"Trilinear" Filtering in Motion

[digitalwanderer]

it would be sweet if epic put an aniso option in game, selectable from the details menu

I'll take it a step further and say it'd be sweet if ALL games shipped with in-game AA & AF options, WAY SCHWEEET! (Hey, I can dream can't I? )

EDITED BITS: Brent, am I understanding your post correctly if I assume this is the first time you've actually gotten trilinear filtering truly enabled on the 9800? If so, how does true trilinear compare to nvidia's pseudo-tri?

 

[Brent]

it would be sweet if epic put an aniso option in game, selectable from the details menu

I'll take it a step further and say it'd be sweet if ALL games shipped with in-game AA & AF options, WAY SCHWEEET! (Hey, I can dream can't I? )

EDITED BITS: Brent, am I understanding your post correctly if I assume this is the first time you've actually gotten trilinear filtering truly enabled on the 9800? If so, how does true trilinear compare to nvidia's pseudo-tri?

First time with AF

I just compared both in movement with AF. Firstcoloredmip 1 shows the 9800 pro having more filtering between mip-maps, as we already knew.

In normal view it depends on the texture if you can tell any differences. In my test map there is only 1 texture out of the 4 you can see any kind of a difference, and it is very very slight, you basically have to be looking for the difference. On the other 3 textures I couldn't notice any difference between the two cards at all.

So I would say if you are looking for the difference, you may be able to find it.

But if you are playing a multiplayer game, it won't negatively impact gameplay.

 

[digitalwanderer]

it would be sweet if epic put an aniso option in game, selectable from the details menu

I'll take it a step further and say it'd be sweet if ALL games shipped with in-game AA & AF options, WAY SCHWEEET! (Hey, I can dream can't I? )

EDITED BITS: Brent, am I understanding your post correctly if I assume this is the first time you've actually gotten trilinear filtering truly enabled on the 9800? If so, how does true trilinear compare to nvidia's pseudo-tri?

First time with AF

I just compared both in movement with AF. Firstcoloredmip 1 shows the 9800 pro having more filtering between mip-maps, as we already knew.

In normal view it depends on the texture if you can tell any differences. In my test map there is only 1 texture out of the 4 you can see any kind of a difference, and it is very very slight, you basically have to be looking for the difference. On the other 3 textures I couldn't notice any difference between the two cards at all.

So I would say if you are looking for the difference, you may be able to find it.

But if you are playing a multiplayer game, it won't negatively impact gameplay.

You're not going to give me a straight answer no matter how much I try to get one out of you, are ya?

It's cool, no problem. I'll quit hounding ya for a while and drop it, but it isn't a done issue by any means Brent. (And yes it is the key. )

Glad to hear ya got it working, it looks a hella lot better with it on for me.

 

[Brent]

it would be sweet if epic put an aniso option in game, selectable from the details menu

I'll take it a step further and say it'd be sweet if ALL games shipped with in-game AA & AF options, WAY SCHWEEET! (Hey, I can dream can't I? )

EDITED BITS: Brent, am I understanding your post correctly if I assume this is the first time you've actually gotten trilinear filtering truly enabled on the 9800? If so, how does true trilinear compare to nvidia's pseudo-tri?

First time with AF

I just compared both in movement with AF. Firstcoloredmip 1 shows the 9800 pro having more filtering between mip-maps, as we already knew.

In normal view it depends on the texture if you can tell any differences. In my test map there is only 1 texture out of the 4 you can see any kind of a difference, and it is very very slight, you basically have to be looking for the difference. On the other 3 textures I couldn't notice any difference between the two cards at all.

So I would say if you are looking for the difference, you may be able to find it.

But if you are playing a multiplayer game, it won't negatively impact gameplay.

You're not going to give me a straight answer no matter how much I try to get one out of you, are ya?

It's cool, no problem. I'll quit hounding ya for a while and drop it, but it isn't a done issue by any means Brent. (And yes it is the key. )

Glad to hear ya got it working, it looks a hella lot better with it on for me.

I did give you a straight answer, especially with the two last sentances up there. I don't know how much clearer I can make it.

 

[Dave Baumann]

Who am I to decide if what NVIDIA is doing is wrong or right, it really isn't my place to decide wrong or rightness.

OK, regardless of the technicalities - is it right to tell us one thing but then remove that on an application specific basis?

 

[digitalwanderer]

So I would say if you are looking for the difference, you may be able to find it.

But if you are playing a multiplayer game, it won't negatively impact gameplay.

Brent, the only clear answer I get out of the above could be roughly para-phrased as: "Well, if you had any kind of clue about what to look for in eye-candy you'd notice; but the average brain-dead consumer who is just buying it for buzzwords won't really notice it."

Since that ain't really what you were probably implying, I'm assuming that his wasn't the clearest answer you could give.

It seems simple to me; yes you can tell there are mip lines in nVidia's almost-trilinear setting, or no you can not.

