The default setting in nvidia control panel shows worse image quality in BF4 for him on his Titan X. In order to fix it, he has to force highest quality settings through the control panel which drops his performance ~ 8% by my eye.
So is that setting off by default?
The default setting in nvidia control panel shows worse image quality in BF4 for him on his Titan X. In order to fix it, he has to force highest quality settings through the control panel which drops his performance ~ 8% by my eye.
and has he confirmed this behavior in any other games? this is kinda fishy, especially if alot of the tech sites were affected by this.
No, just BF4 so far though others have said they feel like they see similar things in GTAV (apparently it's a known issue for nvidia in this game? I don't play the game) and there's a project cars video that seems to show the same, but again, BF4 is the only one he investigated as that is the first video he posted.
Gregster found out what the problem was:
http://forums.overclockers.co.uk/showthread.php?t=18679713&page=7
The default settings/profile of Nvidia control panel don't produce the best quality, at least in BF4, only with "prefer max quality" AF was working as it should at the cost of ~10% performance hit on a Titan X.
I wonder how many more games this happens and how many reviews are done with the default setting in the control panel or profiles.
The updated video with both recorded at the same quality shows the exact same issue as before with the TitanX, a clear lack of AF, texture detail, and possibly draw distance of objects. This has nothing and never had anything to do with the color or gamma settings or "black crush". This is simply that his TitanX video shows lowered LOD and there are very clear screenshots from the video that point this out. You can watch the video and see it clearly. This could very well be an issue with his setup or how he recorded, but the difference is clear.
If you really did read the thread, then to try and call him out is flat out tasteless. The only reason he did the videos is to give back to the community and he had no intention of starting any of this, it was other posters who saw the video and noticed the quality difference. The OP didn't want to be involved in any controversy and that's why he was reluctant to add further material to the discussion.
Lastly, there are only a couple posters (as expected) who are trying to make this into a hit piece against nVidia, the vast majority of people have simply noticed what is clear in the video and are interested to look into it further. Again, this very well could be an issue with a very simple answer, but it deserves a little investigation. Those trying to throw out all these red herrings as to make it seem like it's all a farce are either being ignorant or dishonest.
That is a good question that reviewers should really look into. I just checked my NV Panel and it set to Quality by default. (New install, haven't touched anything in NV Panel).
I don't know how to make it more clear, this is what people were noticing.Are you kidding me? AF doesn't affect texture detail or draw distance in close objects. People were crying the NV side was mixing textures, lines were painted ONTOP of ground textures, objects were missing in the distance. Even the two images provided only saw AF set to either 4x or off. There are no textures or objects missing.
Take the same image and adjust the gamma and you'll see objects missing and textures gone because they'll wash out into surrounding textures/objects.
Color settings had nothing to do with it and his control settings comparison video confirmed this. Even if they had, what you're suggesting is that the missing details on the Titan X were because the Fury X color settings would cause the Fury video to lose detail. That makes no sense. The original issue of recording quality had been addressed very quickly and the updated video where they were both recorded with the same quality continued to show the same issues.I pointed out a major flaw with his first video. An issue that I've showed through various examples that can AFFECT image quality negatively, more so than lower AF. I don't have an account there and I've asked the people posting here if they had one there to ask him to post an unedited video (ie the colors untouched). The end result would be more similar to this video, and magically "textures/objects" still aren't missing.
Yes, please show me an example where having 1 setup suffer from "black crush" causes the other setup not suffering "black crush" to show less details.I'm not even referring to this place, I visited the four other threads where this was discussed. You want me to get you a few charming quotes? And no, the majority of people DIDNT even address the glaring issue of color representation. An issue that can make something like WORSE, should I post more examples?
You keep going back to this when it has nothing to do with the issues involved. This has been confirmed by his video comparing the Titan X against itself with higher nvidia control panel settings.One can not discuss IQ issues until the issue that the original video that sparked this is resolved of it's most glaring issue - one side had the colors adjusted which created other issues. It was no longer a 1:1. People aren't saying now "look, the lines are over the paper textures LOL" now its a more clear "something is up with anisotropic filtering, what gives?"
Not removes objects (again, more than one issue) but umm. . .Unless, you're now going to go on record saying "lowering AF reduces texture quality and removes objects."
Wikipedia said:In 3D computer graphics, anisotropic filtering (abbreviatedAF) is a method of enhancing the image quality of textures on surfaces of computer graphics that are at oblique viewing angles with respect to the camera where the projection of the texture (not the polygon or other primitive on which it is rendered) appears to be non-orthogonal
and has he confirmed this behavior in any other games? this is kinda fishy, especially if alot of the tech sites were affected by this.
Most tech sites when doing reviews have set settings anyway!, You know, AA/AF etc....I dont imagine they bench games with default settings.
10% performance just from AF low -> high alone is a big deal. What about lod distance?
Quick cut and stitch I put together for comparison. You can see the Titan X issues from in game screenshots are fixed and look (quality wise) just like the Fury X samples posted previously.
10% performance loss with the image quality settings that should have been ran on.Quick cut and stitch I put together for comparison. You can see the Titan X issues from in game screenshots are fixed and look (quality wise) just like the Fury X samples posted previously.
10% performance loss with the image quality settings that should have been ran on.
The benchmark numbers on AMD vs Nvidia cards are unfair RIGHT NOW!
I'll jump to conclusions until reviewers debunked it with proper image quality comparisons and benchmarking with proper settings.Be careful not to jump to conclusions.
http://forums.anandtech.com/showpost.php?p=37533029&postcount=36
Some vendors have been accused of "cheating" at benchmarks doing things that give much higher benchmark numbers, but make things worse on the actual likely workload.
The problem is they don't mention what settings they use in the control panel. Quite a few don't even have a "testing methodology" section and don't even show which sections of games they are testing. It's quite a drop in standards. In the past, reviews came with image quality comparisons.
In my experience, AMD and NV look the same when they are at the same settings, there is not one that stands out different over the other.
Driver defaults on all cards, every brand, all the time, we do not change or touch anything in the driver control panel prior to testing.
It is up to AMD and NVIDIA to provide the kind of experience they want out of their cards to gamers. It is our job to evaluate that experience. We take that, at its default, and evaluate the gameplay experience. If we were to change anything in the driver we would be artificially changing that experience and creating an unfair comparison.
To keep it fair, we leave the driver settings alone, and at default, on both AMD and NVIDIA. If we detect some measure of ill looking image quality we bring this to the attention of AMD and NVIDIA.
10% is pretty significant, that's the difference between the 290x and the 980 being 1:1 and the Fury X being neck and neck with the Ti and occasionally tagging TX.
GTA5 needs exploration.
wouldn't Default settings make the in-game AF setting take precedence?