“ are rfactor2 physics broken” video

Discussion in 'General Discussion' started by GTClub_wajdi, Dec 29, 2020.

  1. TJones

    TJones Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,074
    Likes Received:
    257
    Hi, good news as S397 at least looking into it, and try to improve the tires that guickly. I did'nt expect this.

    Looking at the video @Remco Majoor (#155) postet recently where we could see a quick lap at Sebring in a Ferrari 488 SUV (sort of).:)
    Look's like Risto worked with very soft suspension and therefor very high rideheight. With such high rideheight and splitter clearence, i doubt there would some appreciably ammount of downforce generation from undertray in the real car. Of course Sebring is very bumpy and they have to take this in consideration. But there are also some mid / high speed corners, T1 for example. At least I havent seen much SUV like setup's in real races there.;)

    So how is this possible at rF2?
    A few weeks ago, I did some testing with some GT cars at a skid-pad. It seems like, there is still plenty of downforce left even at max rideheight.
    Here's the link to my post about skid-pad tests (Post #45 / #49): https://forum.studio-397.com/index....-updates-includes-bop.67278/page-3#post-10408
    Don't know if this is believable realistic.

    Still hope someone from S397 Vehicle-Physic-Department (short VPD) :) can take some additional look into it, or come here and clear sings up in case i'm wrong.
    Potentially required improvements ("processing of aero maps" as a example) might reqire some code changes then. That of course need usually a lot fo time.:(
     
    Slip_Angel and Vladie like this.
  2. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    Remember that softer spring can get compressed by aero , so the car get low in fast corners and stay there, and is tall in slow corners where aero is unexistent. Remember too that underbody work WITH carbody , so both contribute to down force. For the extreme example of this just look at F1 1979 mod : terrifying downforce on very squishy springs, the car turn almost on packers in fast turns.
    About underbody depending on the design, rake angle can be more important than effective height
     
  3. John R Denman

    John R Denman Registered

    Joined:
    Mar 18, 2018
    Messages:
    512
    Likes Received:
    630
    When the model is built with accurate geometry and physical forces the setup responses are generally dead on the money with very similar MoTec data between real life and rF2. Suspension geometry makes a huge difference in how the car responds to setups. For example anti-roll geometry negates the purpose of AntiRollBars. What percentage range of anti-roll is built into the suspension geometry is built into the car through the expected ride height range?

    Keep in mind a few things.

    First like every other Simulator, be it a $40,000 seat of ANSYS, or a $60,000 seat of Mentor Graphics + Valor there is a limit to the realism of outputs. Both of those tools take quite a while to run their analysis feature, but with a race sim it needs to be realtime. So its never going to be flawless.

    Second S397 is in the business of generating revenue from the the only true race simulator on the market. The largest segment of the market prefers GT cars, nothing wrong with that, its just the way it is. To appease that market means releasing cars with proper authorizations from manufacturers. Some manufacturers may provide dimensions for suspension geometry but frankly that wouldn't be a smart move for any of them to do for a production based car. OEM's tend to prefer not to give that data away, although Prototypes are probably more forthcoming. So that leaves S397 with the task of creating and releasing cars of various designs yet keeping them as equal in lap times as possible. That's no small task. Imagine trying to implement a BOP on many cars in a class across many tracks with a release deadline. Ain't gonna happen with the budget prices on these mods.

    The dreaded Aliens. As with any simulator there are ways to exploit the operations, and when its a realtime based simulator, the gloves are off with exploitation. Fix one thing and two others are broken. Realtime simulator code requires a numerous interrupts through the cycles. If one slightly wiggles the wheel under braking what do you suppose that does to the realtime calculations? Some folks have learned to observe ways to exploit those little ways to cheat the system. Yes it sucks, and no it won't go away until hardware is about 1000 times faster.

    That doesn't say there isn't room for improvement. But it comes back to knowing specifically what needs to be updated, and no one can do that without specific data thats incorrect. Unfortunately that data won't be available with encrypted files. There is your conundrum.

