in [H]'s eyes.

 

[Brent_Justice]

yeah the reason i posted earlier about the 2xAA vs 4xAA stuff is because there is a very small performance loss for the 6800, while the visual quality increases greatly.

I find it odd that the very small performance loss associated with going to 4xAA would cross the line to 'unplayable' in [H]'s eyes.

'very small performance loss' is not entirely accurate

i don't know where you got this

but this is most certainly a game specific behavior wherever you experienced this at

please elaborate on this ‘very small performance loss’ that you experienced

 

[burningrave101]

concerning 4XAA

I did try 1600x1200 4xaa/8xaf in MADDEN 2004 on both the 6800GT OC and X800Pro, performance was to slow at times to be playable. I found 1600x1200 2xaa/8xaf to provide the best performance in the game.

I use 4XAA, when it is playable. I love 4XAA too, I wish all the games would play smoothly at 1600x1200 with 4XAA, but they don't.

There is very little performance loss going from 2xAA to 4xAA on the 6800's. 2xAA is the largest step in performance drop on the 6800's. The X800's however lose quite a bit more performance from 4xAA.

The 6800's were getting an average frame rate of over 50 fps in Nascar 2004 and over 90 fps in Painkiller at 1600x1200 w/ 2xAA + 8xAF. Why wasn't 4xAA enabled? You enabled 4xAA in Painkiller for the X800Pro but not the 6800GT? lol

The 6800GT loses far less performance with 4xAA than the X800Pro does.

Here is one benchmark comparison of the different performance losses for AA on the 6800's.

http://www.nvnews.net/articles/chrisray_iq/comparison/AAgraph.jpg

 

[Shane]

Thats it! This thread has made me decide that my next purchase will be a Matrox ! lol

 

[Brent_Justice]

There is very little performance loss going from 2xAA to 4xAA on the 6800's. 2xAA is the largest step in performance drop on the 6800's. The X800's however lose quite a bit more performance from 4xAA.

The 6800's were getting an average frame rate of over 50 fps in Nascar 2004 and over 90 fps in Painkiller at 1600x1200 w/ 2xAA + 8xAF. Why wasn't 4xAA enabled? You enabled 4xAA in Painkiller for the X800Pro but not the 6800GT? lol

The 6800GT loses far less performance with 4xAA than the X800Pro does.

Please show me where yo have facts to back up this statement.

I did try painkiller at 1600x1200 with 4XAA, it was not playable.

 

[chrisf6969]

In many cases, 2x & 4x AA are practically "free" b/c the CPU/system are the bottleneck, not the videocard. But not in every case.

 

[@trapine]

originaly posted by BurningNVIDIAGrave

There have been a number of users here on Hard trade in their X800Pro's for GT's

Ever heard of the thing called the herd mentality

 

[Brent_Justice]

In many cases, 2x & 4x AA are practically "free" b/c the CPU/system are the bottleneck, not the videocard. But not in every case.

depends on the game, and depends what resolution you are running at

 

[burningrave101]

depends on the game, and depends what resolution you are running at

I posted a new benchmark comparison of AA performance losses in my previous post.

And your trying to say that Painkiller dropped from over 90 fps to a speed that was unplayable from just enabling 4xAA on the 6800? You had 4xAA enabled on the X800Pro and it obviously didn't lose much because the X800Pro isn't that far ahead of the 6800GT in Painkiller. If the X800Pro can do it, i know the 6800GT can.

BTW, why was Far Cry included in the Apples to Apples? It wasn't used in the rest of the test. We NEED apples to apples for each game thats used in the other tests so that we can compare results.

 

[gordon151]

Your free to think any way you want. But you've done nothing to prove your side other then support a review. I have at least gone to the trouble to back up my claims and explaining the reasoning. Something the rest of you are obviously incapable of doing.

There have been a number of users here on Hard trade in their X800Pro's for GT's. There have been a number of end users at nV News with both X800's and 6800's that run the cards head to head in the games they play.

My babble at least has a point and at least has something to do with the technical aspect of the cards. Your constant "Because [H]ard says so" really doesn't prove anything.

Your babble making a point is your own opinion, as I don't find making blanket statements that 4xAA/8xAF is the setting that everyone with a 6800 uses (which is totally ridiculous, as there are many games now that aren't fully playable at that those settings in certain resolutions) and that there is very little performance drop between 2xAA and 4xAA on the 6800 (which is also ridiculous, as the drop is also very dependant on the type of game and even the link you pointed to showed upwards to 45% drops going from noAA to 4xAA) kinda silly.

