"Trilinear" Filtering in Motion

[StealthHawk]

edit: [H] thread = People are saying the comparison is invalid because there is too much compression on the videos If anything, that would mask the bilinear artifacts :?

Yes, I read the comment. I compressed the files in DivX format, which was much worse in quality compared to Microsoft's MPEG4. It's possbile that they aren't playing the video back at it's full resolution of 800x600.

Anyway, I have a couple of uncompressed videos that I'm moving to the server right now. Each file is around 25MB, which is a couple seconds of video but just enough to see the differences. You'll be surprised that the quality isn't significantly better than the MPEG4 files.

I'll update this post when the uncompressed videos have been moved to the server.


http://www.nvnews.net/temp/1-no-compression.avi - 25.9 MB

http://www.nvnews.net/temp/2-no-compression.avi - 25.2 MB


Probably a good idea to right click on the links and choose "Save As..."

What codec do you need to play these? They don't open with MPC or with WMP for me.

 

[andypski]

Can I ask a question?

Do you play the game with No AF?

Maybe some comparison videos with AF enabled would be more relevant to what settings people would use on mainstream - high end cards, no?

Applying AF will tend to make the transitions happen further in the distance, and hence reduce their visibility, so as you rightly point out this should potentially improve things somewhat.

Of course if optimisations such as limiting the amount of AF applied on texture stages above 0 were also being applied to UT2003 then the transitions on the detail textures would not be pushed back, and you would have the same problems as before.

- Andy.

[EDIT - I was probably sounding unnecessarily sarcastic. I should never post in the morning before the first cup of coffee ]

 

[Reverend]

Do you believe it is right for NVidia to take away the option to use the application's settings for texture filtering?

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.

When you have heard/read all sides of the story and take some time to digest them, it is most definitely for the lead video card reviewer of a media outlet that is amongst the most popular on the Internet to decide if something is right or wrong. In fact, it is that person's duty. Unless that person feels he shouldn't have such a responsibility for a reason. For a lead video card reviewer of such an outlet to think that he/she is not in a position to make and declare their stance/decision on an issue or subject matter means that he doesn't think he has a responsibility, or he knows he has a responsibility but he is either shirking his responsibility, indicates his hands are tied, not doing the job the public expects him to, or all. He should stop treating this as a "hobby" because his "voice" reaches out to millions and has a big hand in determining the shape of the entire 3D industry.

I may not be have the lead video card reviewers job in this website but I will seek an avenue for voicing my opinions (eg. NVIDIA/3DMark03 issue = NVIDIA cheating... NVIDIA/UT2003-filtering-issue = NVIDIA not being honest with reviewers... both expressed in forum posts by me... not on the front page because I don't think that would be appropriate because of who/what Dave is or is seen to be by the public that visit this site... the point is I will express my opinion regardless of the methiod/medium). If I see something wrong, I will tell it like I think it is, in the avenues I think best suits me.

According to what I hear developers say they would really like it that the drivers not change up things like that, and render how they intended the game to be. That is also what the end user is asking for. So it seems clear, most people would like it to use application filtering as default, and then maybe have an option for faster performance with the expensive of lowering iq a bit.

Developer expects their app to be rendered as they requested. Public expects the app they use/play to be rendered as the app/game requests as well as what they specify in their video card drivers' control panel. The public comes to know what the various options provided by the drivers control panel mean through what they find out if they bother to investigate.

I really keep questioning myself with this, wondering why the heck we are still having to look at something like Trilinear filtering on these high-end cards of today (the Trilinear filtering method has been around for a long long time now). We should be concentrating solely on Anisotropic methods and higher levels of it with today's cards, but here we are still looking at an old filtering method, we just seem to not be able to get past it

You should not be questioning yourself. You should be questioning the state of the industry (i.e. 3D industry as well as hardware reviewing industry) and a certain IHV.

 

[StealthHawk]

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. According to what I hear developers say they would really like it that the drivers not change up things like that, and render how they intended the game to be. That is also what the end user is asking for. So it seems clear, most people would like it to use application filtering as default, and then maybe have an option for faster performance with the expensive of lowering iq a bit.

This seems to be what Kyle says. However, not wanting to take a stance on something does not mean you have to hide the issues. The issues should be properly exposed to the public, they can make their own decision after they have all the evidence.

I really keep questioning myself with this, wondering why the heck we are still having to look at something like Trilinear filtering on these high-end cards of today (the Trilinear filtering method has been around for a long long time now). We should be concentrating solely on Anisotropic methods and higher levels of it with today's cards, but here we are still looking at an old filtering method, we just seem to not be able to get past it

Well, you do benchmark without FSAA and AF at [H]. If you think trilinear filtering is so un-important, why benchmark it?

