1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
Like RaceDepartment on Facebook.

Why does AMS have fewer jaggies than rF2?

Discussion in 'rFactor 2' started by Marc Collins, Aug 8, 2016.

  1. AMS should look worse, since it's based on rF1...

    Do these lines in the PLAYER file, not found in rF2, make the difference?:

    Mipmap Adjust Mode="1" // 0 = Disabled, 1 = Clamp, 2 = Bias
    Mipmap Bias="-0.25000"

    If not, what explains fewer jaggies and better in-motion look when all video card settings, resolution, etc., are identical?
     
  2. Without knowing your specific circumstance, I'd guess you are running old hardware.
    When creating a track, to make the texture sharper, you decrease bias. Some rF1 based modders cheated by using extreme values which worked in a very specific circumstance.

    With rF2, because of the way the engine works(developed in house for their specific needs) negative bias of -1 is the absolute maximum you should use, and only then when you have a problematic material on an object such as guardrails/fences.

    You can use gJed to load the track and check the texture bias and use the above setting to counter the extreme bias settings.

    Triple screen users are worst hit right now as you just can't get the grunt to run three 2560x1440@144hz. Single screen though, you see no jaggies at that res with default settings on quality tracks... In fact it is often the most stunning thing I've experienced.
     
  3. I think it's just outdated tracks or modded tracks with wrong mip bias. I think all recent ISI tracks use 0 bias so they should have no jaggies.
     
  4. But AMS doesn't have an override or different approach? Maybe it's just that all of Patrick's modded AMS tracks are done properly?
     
  5. No, new hardware, but that doesn't seem to affect things one way or the other. It is more often than not white lines that exhibit the major problem. It may have something to do with albedo and transparency settings, too? Lots of tracks look gorgeous except for a few anomalous textures--exactly as you point out. Is there not a global override that can address these?
     
  6. By old hardware, I really mean 1080p. I used Nvidia's DSR maybe 2 years ago and saw the massive improvement in quality. Since then I've been running 1440p@144hz.

    Can ISI fix it globally? Possibly but as we know with ISI they don't do bandaids.

    Personally... every 3rd party track that has been released, I'll import into 3dsmax with a custom script that sets all bias to 0, I'll spend maybe 6-10 hours doing touch up work and tailoring things to my taste. I'll run the textures through photoshop with a batch I created to reduce luminance and saturation. In the end I have a hobby I engage in every day and dozens of tracks that allow me to drive around without things bugging me.

    It's the beauty of rFactor. If you are the type of person that 'can't be arsed with all that' then I'd ask why you would be interested in the product in the first place( with rFactor being the same) when you could just buy an off the shelf sim like AC/AMS. The whole point of the sim is that you can make your dreams... or wait for others to do it. Two years ago, I knew nothing of modding and I got sick of waiting (we still are) so I learnt how to do some basic things. Bit by bit it all comes together and now, I can fix anything that annoys me.

    Is there a particular track that is annoying you?
     
  7. You've got to be joking... First, that it would be normal for someone to have to go in and fix these things...and second, that you haven't offered the improvements back to the authors so they can publish an update...or let you publish it.

    One of my all-time favourite tracks is Toban. Look at the metal "crowd fencing" off to the side as you enter/exit (depending on track direction). Why is it full of jaggies when the rest of the pits is not? Look at the stairs railings on the walkways that go over the tracks...same issue. Look inside the grandstands in certain select areas...same issue. I have no idea why ISI missed out on those glaring items among an otherwise beautifully done track. If you can fix those, please do so and I'll see if I can get ISI to do an official update. They have no excuse if the work is already done for them.

    In terms of modders tracks, just look at Oulton Park. The usual white lines issues. Again, I am sure the authors would appreciate the help if you can fix them. Don't ask me for a comprehensivr list because it would include every track and you have probably already seen all the same issues and fixed many of them!
     
  8. It can already be fixed globally by editing player.JSON. You can override track creator's LOD bias values by setting "Texture Sharpening" to 0 instead of the default 5 (auto), I think it's mentioned in quite a few rF2 guides online. This will force any track-specific sharpening off.
     
  9. Has no effect on the issues I am talking about :(

    And Mat would have to be pretty silly to spend all that time fixing issues he could just magically make disappear with a line in the PLAYER.JSON file!!
     
    Last edited: Aug 9, 2016
  10. Well @Mat Holloway was exactly talking about the same issue as me, and it is 100% fixable with the line I gave above. It's btw exactly the same setting as "Mipmap Bias" in AMS, just differently named in rF2.