 

[gordon151]

BTW, why was Far Cry included in the Apples to Apples? It wasn't used in the rest of the test. We NEED apples to apples for each game thats used in the other tests so that we can compare results.

Are you even reading the whole review man? It was the very first subjective test that they did.

 

[@trapine]

As was asked by brent burningNVIDIAGrave

Please show me where yo have facts to back up this statement.

I did try painkiller at 1600x1200 with 4XAA, it was not playable.

We are still waiting for your proof? Show us your benchmarks that you have done?As we all know by now you OWN a 6800GT. And i dare say you have Painkiller?Then it should be no problem for you to fire it up an post some of your testing painkiller 1600x1200 with 4XAA with playable framerates?
Not Nvnews reviews?

 

[burningrave101]

Your babble making a point is your own opinion, as I don't find making blanket statements that 4xAA/8xAF is the setting that everyone with a 6800 uses (which is totally ridiculous, as there are many games now that aren't fully playable at that those settings in certain resolutions) and that there is very little performance drop between 2xAA and 4xAA on the 6800 (which is also ridiculous, as the drop is also very dependant on the type of game and even the link you pointed to showed upwards to 45% drops going from noAA to 4xAA) kinda silly.

Once again we have more people unable to interpret graphs.

Look at the graph again and just look at the performance drop from 2xAA to 4xAA. NOT noAA to 4xAA.

Secondly, i know full well what the majority of 6800 users are running for their settings because if you go into a 6800 thread here or over at nV News nearly every person is running 4xAA because there is very little performance loss between it and 2xAA that the IQ is VASTLY improved.

I myself have a 6800GT and i have not had a problem running 4xAA in any game yet. And i play all my games at 1600x1200 @ 85Hz.

Where exactly have any of the rest of you made any points? You havn't. Your just continuously pointing back to one review to do all the talking for you. Where is all the 6800 users that say they can run 2xAA but not 4xAA???

 

[xphantg0d]

I think the review was based on what they saw in the game and the camera was probably moving. what you see is just a screenshot of that video. so i think they base the review off of what they see when they do it not what the screenshots show.

 

[burningrave101]

Are you even reading the whole review man? It was the very first subjective test that they did.

Sorry lol, i didn't see it. They always stick Far Cry at the end of Software and Drivers, Test Setup.

I'm going to TRY and run some tests myself using 2xAA and 4xAA in Far Cry and i'll post the results in a minute.

 

[IceWindus]

Theres alot of bullshit going on here, and not enough fact.

 

[@trapine]

Originally Posted by burningrave101
There is very little performance loss going from 2xAA to 4xAA on the 6800's. 2xAA is the largest step in performance drop on the 6800's. The X800's however lose quite a bit more performance from 4xAA.

The 6800's were getting an average frame rate of over 50 fps in Nascar 2004 and over 90 fps in Painkiller at 1600x1200 w/ 2xAA + 8xAF. Why wasn't 4xAA enabled? You enabled 4xAA in Painkiller for the X800Pro but not the 6800GT? lol

The 6800GT loses far less performance with 4xAA than the X800Pro does.


Please show me where yo have facts to back up this statement.

I did try painkiller at 1600x1200 with 4XAA, it was not playable

Your claim was you can run Painkiller @ 1600X1200 with 4AA and Brent called you on it! not FARCRY So hows about some of your (OWN) benchies to prove your claim of painkiller?

 

[gordon151]

Once again we have more people unable to interpret graphs.

Look at the graph again and just look at the performance drop from 2xAA to 4xAA. NOT noAA to 4xAA.

Secondly, i know full well what the majority of 6800 users are running for their settings because if you go into a 6800 thread here or over at nV News nearly every person is running 4xAA because there is very little performance loss between it and 2xAA that the IQ is VASTLY improved.

I myself have a 6800GT and i have not had a problem running 4xAA in any game yet. And i play all my games at 1600x1200 @ 85Hz.

Where exactly have any of the rest of you made any points? You havn't. Your just continuously pointing back to one review to do all the talking for you. Where is all the 6800 users that say they can run 2xAA but not 4xAA???