Also, keep in mind that as far as we know, all gfFX cards are affected by this, not only the high end ones!


If you could take the time to share your thoughts on my comments in the post I made on page3(the one above yours), I would greatly appreciate it.

 

[cho]

i have the fraps demo version only

http://www.gzeasy.com/ours/cho/fx5600_trilinear_off_msmpeg_4.avi [44.67,quality,trilinear filter disabled in ut2003.ini]

http://www.gzeasy.com/ours/cho/fx5600_quality_msmpeg_4.avi [44.67,quality,trilinear filter enabled in ut2003.ini]

http://www.gzeasy.com/ours/cho/fx5600_quality_44.03_ad_msmpeg_4.avi [44.03, with AD script, trilinear enabled in ut2003.ini]

http://www.gzeasy.com/ours/cho/9600_trilinear_on_app_msmpeg_4.avi [7.91, trilinear filter enabled in ut2003.ini]

IMHO, it is hard to see the mipmap beam of 44.xx+ quality mode without AD.

 

[Dave Baumann]

Of course if optimisations such as limiting the amount of AF applied on texture stages above 0 were also being applied to UT2003 then the transitions on the detail textures would not be pushed back, and you would have the same problems as before.

One thing you will notice about this behaviour in the 44.03 drivers is that the optimisations also completely overides that default LOD requests of the application. The detail textures have a positive LOD values, so you can see in the Radeon shots that the First mip level has changed very close to the bottom of the screen - this is isn't the case with the 44.03 drivers, and I guess thats becuase with AF applied the transition would be pushed back a little and make it noticable again.

 

[digitalwanderer]

What codec do you need to play these? They don't open with MPC or with WMP for me.

MikeC: What codec do I need for the uncompressed ones? Media player ain't having any luck with 'em.

WindowsXP reports the video codec as Fraps. Do you have FRAPS 2.0 downloaded and installed?

It's fraps 2.0 custom codec, you need fraps 2.0 installed which means I gotta boot up into me XP install to see 'em. (Fraps 2.0 don't work on 98se. )

And a quick standing ovation of applause for Reverend's post on reviewers responsibilities, the man speaks the gospel straight!

 

[digitalwanderer]

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

Y'know, I think that might be one of the biggest bones of contention I have with the site you work for right now...it IS your place to call this "right" or "wrong", and you folks just sit indifferently on the sidelines saying, "Well, we can't do anything about it."

You can, you SHOULD! By not doing/saying anything about it you are endorsing it with your silence, which is WRONG.

The truth will out in the fullness of time Brent, how is your attitude going to look then?

 

[Brent]

When you have heard/read all sides of the story and take some time to digest them, it is most definitely for the lead video card reviewer of a media outlet that is amongst the most popular on the Internet to decide if something is right or wrong. In fact, it is that person's duty. Unless that person feels he shouldn't have such a responsibility for a reason. For a lead video card reviewer of such an outlet to think that he/she is not in a position to make and declare their stance/decision on an issue or subject matter means that he doesn't think he has a responsibility, or he knows he has a responsibility but he is either shirking his responsibility, indicates his hands are tied, not doing the job the public expects him to, or all. He should stop treating this as a "hobby" because his "voice" reaches out to millions and has a big hand in determining the shape of the entire 3D industry.

I may not be have the lead video card reviewers job in this website but I will seek an avenue for voicing my opinions (eg. NVIDIA/3DMark03 issue = NVIDIA cheating... NVIDIA/UT2003-filtering-issue = NVIDIA not being honest with reviewers... both expressed in forum posts by me... not on the front page because I don't think that would be appropriate because of who/what Dave is or is seen to be by the public that visit this site... the point is I will express my opinion regardless of the methiod/medium). If I see something wrong, I will tell it like I think it is, in the avenues I think best suits me.

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

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.

This seems to be what Kyle says. However, not wanting to take a stance on something does not mean you have to hide the issues. The issues should be properly exposed to the public, they can make their own decision after they have all the evidence.

Agreed, and I haven't hidden any issues. If I can prove it I've talked about it somewhere.

Y'know, I think that might be one of the biggest bones of contention I have with the site you work for right now...it IS your place to call this "right" or "wrong", and you folks just sit indifferently on the sidelines saying, "Well, we can't do anything about it."

You can, you SHOULD! By not doing/saying anything about it you are endorsing it with your silence, which is WRONG.

