Released DX11 and VR Open Beta, Is Here!

Discussion in 'News & Notifications' started by Christopher Elliott, May 1, 2017.

  1. mixer61

    mixer61 Registered

    Joined:
    Dec 26, 2011
    Messages:
    553
    Likes Received:
    149
    Man thanks, Emery !
     
  2. Marc Collins

    Marc Collins Registered

    Joined:
    Jan 11, 2012
    Messages:
    3,159
    Likes Received:
    162
    OK, did some more testing.

    By applying global setting of supersampling, I was able to confirm that all the shimmering and flickering shadows that look on my rig exactly as bad as they did in DX9 are "cured" by the supersampling, just as they were in DX9. Unfortunately, the frame rate is also halved, just as it was with DX9, so this is not an acceptable result. I hope shadow resolution and performance optimization is being worked on right now.

    The shimmering and flickering shadow issues are effectively unchanged from DX9, but a second more alarming issue is new (at least to me). The "running shadow" problem we all know about is now vastly more pronounced and affects huge and distant objects like grandstands, not just armco. I only noticed it on armco in DX9. But, on my rig at least (AMD GPU with ultra-wide monitor), the problem is only occurring in the extreme left and right portions of the screen. These are the same ultra-wide edges where the former image remains when you go back to the UI. As if the part of the screen that is wider than the UI is somehow being processed differently (or not reset) than the central part. I can see how this would happen when you return to the UI, but something is also happening in the main view, too. If I use my controller to "look" towards the glaring "running shadow" transition line (where for example on one side of the heavily jagged line it is sunny and the other it is shady, even though both sides are under a shaded grandstand roof), the problem disappears because that part of the view is now in the middle of the screen instead of in the ultra-wide edge. If I release and centre my view, the problem reappears in the edges of the view only.

    As mentioned before, I have every setting absolutely at max because I thought that might avoid these sorts of problems that are usually caused by LOD reductions or constraining the quality.

    Would love to know if these are AMD-only issues; whether they are known and being fixed; or what we're supposed to do to get shadows that match the quality of the rest of the scene.

    Also, I am still seeing excessive bloom here and there while using only updated S397 content. For instance, on the Brabham BT20 tach needle. I thought that was only supposed to happen with DX9 content? Or is it just further global lighting adjustments needed?
     
    Alex72 and Emery like this.
  3. Richard729

    Richard729 Registered

    Joined:
    Apr 24, 2016
    Messages:
    28
    Likes Received:
    9
    I'm seeing very bad running shadows at Monte Carlo 1966 2.1. It's very noticeable on the buildings on the start/finish straight and up the hill to Massenet.
    GTX 970
    2,560 x 1,440

    I started a thread in Technical & Support:
    https://forum.studio-397.com/index....raw-distance-is-very-small.55770/#post-894622
     
  4. FuzzyFassbender

    FuzzyFassbender Registered

    Joined:
    Apr 26, 2015
    Messages:
    100
    Likes Received:
    42
    I use a single UltraWide 3440x1440 or triples @ 5780x1080. For my Ultrawide I can run all settings on highest and maintain 80-90 FPS average with 17 AI (Sunny/clear day, less for weather and night), for triple screens I need to turn things down quite a bit. I don't use NVidia Inspector for rF2, I think the recommendation is to use the built in settings otherwise you may run into shader issues. My hardware: i7 3770 3.4/4.0 GHz, GTX 1070 8 GB, 16 GB ram.
    upload_2017-5-19_19-33-15.png
    upload_2017-5-19_19-33-35.png
     
  5. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,578
    Likes Received:
    1,071
    I just discovered Palm Springs
    -nice layout I wonder if this track will be included in the dx11 upgrade ?4
     
  6. 2ndLastJedi

    2ndLastJedi Registered

    Joined:
    Nov 3, 2016
    Messages:
    1,873
    Likes Received:
    1,198
    I have a i5 6600k and GTX 1070 and have all maxed but PP off or low and get a very solid v-sync at 3840x2160 .
    But i've never seen benchmarks for a GTX 1070ti (lol) but assume it to be better than my 1070 so just max it out and see then lower accordingly , i wouldn't start low and go high because it would be able to run max or near settings depending on fps required .
    Also i dont think inspector changes have any effect on visually but still hitting performance as if it was atleast for me thats what ive found .
     