Maybe if we all interpreted internet graphs like you, we would believe there is barely a 9% drop in performance going from noAA to 4xAA and a 4% drop going from 2xAA to 4xAA. I mean, what is with you, do you like making gross generalizations from one source or something? Also is the fact that the drop is different from game to game a lost concept on you?

Secondly, please list the games in which you have no problem running at 1600x1200 with 4xAA/8xAF on (list your minimum and average fps as well please). If we're going to be making blanket statements here, we need at least one source formulate them from.

 

[fugu]

burningNVIDIAGrave

there's no need for that

 

[IceWindus]

there's no need for that

Its way, waaaaay to late for that.

 

[burningrave101]

Your claim was you can run Painkiller @ 1600X1200 with 4AA and Brent called you on it not FARCRY So hows about some benchies to prove your claim of painkiller?

I dont have Painkiller or i would. I'm working on setting up the Far Cry benches now. Brent only ran 2xAA in Far Cry as well.

Painkiller is not going to drop from 90 fps to an unplayable frame rate all because of going from 2xAA to 4xAA. Look at the X800Pro! It has far worse AA performance than the 6800GT and yet it was able to run 4xAA just fine.

How about you guys try and come up with some results on other sites that show 2xAA playable but 4xAA not on the 6800's. You guys are so lazy you just sit back and point to one review and agree with it and not even able to help support it and yet i'm completely wrong even though me and everyone else that actually have 6800's are able to play at 1600x1200 w/ 4xAA in every game we've tried lol.

 

[Xrave]

I dont have Painkiller or i would. I'm working on setting up the Far Cry benches now. Brent only ran 2xAA in Far Cry as well.

Painkiller is not going to drop from 90 fps to an unplayable frame rate all because of going from 2xAA to 4xAA. Look at the X800Pro! It has far worse AA performance than the 6800GT and yet it was able to run 4xAA just fine.

It's going to be different for different computers. Also it's hard to make that statement if you don't have the game/haven't tried those settings for yourself.

 

[burningrave101]

It's going to be different for different computers. Also it's hard to make that statement if you don't have the game/haven't tried those settings for yourself.

Explain how the X800Pro can run 4xAA in Painkiller without a major performance loss and the 6800GT can't when the 6800GT has better AA performance.

 

[@trapine]

How about you guys try and come up with some results on other sites that show 2xAA playable but 4xAA not on the 6800's. You guys are so lazy you just sit back and point to one review and agree with it and not even able to help support it lol.

Hows about you actually show us proof of (YOUR) card running 1600X1200 with 4AA instead of posting Links to NVnews?
You have been sprouting off alot of supposed noladge of why the 6800 is vastly more superior to the X800.and you seem to know this or come to this conclusion from reading NVnews?
So all we are asking is please post some benchmarks of (YOUR) running 1600X1200 with 4XAA in a cross section of games as you claim the 6800GT can?

(edit post) im shore you can post at least 1 or 2 game benchmarks from your own card running 1600X1200 with 4xAA to at least go someway to support your claim?

 

[burningrave101]

Hows about you actually show us proof of (YOUR) card running 1600X1200 with 4AA instead of posting Links to NVnews?
You have been sprouting off alot of supposed noladge of why the 6800 is vastly more superior to the X800.and you seen to know this or come to this conclusion from reading NVnews?
So all we are asking is please post some benchmarks of your running 1600X1200 with 4XAA in a cross section of games as you claim the 6800GT can?

I've came to that conclusion from all the other reviews that have been posted. I've linked to nearly every one thats been available for the 6800's vs X800's. They all show about the same thing. The 6800's are tied for performance with the X800's in D3D while the 6800's are faster in OpenGL and support more IQ enhancements. Thats not from nV News. What i'm getting from nV News is ACTUAL 6800 users. There aren't very many here on HardOCP and if there are they aren't posting any results like they are over at nV News.

And like i said, i'm trying to get some games set up to test with. I have Halo, Far Cry, UT2004, Splinter Cell, Tomb Raider: AOD, Morrowind, Call of Duty, Unreal II, Quake III and a few others that i will try and use for comparison.

Your all quick to judge what the 6800 can and can not do when you dont even have one. I do. And if you will look at the rest of the posts by Hard users with 6800's you will see that most of them mention running 4xAA + 8xAF at 1600x1200 without any problems. There have been a number of them say that about Far Cry.