    Its also part of why I don't do much with GT cars other than buy them to give S397 some revenue. The clock shows I've got something like 10,000 hours rF2, probably 1/2 that as idle time. My expenses to S397 are probably something on the order of $400. I can't think of any other product thats provided me with that kind of value, can you? Heck I've spent more on go kart tires at just one of my son's races than I've paid to S397.

    Try testing on some of the quality 3rd party mods that don't have encrypted files. Look at the suspension geometry in PTool (after reading Carol Smith's Tune to Win). Install MoTec and learn how to use it. You'll be amazed at how accurate this sim really is. And yes an honest simracer will still be chasing the aliens. There are a few really good alien free leagues although not many of them run GT cars.
     
    mixer61, Bruno Gil, instinkt and 3 others like this.
  4. romaFC

    romaFC Registered

    Joined:
    Oct 25, 2019
    Messages:
    32
    Likes Received:
    29
    You do realize what you wrote is false. Like very false.

    When S397 or iRacing or ACC get license for say AMG GT3 they get access to document like this. FIA GT3 homologation document. Along with manufacturer booklet.

    [​IMG]

    They will get everything all they way to aero map which will depend on manufacturer if they have full aero map. Even real teams sometimes will just get balance chart.

    Only thing that will separate the 3 games in same car is the tyre model on larger part. (Also hint chassis flex that rf2 does is made up numbers real life GT3 don't have that info.)

    I cannot believe that you keep peddling this garbage that rf2 is the be all true all simulator.

    Based on what? That pro diver said rf2 is realistic, that pro driver said acc is realistic.

    Unless you know exactly whats under the source code of rf2 where you can log everything and compare it to real life you will never ever know.

    Even S397 doesn't know that is why they keep improving the tire mode, aero, etc etc.

    Like every other rf2 competitor is doing.

    Its called evolution.

    10 years from now something better will come along and so on.

    5 years from now s397 could get better data and say oh wait we been doing slip wrong this whole time. Its not dramatic drop on curve.
     
    Vladie and FORMULAEFAN71 like this.
  5. Slip_Angel

    Slip_Angel Registered

    Joined:
    Jun 15, 2019
    Messages:
    588
    Likes Received:
    297
    Look i understand simulator are no joke, but can we not expect these thing to be even little bit more realistic ? you seen the video, the difference in ARB is basically nothing , even wrong is most places.
    I'm not chasing after aliens or lap times none of that BS. I play simulator for their realistic vehicle simulation. This is why we are here in the first place...if I/we wanted good A.I even console arcade games has good A.I, if people wanted good online then gran turismo has good online. you see my point ? This is the place where we can use our REAL skills and REAL knowledge to test and we need correct physics for that otherwise no point of a simulator... as even arcade games like forza has got other aspect of racing game covered i.e ai,online,content, ui .........expect PHYSICS
    people need to understand the priorities of a simulator and stop asking things that only take out resources in wrong direction.

    Edit : it is very disappointing that people here are mostly moaning about U.I, graphics , direction of a flag on a car , online system WHILE physics are straight up BROKEN for YEARS...the whole point of a simulator is lost when physics are bad.
     
    Last edited: Jan 5, 2021
  6. Slip_Angel

    Slip_Angel Registered

    Joined:
    Jun 15, 2019
    Messages:
    588
    Likes Received:
    297
    it isn't like that , softer spring means more roll and other side effects ...which are NOT good for underbody aero.
    True that soft spring car will get compress more BUT in straight line.... corners are totally different they induce roll which INCREASES the difference between left side height and right side height especially BIGGER issue for softer springs.
     
  7. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    Aero load is not subject to inertia, and car body roll is limited if there is limited suspension travel available. Weaker springs are slower to push up the body, another effect that help counter roll, together with damper. Race car have stiff setup to improve handling, not grip , aero is a factor but the amount of sensitivity of aero to roll is speculation without data or at least calculations ( it's a simple bernoulli problem , at least for a guesstimate level).
     