The truth will out in the fullness of time Brent, how is your attitude going to look then?

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.

I suppose I can put my OPINION in there though, I have to make sure to label it clearly as that.

What truth is coming out in the fullness of time? I don't know what my attitude will be in the future, this whole scene is dynamic, and we change with it.

edit: It also seems strange to me that now you are claiming that there is no visible difference between bilinear AF and trilinear AF. Especially when you said this about Performance vs Quality AF in UT2003:

I never said anywhere I can't tell the difference between Bilinear AF and Trilinear AF.

And yet are you now implying that there is no difference between bilinear and trilinear when only 8x AF is used?

I am not implying that, I don't know where you picked that up.

Interestingly enough, you gave the nod to the gfFX when no AF was applied in your UT2003 article. Are you saying you cannot see the difference between video 1 and video 2

I can see a difference in those videos, yes. A few things to note about those videos though. A much slower movement speed was used then what you normally move at in a real game. The difference is noticeable without AF, but when you apply AF, as most people would enabled with these cards then the difference is comparable. 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.

And of course, I am still wondering whether or not trilinear AF was really applied for the r9800 since you configured UT2003 this way:

The goal was to set the game up as the end user would. They would use the driver control panel to enable and disable AF. So on the NVIDIA card we set it to "Quality" and then tested with NoAF, then moved the slider over to 8X and tested. Then for the 9800 we ticked app preference and tested, and then unchecked it and move the slider to quality AF and tested. These are typical setups. Only advanced users know the options in the INI file and what you can do there. For the typical gamer they aren't going to mess with the INI file, they are going to use the driver control panels. That was the goal, to test in a typical gaming environment.

 

[K.I.L.E.R]

Only advanced users know the options in the INI file and what you can do there.

I'm mentally retarded and I know how to configure ini files in any game. :?
Anyone who is "normal" should be able to figure it out in less than a few seconds.

 

[digitalwanderer]

Y'know, I think that might be one of the biggest bones of contention I have with the site you work for right now...it IS your place to call this "right" or "wrong", and you folks just sit indifferently on the sidelines saying, "Well, we can't do anything about it."

You can, you SHOULD! By not doing/saying anything about it you are endorsing it with your silence, which is WRONG.

The truth will out in the fullness of time Brent, how is your attitude going to look then?

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.

I suppose I can put my OPINION in there though, I have to make sure to label it clearly as that.

What truth is coming out in the fullness of time? I don't know what my attitude will be in the future, this whole scene is dynamic, and we change with it.

How you give your opinion is just as important as what your opinion is. I know what your opinions are from reading here, but burying your true opinions in forum posts while spouting PR BS on your front-page is NOT reporting, it's selective reporting.

What truth will come out? You'll have to wait to find out just like everyone else Brent.

 

[Brent]

Only advanced users know the options in the INI file and what you can do there.

I'm mentally retarded and I know how to configure ini files in any game. :?
Anyone who is "normal" should be able to figure it out in less than a few seconds.

I didn't say they couldn't figure it out, I said the typical gamer doesn't know about it, or the settings in there and what they can do. They will use the control panel in display propertities to enable AA and AF. And they will enable and disable Trilinear from inside the game as well as all the graphics settings from within the details tab in the game.

Anyways, thats how we did the article, and we made it pretty clear that we didn't change AF via the INI file.

 

[digitalwanderer]

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.

(bolding mine)

A question Brent, could you say the above sentence without the bolded bit in it? Because "that distracted me from gameplay" turns it in to a hell of a subjective call rather than an objective one on visual differences.

 

[Brent]

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.

(bolding mine)

A question Brent, could you say the above sentence without the bolded bit in it? Because "that distracted me from gameplay" turns it in to a hell of a subjective call rather than an objective one on visual differences.

IQ is mostly subjective to begin with. I do understand what one person sees another may not, everyone's eyes are different and notice things at varying levels.

You have to understand the focus of the article, and then you will see that makes sense.

 

[digitalwanderer]

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.

(bolding mine)

A question Brent, could you say the above sentence without the bolded bit in it? Because "that distracted me from gameplay" turns it in to a hell of a subjective call rather than an objective one on visual differences.

IQ is mostly subjective to begin with. I do understand what one person sees another may not, everyone's eyes are different and notice things at varying levels.

You have to understand the focus of the article, and then you will see that makes sense.

I think I understand the "focus" of the article a lot better than you seem to Brent, or at least a whole lot better than you're willing to admit to at this point.