  7. 2ndLastJedi

    2ndLastJedi Registered

    Joined:
    Nov 3, 2016
    Messages:
    1,873
    Likes Received:
    1,198
    For me personally i think the Sim actually look a whole lot better the first time i fired up DX11 but ran not so well , but still was able to manage high settings in 3840x2160 but as the optimization has evolved the visuals have gotten worse . I have a feeling that the way the optimisation is working is by reducing all the good stuff that made it look fantastic initially by slowly adjusting the heavy hitting things (I guess thats what optimisation is but it's really affecting that awesome look from the beginning) .
    Flicker and shadow crawl imo have gotten worse . Maybe its just me adapting to the new look but i don't think that is the case .
     
  8. bwana

    bwana Registered

    Joined:
    Dec 25, 2011
    Messages:
    2,139
    Likes Received:
    1,040
    Its already been DX11 ed. ;)
     
  9. Marc Collins

    Marc Collins Registered

    Joined:
    Jan 11, 2012
    Messages:
    3,159
    Likes Received:
    162
    I have been away, so am really coming into it for the first time and just comparing to DX9. But overall, the visuals are improved, but not particularly impressive. I just ran a session at NOLA at 9:45 am where AI had their headlights on just because it was cloudy. Not even rain.

    I haven't tried AMS yet, but I hope the time zone was fixed so it isn't dark there at 8 am local time.

    Many, many glitches and issues. I'll try Portugal next to see if the half dozen major reflections and z-fighting issues have been addressed, but I am not holding my breath. [EDIT: some things fixed, but many obvious ones not. Transparency issues that are new with some trees. Shadows for objects that you cannot see.]

    After 6 months, things still seem a bit chaotic and rushed...because they felt they had to release at beginning of May after promising end of March. Lots and lots and lots of work still left to go.

    And I don't even care that much about graphics. I would have been happy with better shadows in DX9 and R3E quality track and car graphics. The lack of updating on the cars themselves, the wonky FFB on the Civic, the science experiment FFB overall where very few cars actually feel authentic or as though they are all in the same sim (contrast with AMS, where every single car feels close to perfect for what it should be).

    I am not giving-up yet, but close. If Reiza continues on their path and soon has a new graphics engine with the current content at its current or even better quality, I am not sure what hole rF2 is filling? Reiza progresses in a week the same amount ISI/S397 does in a month or two months. A couple of things in rF2 are superior, but they are implemented so inconsistently or badly that it's a moot point.

    My favourite is FFB, but the same theme applies to track art and car physics. We all know rF2 has the best FFB of any available sim. The USF2000 is sublime. Reiza AMS cars come close (now that they also upped the FFB resolution), but can't quite hit the peak. But then there are a dozen rF2 cars that are a joke. Every car in AMS beats them. So you can have a couple of cars at 10/10 FFB in rF2, with most others spread out between 3/10 and 8/10. AMS you get 20 cars at 9/10 and none that are below 9/10. Sorry, bit for the normal person, having the product at 9/10 across the board is better. There is a joy to feeling the proper differences from one car to another as you drive each one. Reiza now has proper drivetrain programming; proper dirt on wheels programming (how come rF2 has the sound accompanying dirt on the wheels, but not the actual effect??); etc., etc. These are not insignificant items.

    Long, long way to go for rF2 and it is still slow progress by all appearances. I wish S397 the best, though. I truly hope the new Reiza product and iRacing will have a worthy competitor.
     
    Last edited: May 20, 2017
  10. bwana

    bwana Registered

    Joined:
    Dec 25, 2011
    Messages:
    2,139
    Likes Received:
    1,040
    Your seeing a feeling a lot that most aren't.. Hope you find your xanado one day
     
  11. 2ndLastJedi

    2ndLastJedi Registered

    Joined:
    Nov 3, 2016
    Messages:
    1,873
    Likes Received:
    1,198
    AC's latest update addressed my biggest issue for me, Ai, so i might test that for the next month and let rF2 update and come back to a more complete product.
     
    gabrielof likes this.
  12. Brack Jabham

    Brack Jabham Registered

    Joined:
    Oct 5, 2010
    Messages:
    585
    Likes Received:
    347
    I guess it is useful to have people who can provide deep insight and feedback, plus comparisons of close rivals. I wish I had the time! Things are certainly progressing, and my personal view is that S397 updates can really push rF2 along, when previously it was " a long time between drinks"
    As I personally don't have the time to spend on other sims too, I'll stay hopeful that parity, then better than parity, can be achieved
     
  13. MarcG

    MarcG Registered

    Joined:
    Oct 4, 2010
    Messages:
    6,854
    Likes Received:
    2,234
    Oh no the AI have lights on in cloudy conditions....just like in real life...THE HORROR!
     
