Cars seem too light in accidents.

Discussion in 'General Discussion' started by PedroJ, Oct 25, 2023.

  1. PedroJ

    PedroJ Registered

    Joined:
    Sep 15, 2021
    Messages:
    25
    Likes Received:
    24
    Rf2 for me continues to be the best simulator on the market, however, it has some flaws, which we all know.

    But what bothers me most (maybe I'm wrong), is the fact that when a car comes into contact with another car, no matter how light it is, we generally lose control of the car, it reacts in a way that indicates much less mass than it should have. At least that's the feeling I have in some contacts with AI, or even online.
     
  2. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,927
    Likes Received:
    3,879
    Absolutely. Also few days ago I just barely touched barriers exiting Vialone at Monza 1966 1000km at around 250km/h, and I lost control.

    Also, if contact happens does both cars go out of control in your situation, or is it yours only ? Do you think those contacts happen at the limits of adhesion/traction, when even a butterfly landing on your car could unsettle it, or does it happen in moments when car is much less stressed like going through straights ?
     
  3. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,479
    Likes Received:
    4,393
    Durge has long espoused that the rF2 car is too light in regards to inertia. His testing involved jumping cars over curbs at similar tracks with different sims. The rF2 car almost always got more air than the other Brand X's.
     
  4. Bernat

    Bernat Registered

    Joined:
    Jun 15, 2020
    Messages:
    757
    Likes Received:
    585
    I haven't come across a situation which made me think this. Can you give some visual example in a video?
     
  5. MikeV710

    MikeV710 Registered

    Joined:
    Mar 29, 2022
    Messages:
    394
    Likes Received:
    247
    Great band and drumming by Phil Collins :)
     
    davehenrie likes this.
  6. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    7,479
    Likes Received:
    4,393
    Phil was always a bit heavy handed on his drum technique. Like many Rock/Prog drummers, he suffered later in life from this.
    Near the end of his live career, he put out a great farewell to his drum kit with this tune. You really have to have the patience to get to the Mike Tyson part to see what I mean.
     
    pkelly likes this.
  7. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,927
    Likes Received:
    3,879
    I wonder how he did account factors like aerodynamics. And suspension/tires/chassis that are also important in case of impact for how large resultant force is.

    Aerodynamics are very important. In my opinion sometimes in simulations aero doesn't stall after driving over harsh curbs, while it should stall imediately, all the ground effects - gone, good part of front downforce - gone, maybe significant portion of rear downforce too depending on angles. You can sometimes see in sims cars after hiting quite big kerbs to be sucked back on the road very unrealistically.

    Here is 500kg in the air after jump with reasonable amount of aerodynamic lift. No lift and it woul have probably stayed in touch with road.
    [​IMG]

    When talking about hitting other cars, we talk about other car that has plenty of mass too, right ? And if it has speed considerably greater it will also generate greater impact force. And if both were traveling in same direction, one will speed up, and other will slow down. Usually one that speeds up suffer more, but it might not always be the case due to many factors.

    And when racing agaisnt AI, keep in mind thatthey run different physics, and has all sorts of differences that may help them remain more stable. I know this well, because recently I have been pit manouvered and rear ended by Ai way more than I am used to, however I think AI is very awesome now, just turn down their aggression.
     
  8. vava74

    vava74 Registered

    Joined:
    Dec 14, 2020
    Messages:
    845
    Likes Received:
    1,105
    That will vary significantly from car to car / mod to mod / DLC to DLC and is not - in my opinion - accurate if used as blanket statement.
    I do agree that in some cases this is true in rF2 but it is also true that there are classes where that happens in real life as well.
    For instance, high downforce cars are extremely "light on their feet" at low speeds as there is no aero pushing them down so, their mechanical grip will not be enough to make their rotation/spin more difficult to happen (smaller contact patch, harder compounds, ...).
    As an example, LMP3 and GT3 cars tend to spin quite easily when bumped on one of their rear corners...
     
  9. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,379
    Likes Received:
    6,599
    Without video it's really impossible to make any judgement, and even then things aren't always very clear. There are still people highlighting certain Apollo astronaut footage from the moon and insisting it shows it's fake, while some physics professors use the exact same footage to illustrate Newton's laws of motion.

    Durge's tests would appear to be trying to isolate a specific property (or two) from a scenario with many variables. In the game-physics world of little or no compliance and exponential bump rubbers a car having too much mass/inertia could cause too-big jumps just as much as a car being too light.

    With car collisions, aside from the multitude of variables involved in the motion/'grip' of both cars, the collision itself may or may not be reduced by factors like compliance and energy loss. Compliance and physical deformation (or in this case, the lack of them) is probably the main one. I don't think I've ever tested collisions to see what proportion of momentum is conserved, and I'm almost certain the game won't factor different parts of the car body into energy transfer calculations during collisions. So some will probably be wrong, but again how many and by how much isn't all that easy to determine.
     
  10. PedroJ

    PedroJ Registered

    Joined:
    Sep 15, 2021
    Messages:
    25
    Likes Received:
    24
    I should have been more precise with my comment. What i mean, driving, loosing the car, I don´t feel that the cars have less mass than they should. It is always in corners when we have less traction, and if some car just touch you, even if slight you loose control of car, usually with GT cars, like GT3, GTE, etc... in the open wheelers never felt that, even if we touch wheels.
     
  11. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,379
    Likes Received:
    6,599
    I can't really comment on AI as I haven't spent much time with them, but online I've very rarely felt like contact caused too much reaction. I've been knocked off track a few times, but replays showed it all made sense...
     