They are either noticable or they are not noticable, it's a straight questiong that deserves a straight answer.

Brent wrote:

When I tested for movement in the article I played real multiplayer games on just about every map looking for mip-map boundary transitions with NVIDIA's Quasi Trilinear, and did not notice any that distracted me from gameplay.

Can you say the above without the bolded bits, yes or no?

One word please, either "yes" or "no".

 

[Brent]

Who am I to decide if what NVIDIA is doing is wrong or right, it really isn't my place to decide wrong or rightness.

OK, regardless of the technicalities - is it right to tell us one thing but then remove that on an application specific basis?

I don't think it is right to say one thing (especially in an official pdf), and then deliver another without explaining that you are doing something different then originally intended.

But how am I sure that it was intentional to begin with. From the way I understand their driver system to work this could very well have been a mistake that wasn’t intended. I think they have shown they have restructured their system and things like this won't happen again. But the proof will be in the pudding. We will have to see how their next official driver fairs, to see if they really did fix their problems.

 

[Brent]

Brent, the only clear answer I get out of the above could be roughly para-phrased as: "Well, if you had any kind of clue about what to look for in eye-candy you'd notice; but the average brain-dead consumer who is just buying it for buzzwords won't really notice it."

Please don't put words in my mouth, I stated what I stated, and implied nothing.

It seems simple to me; yes you can tell there are mip lines in nVidia's almost-trilinear setting, or no you can not.

They are either noticable or they are not noticable, it's a straight questiong that deserves a straight answer.

There is not as straight an answer as you would like there to be.

On some detail textures you can see if the difference if you are looking for it. On others you cannot, it depends on the textures according to my testing.

Can you say the above without the bolded bits, yes or no?

One word please, either "yes" or "no".

Again, it is not as clear cut as you would like it to be. I cannot say a Yes or No, I can say, it depends on the texture, depends if you are looking for problems, depends if it effects gameplay or not.

It doesn't effect gameplay negatively.

 

[Reverend]

I just don't think it is up for me alone to decide what is wrong and right with what a company is doing concerning video cards.

Making a stance or expressing an opinion doesn't mean your stance/opinion is the only one or the only one that matters. What I'm saying is simply that the lead video card reviewer (or "reporter" as you put it) should make a stance.

I sure can't run that company better then they already are.

"Reporters" don't run a company by reporting things. You're being a mite presumptious I think.

Now, there may be times where it is clear cut and one can make those kind of decisions. But I believe the wrong and rightness of what they are doing should be decided by the cummunity as a whole. After all it is the end user that are buying these cards.

So every reviewer should not have a "Conclusion" to a video card review? In every video card "conclusion" I have read, the reviewer has expressed what he thinks is right or wrong about that video card. Why would this be any different? Or does this have more political undertones?

Also, I believe you meant "community", not "cummunity"... the latter refers to the pr0n community.

I see myself like a reporter, report all the facts as they are, let the consumer decide for themselves what they want out of the video card.

I can post my OPINION about what I would LIKE to see though, have to label it clearly as just that, an opinion.

Here's my example of how I treated this NVIDIA/UT2003-filtering matter :

It is my "opinion" by stating that this is a nice optimization.

I made a "stance" by stating NVIDIA was being less than honest with reviewers wrt their filtering explanations to reviewers in their documentation.

I am willing to do this and I see absolutely no ill side-effects by doing this. I see no reason why you shouldn't be doing this other than the reasons I specified in my previous post (responsibility, and particularly hands being tied).

It is my place to report, not to call right or wrong, I believe that to be in the hands of the community. Report the facts as they are, objectively, thats the way I see it anyways.

Would it be right to say that a reporter has reported things objectively if he states that during gameplay he doesn't notice the NVIDIA pseudo-trilin? Or is that subjective reporting? Why not just post screenshots, make some videos, etc+whatever and say nothing (about whether he notices the pseudo-trilin)?

PS. I would like to comment more on your responses but I'd rather go to bed... it's 1.40AM here. Suffice to say I respectfully disagree with the way you appear to approach your position as the lead video card reviewer at one of the most poupular hardware review websites on the Internet. Specifically, I respectfully disagree with what you seem to think a "reporter" in the capacity of a video card reviewer should report on, the appearance that you do not seem to think "reporters" should make a stance (I read a hell of a lot of "stances" made by "reporters" in the newspapers) and the apparent fact that you do not think highly enough of yourself in what you do and in this industry. I may be a nobody in this industry (I don't run a IHV company) but I see no reason why I should not make a stance anyway. Nighty night!

 

[digitalwanderer]

it is not as clear cut as you would like it to be. I cannot say a Yes or No, I can say, it depends on the texture, depends if you are looking for problems, depends if it effects gameplay or not.

It doesn't effect gameplay negatively.

How can sometimes apparent miplines NOT effect a game negatively?

