Track Simulation

Discussion in 'General Discussion' started by bravotangosix, Oct 29, 2018.

  1. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,118
    Likes Received:
    537
    Yes, a laser scanned track is only valid in one point of time. The time it got scanned. (For me that is good enough). After that time the real track changes every year. So it is important to know when the track got scanned. But at least you can discuss with people or do some research what got changed after scanning (like repavement of parts of the track).

    It is generally a bad idea to think you can drive the real Nordschleife fast in real life cause you have driven it so often on a simulation + laser scanned track. You will wreck your car. But I think you are a lot quicker learning the track.
     
    Adrianstealth and bravotangosix like this.
  2. KittX

    KittX Registered

    Joined:
    Jan 11, 2012
    Messages:
    147
    Likes Received:
    62
    Yup I was talking just about pure habits that you even don't think of.
     
  3. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,118
    Likes Received:
    537
    Yes, I know what you mean. For instance my simracing manual shifter's reverse gear is top left. On my real car the reverse gear is bottom right. After a lot of simracing i forget that my real car reverse gear is bottom right.

    When I was driving the Nordschleife in real for the first time (and before racing a lot in simulation + laser scanned track) I had the constant feeling that I know the track exactly (although it is so difficult and driving it for the first time in real). That was really a strange feeling, because I was never there in real life. And I noticed that I know nothing about my car (an old 110 PS VW Passat).
     
    bravotangosix likes this.
  4. KittX

    KittX Registered

    Joined:
    Jan 11, 2012
    Messages:
    147
    Likes Received:
    62
    Heh, another story of autopilot habits: when making an ice version of Buskerud a year ago, I was in the rig testing the surface a lot, making changes to the geometry noise and some grip/resistance levels to match or at least pretty much resemble an experience driving my own car on frozen lakes. After two or three days doing that I was taking my friend home, the roads were snowy and icy, we've been discussing something and I've done a U-turn doing some scandi flick, then on throttle and on full opposite lock, basically how I've used to treat loose surfaces all the time past couple of days, at the same time talking to friend. Then saw his face, then I thought "oh wait a minute, what have I just done?" :D
     
    bravotangosix and vittorio like this.
  5. FAlonso

    FAlonso Registered

    Joined:
    Jul 2, 2013
    Messages:
    128
    Likes Received:
    71
    It would be more important for me that the temperature of the track has effects on the tire.

    It would also be important to improve the tire model, at this time the minimum pressure always gives the best result.

    It would be more important for me that puddles are created in the circuit and that aquaplaning occurs.

    just a few examples ... the fidelity of the track is not so important, 99% of the players are not going to run on the real track.
     
  6. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,118
    Likes Received:
    537
    @FAlonso and you think that track modeling artists will solve your listed problems?

    EDIT: But I agree. Every problem you listed is important. For me too.
     
    Last edited: Oct 29, 2018
  7. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,118
    Likes Received:
    537
    Out of interest and not on topic. Are puddles correctly created for you when racing Sebing?
     
  8. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,529
    Likes Received:
    1,028
    laser nords is coming that will be brill , i hoped i'd be able to experience it in rfactor2

    know it well but using iracing
     
  9. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,167
    Likes Received:
    2,994
    I haven't noticed yet, but does track temperature really does not affect tire temps, and minimum pressures gives best results ? I have not been tweaking tire pressures yet in rF2, as it usually feels good by default. If there is such problem, then are you sure it is global rF2 problem, or just perhaps related to some mod ? First time I ever hear something bad about rF2 tire model.

    Spot on about aquaplaning, there is no aquaplaning for sure. And whole grip simulation is a bit strange, seems like just grip lowered and increased tire load sensitivity, I don't think rF2 simulates wet surfaces very well. Hard to tell, but seems like contact patch water accumulation might not be properly simulated, so you can feel rather similar grip level on low speed and on high speed. Couldn't tell if slip angles even change in the wet, I think slip angles should increase slightly in wet, to be more precise the peak should stretch out a bit while also being lower, not just get lower.

    Viscous hydroplaning is not simulated. So you never have to change the line in wet track.

    Regarding puddles accuracy, they have to be accurate in laserscanned terrain, because you basically have data of low local areas, however it has to be double checked with draining systems of the track and so..

    But I very enjoy wet racing anyway, it is already a blast. The way water builds up, the spray, effects on windshield, track getting gradually more slippery, then drying out, putting on slicks as soon as it dries enough and trying to place loaded tires always on dry surface... it is very fun already. But with better contact patch water accumulation and dynamic hydroplaning it would be superb, and then viscous hydroplaning but it would be extra i suppose, mostly because AI problems.

    Fun fact from Sebring again. It held probably wettest race ever in 1965
    [​IMG]
     
  10. FAlonso

    FAlonso Registered

    Joined:
    Jul 2, 2013
    Messages:
    128
    Likes Received:
    71
    They are problems of the game that have not been solved since its first days.
     