    F1Aussie likes this.
  14. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,578
    Likes Received:
    1,071
    thanks, i must have old version i'll grab the new one !
     
  15. jayarrbee36

    jayarrbee36 Registered

    Joined:
    May 9, 2016
    Messages:
    658
    Likes Received:
    2,359
    Palm Beach has, Palm Springs hasn't.
     
    bwana and 2ndLastJedi like this.
  16. Gonzo

    Gonzo Member

    Joined:
    Oct 5, 2010
    Messages:
    954
    Likes Received:
    303
    @Marc Collins: i am sharing most of your thoughts, but i have to add: have you ever used the report bug function in the launcher (upper right) and fill a single report for each bug? The team is working like crazy on all that stuff, but they cannot find and fix all bugs if you do not report them. I know that reports got ignored in the past, but that was before S397! The internal testing process has incredibly improved. It also takes time to hire the right people to increase the progress speed. If the focus moves from the main things (graphics and gui) you might see great steps in other parts. The best thing we can do is to play rF2 and invite as much people as possibe. As you can see here the usage has increased since the overtake. If we all participate online, accept the situation and give the new team/company some time rF2 can climb to be a major title again, the steps which are gone since the overtake are huge compared to before but might be small if you compare to competitors.

    Please do your job and report those bugs one by one and see what will happen.
     
    peterchen, John, Louis and 3 others like this.
  17. Marc Collins

    Marc Collins Registered

    Joined:
    Jan 11, 2012
    Messages:
    3,159
    Likes Received:
    162
    The Portugal bugs, for example, were reported years ago when the previous update of that track was released.

    The other major graphical issues cannot possibly have escaped the devs unless they are blind or using only a very narrow set of test equipment. I agree they are working hard to fix everything they can. If they need my list of things that are glaringly noticeable after a 5 minute test then I will lower my hope and expectations.

    You could help by responding to my message with confirmation that your rF2 has the same issues, or no it does not. I already tried that earlier with the Civic FFB. Is it wonky on my rig--yes. Has the F2 FFB been wonky on my machine sine that car was released years ago--yes. But some others love the F2 FFB. ISI was never interested in delving deep enough to figure out that hardware responds differently to FFB programming, so the trick isn't to get a car to feel great on one hardware, but to get it to feel as good as possible across the whole range of commonly used hardware. A much more difficult task. And made even more difficult when you are "building the plane while flying it" and updating and altering basic FFB parameters and code over the years in between releases of vehicles. How can a car be released now with FFB that is reported within an hour of release as odd at best and dysfunctional at worst by a huge proportion of users? Lack of testing? Lack of testing on the proper range of hardware? No one cares? The USF2000 is the FIRST car in a long time released for rF2 that had almost universal praise of its FFB, including by me. I mistakenly thought that was the influence of S397 and that all the other cars would be reprogrammed accordingly. Nope, it was just good work by the apparently third-party modder who put the car together for ISI/S397.

    This is why I am saying there is a long way to go.

    And to MarcG, with your always unproductive commentary, the issue isn't that the AI had their headlights on because it was dark; the issue is that it was artificially dark in the first place. In real life no one would be using headlights in those daylight and weather conditions. The variation in light is exaggerated relative to real life. Sorry if after waiting 3 years for new lighting I expect better than a few clouds mid-morning to cause late dusk-like darkness. I can see the potential of the new DX11 lighting, but why is it still in this state of experimentation? We were told months ago that the DX11 build was ready to go and just a few tweaks were needed. Not true based on the evidence of what has happened this month.
     
  18. rer8

    rer8 Registered

    Joined:
    Dec 19, 2010
    Messages:
    492
    Likes Received:
    23
     
  19. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    Have you tried varying the rf2 files within steam,right click rf2,properties,local files,verify
     
  20. Alex72

    Alex72 Registered

    Joined:
    Dec 27, 2016
    Messages:
    1,317
    Likes Received:
    992
    @Marc Collins Yeah the FFB is a little wonky for me too on the modern F2's. Its like it has a life of its own. Its not even. Rolling out and turning the wheel it can be hard-soft-hard seemingly a bit random. Its not that i lose grip it just seems to change between heavy and light FFB at its own will. Hard to explain but its not a solid/tight/smooth as USF2000/Skip or other older formula. And the Civic, i tested it quickly and i almost lost my arms when accelerating out of a corner. :D FFB violently throws the wheel when throttling. Maybe its supposed to be like that on a FWD race car. Im no FWD expert.
     

Share This Page