    Anyhow, no idea how to fix that then, all jaggies complaints I've heard about rF2 have related to the bias setting being wrong.
     
  11. Do you see the specific examples I gave regarding Toban? Does the Texture Sharpening adjustment affect them? It doesn't on my installation, nor does it affect all the other "white line" jaggies that are problematic. Texture Sharpening can affect jaggies, though, by making texture edges "fuzzy" so they are not stair-stepped. But this is not much of a problem in rF2. I'd be interested to see if you can identify the Toban issues--because if you can't see them at any Texture Sharpening setting, let alone all of them, you have some other influence that is helping your picture quality. Also, be aware that 99% of the problem only occurs when the scene is in motion/the car is moving. Static screen shots look fine.

    For another example, look at the video of the new Snetterton release.

    http://www.onwheelstv.com/rfactor/snetterton-300-for-rfactor-2/

    The problem with white lines and the tops of armcos is easily apparent (especially full quality and full screen) in the first two or three minutes.
     
  12. Sure, i'm not well so only around once/twice a day. I'll look into it and report back.

    To be clear, I import to fix a range of issues not just this particular one.

    It is not normal, not in first party content... that I have been aware of.

    If an author knows how to make a track, create materials for it and export the instances with rFactor specific properties, they do not need to be told their bias is wrong. They also don't need to be told their AI is rubbish, the textures are way to saturated and the road mesh is lumpier than cottage cheese thighs.

    *just had a look
    What you are seeing are fresnel reflections(different to the track lines). If I was pressed to have an opinion, I'd say this is one of the final limitations of the graphics engine and certainly one that would be scheduled to be looked at. I don't think you can compensate for this, but you might run cloudier days to minimise it.

    I'd never notice such a thing at speed but if you park yourself next to that fence and watch as the sun moves, you can see it.
     
    Last edited: Aug 9, 2016
  13. Ari Antero

    Ari Antero
    Premium Member

    Nvidia Inspector texture filtering options improves rFactor2 picture quality. I am using settings below and don`t have jaggies or flashing track lines.

    4 - Texture Filtering

    Anisotropic filtering mode > User-defined / Off
    Anisotropic filtering setting > 16x
    Texture filtering - Anisotropic filter optimization > On
    Texture filtering - Anisotropic sample optimization > On
    Texture filtering - Driver Controlled LOD Bias > Off
    Texture filtering - Negative LOD Bias > Clamp
    Texture filtering - Quality > High quality
     
  14. @Marc Collins Ok, I probably see the same thing now what you mean:

    [​IMG]

    I mus say I never noticed these white dots before from the cockpit, maybe because I'm using a relatively small 23 inch screen. I think this is more to do with standard aliasing and specular as Mat said, surely nothing you can fix by LOD bias, as that only deals with flat texture jaggies like white lines. This seems more like an object level problem. I'll try to play more with AA settings to see if it makes any difference.
     
  15. That issue you have circled seems to be generic at all tracks and often affects armco and grandstands, too.

    Mat discusses the fresnel reflection issue (presumably the stairs railings and metal fencing in the pits) at Toban.

    Ari shows that you can Clamp the Negative LOD Bias using NVIDIA Inspector, but what if you are using AMD??
     
  16. Gijs van Elderen

    Gijs van Elderen
    Premium Member

    Those jaggies are from horizontal (or vertical) geometry (mesh). It has nothing to do with AF (texture fitering) or texture LOD bias.

    Use a high AA level.

    Texture sharpening at "0" won't fix jaggies at ISI tracks: Trees, white lines will be more blurry with texture sharpning at "0".

    Just leave at "5".
     
    • Agree Agree x 1
  17. Yep, I tested now with AA level 8 and it mostly eliminated this issue. I was not the one concerned about texture sharpening, but setting that to 0 is an easy fix for those who see texture jaggies on modded tracks, btw it does exactly the same thing as the LOD clamp setting from Nvidia Inspector.
     
  18. Except as I said, setting to 0 has no effect on white line jaggies at modded tracks :(
     
  19. If the modded track has separate lines and road, then simply load it into gjed, set the bias to 0 or -1, export, repack and away you go. If the lines are part of the road texture, it likely is not a track worth any time.
     
  20. A new piece of info that could be useful. I noticed that some of the same texture edge issues I get in the front view of rF2 show up in the rear view mirror of AMS. We know there is less processing happening in the mirror (in both rF2 and AMS), so does anyone know exactly what is not being done in the mirror?