  11. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    9,154
    Likes Received:
    4,233
    If that flat pancake track isn't a flat pancake track in real life, then it's a poorly done track. And what puzzles me is that if it isn't a 'quick' conversion from another game, it's probably taken hundreds of hours to build - so how did a trackmaker turn a characterful bumpy track into a pancake track? It's wrong from the moment the physical track surface is complete, but it's made pretty and trackside objects are designed and built and all the new tech features put on top.

    We had ORSM Bathurst back in rF1, which was a pretty good representation. You could argue about how steep the dipper was (and I sometimes did) but it wasn't particular easy to prove one way or another. However, the Turn 1 exit kerb was totally wrong, should have been seen to be totally wrong from the moment it was first made, if you watch the Bathurst telecast one year you'll see that kerb hundreds of times, so again you have this fully made track with wonderful details and that kerb is wrong. Later on they patched it... but the kerb over the top of the mountain (inside of McPhillamy park) was still wrong. Again that's not a technical detail which is hard to judge - because, again, some of the height and slopes were a bit interesting on that track - but the size and placement of a kerb should be relatively easy to get right.

    So, for me, if you watch a real race and then want to go race on it yourself, you need it close enough that you don't notice anything wrong with it. If you see the real life cars hopping under braking into a particular corner, probably hear commentators talk about the bumps, probably see a car or three miss the apex or abort the turn because they locked up under brakes, and then your sim version doesn't have any bumps there, you start to lose connection with the real track. If you're doing a series and that version is all you've got, you'll use it anyway, but it's not what you really want to do.

    Laser scanning does at least provide some solid data that can be used to get the layout very correct, and reproduce the real surface character. It starts to lose relevance over some years if the track was smooth when it was scanned and has since developed bumps (or was scanned bumpy, and then got resurfaced) but it's still better than a modded track which has some corners wrong.
     
    bravotangosix and ApexModding like this.
  12. PLAYLIFE

    PLAYLIFE Member

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    Most publicly available simulation software is still a fair way away from being super accurate. Even commercial stuff still has a long way to go. Software models are very limited by an array of various things such as PC power, physics understanding, number of people a company can employ to develop it etc. One software code I have been involved with has been in development for 22 years and can still only predict structural integrity to within 40% of experimental data consistently!

    One can have an accurate circuit layout (corner radius, straight length, inclinations etc.) and that makes for an improved experience. Bumps are an interesting one. I pose the question, does it matter if bumps are accurately represented if the suspension or tyre or aerodynamic models are wrong? Bumps may just be a little too far in the weeds given the limitations of physics engines.

    My personal opinion is that the whole sim experience is limited by the most inaccurate aspect.
     
    juanchioooo and bravotangosix like this.
  13. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,167
    Likes Received:
    2,994
    Thats right if you get bumps right, then you make good use of physics, if you don't even make bumps, then you don't exploit physics to the fullest. And rF2 is very worth that. It is so good, you can even see it:


    Interesting idea about sim being limited by most inaccurate aspect, though I doubt if it works like if something bottlenecking anything. At least not in general. In specific issue maybe. Perhaps if a certain sim has suspension problems, then it might need smoothing out the surface even if it is laserscaned, it is not imposible. In this case physics would be bottlenecking track accuracy.

    P.S. for those who doesn't feel bumps/road. Make sure you don't use too much FFB smoothing, I don't use it at all with T300.

    Little bit offtopic. I am so in love with rF2 lighting and sky, the way it works together and affects shading. In Oulton park I sometimes get GT Sports look feeling, and GT Sport is prebaked lighting...
     
    Last edited: Oct 31, 2018
    Mauro and bravotangosix like this.
  14. PLAYLIFE

    PLAYLIFE Member

    Joined:
    Dec 26, 2010
    Messages:
    743
    Likes Received:
    126
    Put it this way, if a Formula 1 team like Ferrari with their vast amount of experience and resource cannot correlate their aerodynamic models with their wind tunnel and real on-track car performance, what hope do any lesser organisations have.

    I think the suggestion that any of the 'simulations' we play are accurate enough that one can consider extremely small bumps in a track surface and an accurate response from the physics engine (suspension, structural, tyres, aerodynamic models etc. all working in unison) is more hopeful than anything else.

    The interesting question is how accurate is 'accurate'? Are we talking within 50%? 20%? 5%? The only way to begin to know is take the physics model and run the real-world equivalent to the same conditions measuring every possible parameter and see how well it correlates. It's extremely difficult to correlate CFD models to wind tunnel tests let alone the far more dynamic real-world track testing.
     
  15. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,167
    Likes Received:
    2,994
    Maybe aero is a bit of a stretch for the topic ?

    Very small bumps actually aren't even coming from physics.

    You are right that we have not defined the accuracy. It is serious issue we have. I have written one long "rant" in one forum about how simracing as a game genre is not safe because it is simply nowhere defined what simulation is good enough, in order for a software to be labeled as a simulator. It is only supported by common sense I'm afraid.

    It would be really awesome if someone will come up with some kind of working system how to rate simulation element by element, but I imagine it would be super difficult, perhaps impossible and impractical. So it will probably always hang on common sense.

    If we could pick up fantasy percentage, then I would say 90-95% for essential stuff is enough :D Well but it would certainly depend on an elements, and so many things... there are so many variables, it would be best to stick to the basics...