Again, could you just please answer my rather simple question from above of "Could you say the above sentence without the bolded bit in it?" ? I'd be more than happy to discuss the article's "focus" with you, but I'd really like to know your answer first.

 

[K.I.L.E.R]

I knew it. Brent's colourblind.

j/k

I really did have to pull the CB card out this time.

I guess Brent makes a point.
Considering I believe Quincunx FSAA on hte Geforce 3 is the best mode of IQ that has ever appeared. Not even Ati's 6xAA matches QCX on the NV20. They ruined the QCX IQ on the NV25+.

 

[Brent]

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.

(bolding mine)

A question Brent, could you say the above sentence without the bolded bit in it? Because "that distracted me from gameplay" turns it in to a hell of a subjective call rather than an objective one on visual differences.

IQ is mostly subjective to begin with. I do understand what one person sees another may not, everyone's eyes are different and notice things at varying levels.

You have to understand the focus of the article, and then you will see that makes sense.

I think I understand the "focus" of the article a lot better than you seem to Brent, or at least a whole lot better than you're willing to admit to at this point.

Again, could you just please answer my rather simple question from above of "Could you say the above sentence without the bolded bit in it?" ? I'd be more than happy to discuss the article's "focus" with you, but I'd really like to know your answer first.

Then I would say yes for the most part. There are some instances where their lower filtering shows through and is noticeable.

Actually, I just made a test map for testing filtering in UT2003, a very long hallway with different detail textures on all planes.

I can see a difference between NonAF 5900 and 9800 trilinear filtering quality with movement.

When I enabled quality 8XAF I can notice a distinct mip-map line on the 9800 moving in front of me though. I am assuming because it is doing Bilinear AF on this detailed texture?

 

[madshi]

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

Well, although I don't really agree with you here, I can respect your opinion. But are you aware how HardOCP's behaviour in the last few weeks looks like to a normal community member? Please go with me through the past:

(1) Dave finds out that there seems to be something wrong with the UT2003 filtering.
(2) Two weeks later HardOCP publishes an article benchmarking UT2003 and giving NVidia the trophy for clearly winning this bench without even the slightest hint that there might be a filtering issue needing further analyzation.
(3) Lots of people complain. Lots of posts get deleted/edited in the HardOCP forums. Lots of people get banned at HardOCP, even including Dave.
(4) No further investigation is done by HardOCP, not even an update appears to the benches done in (2).
(5) ATI itself complains.
(6) Finally HardOCP is doing some additional research. But the conclusion is this: "Bottom line is that we are still very comfortable with our recent benchmarks and thoughts".

Not to mention some strange comments made in the forum by a well known HardOCP guy.

This all doesn't look like an unbiased report. You don't leave it up to the community to decide. You are actually benching NVidia's filtering against ATI's filtering and after the analyzation you say you still feel comfortable with that.

For any impartial community member this looks like you have taken a stance pro NVidia here. It seems that you think that what they did is RIGHT. You sure don't come over as being impartial in this story.

Do you see what I mean?

And this is just one thing that has happened in the past few weeks/months. We had the NVidia centric Doom3 benches. The whole 3DMark stuff didn't add to your image as being impartial, either. Then what about other possible cheats? There were some articles which hinted strongly into the direction that NVidia might be cheating in a lot of well known game timedemos. Where is HardOCP's analyzation about this? All this together looks like you would try to protect NVidia. Maybe it only *looks* this way, but that's why a lot of people are a bit angry with you...

 

[digitalwanderer]

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.

(bolding mine)

A question Brent, could you say the above sentence without the bolded bit in it? Because "that distracted me from gameplay" turns it in to a hell of a subjective call rather than an objective one on visual differences.

IQ is mostly subjective to begin with. I do understand what one person sees another may not, everyone's eyes are different and notice things at varying levels.

You have to understand the focus of the article, and then you will see that makes sense.

I think I understand the "focus" of the article a lot better than you seem to Brent, or at least a whole lot better than you're willing to admit to at this point.

Again, could you just please answer my rather simple question from above of "Could you say the above sentence without the bolded bit in it?" ? I'd be more than happy to discuss the article's "focus" with you, but I'd really like to know your answer first.

Then I would say yes for the most part. There are some instances where their lower filtering shows through and is noticeable.

Could you just say "Yes" and drop the "for the most part" please? It's really annoying that you can't answer a yes/no question without adding some form of qualifier to it.

 

[Brent]

oh this is just kick ass

i left the 9800 at app preference and changed the aniso in the INI file

the mip map line dissapeared on the 9800 in my test map

freakin awesome

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