  8. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    It shouldn't do anything. rF2 doesn't interrupt or interject into realtime when it receives inputs, it's constantly running with driver input in the loop. That's why it goes into slow motion if the CPU can't keep up.
     
  9. John R Denman

    John R Denman Registered

    Joined:
    Mar 18, 2018
    Messages:
    512
    Likes Received:
    630
    Fair enough. If S397 can't build a mod exactly your way, than go play with the racing games instead.
     
  10. Remco Majoor

    Remco Majoor Registered

    Joined:
    Apr 22, 2014
    Messages:
    995
    Likes Received:
    1,068
    This statement gave me cringe
     
  11. John R Denman

    John R Denman Registered

    Joined:
    Mar 18, 2018
    Messages:
    512
    Likes Received:
    630
    That's interesting. If they can avoid interrupts in realtime I would dearly like to contract S397 to upgrade one of our missile simulators.
     
  12. romaFC

    romaFC Registered

    Joined:
    Oct 25, 2019
    Messages:
    32
    Likes Received:
    29
    ah yes.
    when talking out of you realm about something that you have no knowledge.
    you revert to silly replies.
    I don't go to my dentist and tell him how to fix my cavities.

    You don't even drive GT cars in rf2 but you are assuming they are realistic because you seem to think this is the only 'true' simulator in existence.

    Which frankly is the most ridiculous statement ever posted.
     
    Vladie likes this.
  13. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    Realtimes means predictable time of reaction. It does not mean "immediate" like reality (where we dont know how small the least reaction time is, maybe time is quantified too?). So you can have a cycle every 1sek only, its realtime somehow as long as it predictable happening.
    In case of rf2, for a physic cycle which run at a constant rate, the inputs are queried once and new states are calculated with it probably. I dont know for sure but thats the only way I could think about it. The code certainly does not react to each input change (interrupt) and calculate a new state.
     
  14. Martin Dyrlund

    Martin Dyrlund Registered

    Joined:
    Oct 1, 2013
    Messages:
    78
    Likes Received:
    53
    Here is a small picture of a official gte car at the start of braking.
     

    Attached Files:

    Sim_Player and romaFC like this.
  15. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    a real one? the gte are locked for downforce in rf2??
    The steep reduction of drag in the intial braking does not make sense. The rake is higher, AoA is higher, more drag
    Higher data rate would be useful, this is not 10ms rate?
     
  16. Slip_Angel

    Slip_Angel Registered

    Joined:
    Jun 15, 2019
    Messages:
    588
    Likes Received:
    297
    I missed the point genuinely, could you explain please ?

    If i have to guess from graph it looks like front reached peak downforce when car was rising up ? and pitch was stabilising ?
     
  17. hitm4k3r

    hitm4k3r Registered

    Joined:
    Jun 15, 2016
    Messages:
    1,320
    Likes Received:
    3,121
    Isn't drag dependend on the speed of the car, even with exponential factor? With initial braking you have the most speedreduction, thus the most abrupt drag reduction. The step looks indeed a bt wierd but that might be the samlping rate. Or am I seeing this wrong? From what I gathered from his post he is referring to the strong increase in front downforce. Wich is to be epxected as all the force is shifting from the back to the front of the car.
     
  18. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    I think the front downforce increase is due the front spoiler get near the ground preventing air to go under the car from the front, the rear lose downforce for the increased height that slow down air under the car.
     
  19. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    I hope you dont believe that there is a realistic air flow model in rf2.
    I think the primitive model in rf2 just shifts downforce forward by increased rake angle.
     
  20. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    Oh btw., maybe some can show also in motec the formula ride height - vertical tire deflection.
    As ride height does care about suspension but not about tire deflection.
    Im sure you are way negative breaking hard lol...
    There s so much off in the ride height, tire deflection, ground hit, packer / bumpstop begin modelling, what you guys talk about is just a drop of water on a hot stone
     

Share This Page