    As for undulations/bumps of the track, it depends on reference scale. 5cm off per 100 meters might be acceptable. But to have a rather localized undulation/bump which takes only about 10-15meters off by 5cm is very significant, 5cm in such distance is significant when you hit it with a racecar at high speed. I have modeled Jaguar XJ13 with extreme attention to body lines, 2-3mm made significant difference, and even less in some cases.
     
    bravotangosix likes this.
  16. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    9,154
    Likes Received:
    4,233
    @PLAYLIFE I think aero should be excluded from such discussions, and the limitations in that area certainly shouldn't be used as an indicator for things like bumps. If we look at the glass as half-full instead, 15 years ago there were games that some people considered good representations of driving a car (very subjectively), but after playing rF1 and sims of its ilk those other games showed their deficiencies. We still see that today with rF2 (and even still rF1) compared to more mainstream titles that some people consider a decent driving experience - in rough rubbery percentage terms you could say the games require 50-70% correct driving technique to do well, so rF2 and other better sims would have to be judged higher.

    What percentage accuracy you place on each subsystem becomes irrelevant to some extent, what matters is the whole experience. That's nearly impossible to measure, and the majority of players have no realistic reference to compare to (but are usually the most vocal in their opinions), but with an accumulation of A-B comparisons the cream steadily rises to the top.

    And as usual that's why it's infuriating that the better sims lack many real life aspects of racing series or cars, but I digress.
     
  17. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,167
    Likes Received:
    2,994
    I said that aero is a bit of a stretch when focus is track simulation, but there might be some link as downforce ispressing car down and aerodynamic lift does the opposite. I suppose higher downforce restricts vertical movement but it results in higher stress for chassis per bump, and lift is the opposite. Though higher downforce logically requires harder springs which will result in more vertical movement... yeah might be a bit of a stretch in track accuracy discussion, but at a very high level of accuracy could be worth to take in account.

    @Lazza I agree, I'd say the same purely by guestimating percentage, rFactor 2 is very likely highest, and it doesn't seem to change soon. I thought that ACC can make a good run, but seems like it hits the wall of its own popularity and community habbits as comparing their R1 to R2 releases, people want "exciting" drifts, and VERY specific FFB to do the hard work for them- this way it is so obvious how slipangle curve dropoff is made to be easier, FFB SAT dropoff also seems to come earlier, tire responsiveness has also been increased a bit, marbles off line less dangerous, and grip is not so much lower off the rubbered normal racing line............

    There is no perfect sim, but at least there would be a tendency of gradual improvement with an ultimate goal of maximum realism. It is simply exciting. Thats what I hope from rF2. The way I see see rF2 at the moment is the only simracing title which can really make true simulation popular. Well perhaps also Reiza with their fine projects, I think Automobilista is outstanding sim.

    Speaking of Reiza I am really looking forward to their pack in rF2, they makes just outstanding tracks, something perhaps rF2 lacks the most.

    So far my favorite rF2 tracks are: Monaco 66, Spa 66, Atlanta MP, Mores, Lime Rock Park, the new Sebring (so good!), Matsusaka, Zandvoort, Silverstone and perhaps the most favorite Oulton Park, not only blast to drive, but it also sometimes looks incredibly realistic, more than most other rF2 tracks.
     
    bravotangosix and Mauro like this.
  18. ECAR_Tracks

    ECAR_Tracks Registered

    Joined:
    May 1, 2016
    Messages:
    456
    Likes Received:
    465
    RF2 is the best one at physics but let's face the fact honestly: there's no an accurate racing available. Every single one has their fatal flaws. I can't consider a proper simulation a virtual car that not needs clutch to be geared or can reduce 8 gears in a blimp without damage the clutch, gearbox nor the flywheel. It's like a flight simulator without any fundamental fly aspect as flaps for example. Needs fixing.
     
  19. ECAR_Tracks

    ECAR_Tracks Registered

    Joined:
    May 1, 2016
    Messages:
    456
    Likes Received:
    465
    Couldn't agree more People concerned if the track is + or - 2 cm off the real life while run at 10°C or 50°C has no effect Everyone with a minimum real life racing experience knows how absurd is that. Yeah, it's off topic but I'm always amazed how something like that is barely recalled in this Forum
     
  20. Alex72

    Alex72 Registered

    Joined:
    Dec 27, 2016
    Messages:
    1,315
    Likes Received:
    985
    A lot of modded tracks have had flat pancake Surface. Not directed at rF2 modded tracks so much but i am a sucker for feeling the Surface in the FFB and when tracks are smooth like an Indoor floor im not enjoying the track no matter how well done it is in Everything else. I remember many years ago when LilSki mentioned the (for AC at least) extra physical layer that gives the bumps, and many had skipped that layer. To be fair many probably didnt even know you could do it or that you should do it to give the track FFB character, but my Point is that personally i couldnt enjoy myself on tracks without any shaking/rumbling in the FFB. Thats why i Always loved LilSki's tracks - Always knew he put a lot of love in the unseen as well in the seen parts. :)
     
    bravotangosix likes this.

Share This Page