  12. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,379
    Likes Received:
    6,599
    The OP is talking about car to car collisions. I've had those in modern F1, 60s Formula cars, and GTE/GT3 without feeling that anything was amiss, having checked replays. When in a pack it's easy to lose track references (or an understanding of where and how fast another car is moving) and collisions and their effects can be unexpected, but for me it's very rare to think an interaction was wildly wrong.

    Cars rolling is a different matter, and like the jump tests really have many more variables (and game simplifications) than inertia and weight.
     
    Emery likes this.
  13. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,927
    Likes Received:
    3,879
    At least there is a way for physics to be intriguing people...

    At the limits of grip there is no wonder that even slightest touch will set car off and out. There really are so much variables, including very much the most fun one - the tire. I suppose if tire has a lot of reserve over the limit, then slight push of other car, slight change in tire loads and slip can be possibly handled without loosign car completely. This itself is about 50 variables alone. And besides that driver skills also can help. Problem is tht realistically most tires, especially high performance modern tires shouldn't have huge amount of "safety" over their peak performance points.

    Such basic things like mass and inertia takes part in most dynamics of car, quite easy to tell about both being higher or lower. I have never tried it, but it would be fun game to guess cars mass simply by just driving it fast.
     
  14. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,219
    Never had problems "believing" car on car collisions, but we know that there are limitations: in R/L materials can an will absorb energy, cars are not made of a single collider box, and carbody can have "interactions". This help reducing a bit the amount of energy that is transferred to the contact patch.
    In simulations, excluding a single product on the market (beam-ng) all those "dampers" are non existent. So I accept that car behaviour is more harsh and even our reaction to contact is late, as we understand we have been touched only AFTER the car has already started spinning.
     
    PedroJ and Emery like this.
  15. Bernat

    Bernat Registered

    Joined:
    Jun 15, 2020
    Messages:
    757
    Likes Received:
    585
    I agree with what has been said. The physics of collisions is too complex to simulate and also guessing what's right and what's wrong. We've seen cars IRL doing things we thought were weird. Cars flying and such.

    There's also the effect of an impact in the grip. The effect on the tire isn't progressive as when we brake or turn, it's a sudden change in grip that the tire might not be capable to handle so fast. Rubber can deform so much that it's no longer a tire. And then there's also other effects like the impact on the suspensions that might unsettle them in ways they're aren't designed to handle.

    In some cases I think the car could have absorbed the impact better depending on where it hit. The damping of the materials in a collision is very hard to simulate and I don't think any simulation does it completely right. They will probably set a damping for the whole car, some will set a higher value some lower, but it's an approximation that will never match reality. Except perhaps BeamNG which tries to model such things to some degree.

    For these reasons, even if we saw a video example we wouldn't probably agree about how the cars should have behaved except the lack of deformation in extreme crashes.
     
  16. MikeV710

    MikeV710 Registered

    Joined:
    Mar 29, 2022
    Messages:
    394
    Likes Received:
    247
    Never thought his Illness was due to his heavy hitting to be honest. He might not have the finesse of say Dave Weckl but I don´t see him to much as a heavy hitter. The drums on that video are very processed to match how they sound on the album which sounds like they are being hit harder then what he Is actually doing live. Plenty of heavy handed Prog/Rock drummers out there with no health Issues like Phil has so I am a bit surprised. Simon Phillips (Protocol) who played Rock, Metal, Prog, Jazz Fusion for example and one of my faves Is quite healthy, as Is Virgil Donati another fave :D

    Sorry for the of topic folks! Maybe we should have a music thread In off-topics :cool:
     
    davehenrie likes this.
  17. azaris

    azaris Registered

    Joined:
    Aug 2, 2022
    Messages:
    20
    Likes Received:
    11
    Collisions and crash physics in most racing sims are plainly bad. There are many reasons for this:

    1. Sims except for BeamNG do not model soft body deformations so all collisions are fully inelastic, which means cars conserve too much energy in collisions. This leads to ping-pong crashes or cars getting wedged into trackside objects and coming to an instant stop. Also the bodywork being rigid leads to silly phenomena like LMP cars launching each other into the air by driving under them.

    2. Numerical models of inelastic collisions at high speeds need to accurately capture the exact moment of contact, whereas sims have constant physics update rates that do not capture the exact moment of contact. This can lead to a situation where the car gets a spurious momentum vector added to it and is launched into the stratosphere. Also numerical errors due to netcode lag and cause massive added momentum.

    3. Cars are not modelled with any rotational air resistance and little if any friction once the tyres lose all grip and the cars start sliding around freely. You can see this in ACC or iRacing where a car gets hit and starts spinning like a beyblade and seemingly never stops.

    4. The rotational moment of inertia is approximated or just wrong. AC famously uses a bounding box to calculate the moment of inertia instead of using the individual component masses, which means all cars in AC behave like soap box racers from a certain point of view.
     
    Bernat and RaceNut like this.
  18. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,927
    Likes Received:
    3,879
    Anyone bothered to look up how similar incidents look IRL ?
     
  19. RaceNut

    RaceNut Registered

    Joined:
    May 21, 2013
    Messages:
    663
    Likes Received:
    541
    Makes sense. We are playing Racing Simulators, not crash simulators, after all. If combining both physics aspects were easy, developers would already be doing it. It's been said, even high-spec PC's will struggle to simulate advanced tire physics alone.

    Regarding BeamNG; I think it's a fun sandbox type of title but, as impressive as the crash physics are, it's not a great Racing Sim, imo.

    As sim fanatics, our expectations tend to far outpace speed of simulation advancements.
     
  20. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    3,035
    Likes Received:
    1,654
    Pretty certain the parameter is available in rF & rF2. Whether anyone, including devs, bother to use it is the question. Available data is minimal, at best. The only car I'm aware of having publicly available data is the GT40 and that only went to 15 deg of rotation.
     
    mantasisg likes this.

Share This Page