You can say they have a very minor to barely noticable negative impact on gameplay, but the simple fact of the matter is they do have some impact on the games visual quality; yes?

 

[Ichneumon]

But I believe the wrong and rightness of what they are doing should be decided by the community as a whole.

What community are you talking about Brent?

The Community at [H] which is only allowed to express assenting opinions? The community where if you dissent with certain site staff opinions or stances the post is deleted, the member is banned?

What kind of community decision is being made there? Given [H]'s policy on having a community saying you believe the community as a whole should decide something is rediculous.... there is no thorough discussion allowed at [H]... only members which agree with the staffs current stance.


Are you talking about the community as a whole as in the various enthusiast site communitys around the 'net? You mean the ones that are regularly derided by [H] staff as either being "police" or "fanboi" sites? The ones where actual discussion among the members are allowed and points to be made and argued? The ones that [H] pointedly ignores or questions what their motives are?

Doesn't seem to me like there is much "[decision] by the community" going on that in any way affects what gets published on [H]...

 

[Brent]

I respect your views Anthony, and I do agree with some of your points for sure.

hehe, we are like 24hours apart, it is 1PM here

 

[Brent]

But I believe the wrong and rightness of what they are doing should be decided by the community as a whole.

What community are you talking about Brent?

The Community at [H] which is only allowed to express assenting opinions? The community where if you dissent with certain site staff opinions or stances the post is deleted, the member is banned?

What kind of community decision is being made there? Given [H]'s policy on having a community saying you believe the community as a whole should decide something is rediculous.... there is no thorough discussion allowed at [H]... only members which agree with the staffs current stance.

I can't reply regarding that, I have no control over those things.

I mean community in the sense of everyone, end user, audience, the gamer, people in the forums etc....

 

[digitalwanderer]

But I believe the wrong and rightness of what they are doing should be decided by the community as a whole.

What community are you talking about Brent?

The Community at [H] which is only allowed to express assenting opinions? The community where if you dissent with certain site staff opinions or stances the post is deleted, the member is banned?

What kind of community decision is being made there? Given [H]'s policy on having a community saying you believe the community as a whole should decide something is rediculous.... there is no thorough discussion allowed at [H]... only members which agree with the staffs current stance.

I can't reply regarding that, I have no control over those things.

I mean community in the sense of everyone, end user, audience, the gamer, people in the forums etc....

That may be your intent on using the word "community" Brent, but because of your affiliations it tends to imply Ichy's definition.

Simple equation of what [H] means when they say "community decision" it means "what Kyle wants", the bit I want to know about is why [H] feels the need for it to be that way.

 

[Doomtrooper]

On some detail textures you can see if the difference if you are looking for it. On others you cannot, it depends on the textures according to my testing.


Again, it is not as clear cut as you would like it to be. I cannot say a Yes or No, I can say, it depends on the texture, depends if you are looking for problems, depends if it effects gameplay or not.

It doesn't effect gameplay negatively.

This contradicts [H] conclusion in the NV31 and NV34 article, which is total Hypocrisy:

http://www.hardocp.com/article.html?art=NDQ0LDI=

I have seen it myself, and it is VERY noticeable..and ATI cards SHOULD NOT BE benchmarked using quality settings if you now think this is 'ok'.

Otherwise get some glasses.

 

[Bouncing Zabaglione Bros.]

But how am I sure that it was intentional to begin with. From the way I understand their driver system to work this could very well have been a mistake that wasn’t intended.

Probably because Nvidia put out slides (via [H]) championing the exact same speed increase in the exact same detected app. Do you think the slides and the marketing PR surrounding the speedup were also not intentional? Do you think the slides were also a "mistake"?

 

[Myrmecophagavir]

Brent, now that you know the 44.03 don't do proper trilinear in UT2003, and neither does ATI's on certain texture stages if you force AF in the control panel instead of using the game's control, when will you go back and comment on this in the previous articles which used these setups?

You need to go back and comment on the results if you find out they're not a [grimace] "apples-to-apples" comparison. People are still looking at the old articles where you're effectively presenting an unbalanced test.

 

[Dave Baumann]

Brent - ask Kyle. As we've already mentioned here, NVIDIA have already acknowledged that these "optimisations" and are telling editors they sought approaval from Epic on them - does that sound like a mistake?

 

[Althornin]

Brent, WHO CARES if it is hard to see, or if it is only visible in one texture out of four - how is it a fair way to compare two cards performance if it is visible at all?
Just answer that one question please.

 

[Brent]

Brent - ask Kyle. As we've already mentioned here, NVIDIA have already acknowledged that these "optimisations" and are telling editors they sought approaval from Epic on them - does that sound like a mistake?

are they really saying that?

i saw where firingsquad said: "(in NVIDIA’s defense, we’ve been told that they did have Epic’s approval for this optimization)." http://firingsquad.gamers.com/hardware/asus_v9950_ultra_review/page4.asp