“ are rfactor2 physics broken” video

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

  1. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    @Jack Dyson just to be clear, I'm not involved in the game development. My questions are to help clarify your meaning and wishes. I suspect on the aero side you have a limited understanding of how the game works, so trying to confirm if that's the case or you have specific targeted improvements (or current shortcomings) in mind.
     
  2. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,926
    Likes Received:
    3,876
    What cars do you drive ? Slick tires should be more difficult to overdrive than treaded tires.

    I personally don't observe things you are talking about. I think thats because we talk about same things differently.

    First of all, I am confused why you assume amount of deviation from racing line compared to small steering input as instability. What I'd call an instability of a car would be some type of oscillation which is hard to prevent, and more importantly hard to control. What you essentially describe is responsiveness of the car, and if it is too responsive then you say its unstable.

    Secondly, its confusing that you compare slick vs radial tire. That makes no sense, I guess what you actually mean is either slick vs treaded, or radial vs bias-ply tire. By the way, this morning i just read an interesting article about bias-ply vs radial tires history.

    Third the responsiveness of a car will definitely depend on responsiveness of a tire. However the suspension and chassis stiffness does have a role in it. Wheel alignment has a role. And steering ratio has a role in it. Actually road surface characteristics also has a role in it.

    BY THE WAY, what do you mean when saying that car should "bite" ? I have always been confused by people using term "bite" talking about cars. In one way it could be understood that tire "bites" in the tarmac, which could be a metaphor for tire gripping up. In another way it could be understood as a car punishing you for not being good with it.

    And finally, are you suggesting that rF2 tires are too unstable at and over the limit, or the opposite ?
     
  3. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    @Lazza Hello again, thanks for making that clear. Yes I confirm I have no detailed info about how aero is implemented in rF2. I have some improvements that I can suggest for the aero part - but as you say I want to be sure and not waste anyone's time. Please can you direct me to some documentation (or academic papers) on the way rF2 renders its aero model and the way the aero forces are integrated into the dynamic traction model ?
    Cheers
     
  4. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Hello, thank you for your reply.

    I drive mostly the classics like the Formula Vee, Nissan BTCC, BB20 presently. I have played around with other cars like the F1 McLarens but not on a regular basis.

    Yes : slick vs treaded (sounds interesting, please give me a link to the article)

    I am using control systems terminology when I say "stable" or "unstable". I can use a thousand words to describe it but I tried to simplify the answer. A racing car at the limit is balanced on the maximum tyre grip ("lateral G vs slip angle" curve) which has almost everywhere negative curvature. This is a mechanical instability or a dynamically unstable system.

    If you try to rotate the car on a quick lap, one method is to quickly and gently "nudge" the front tyres into the apex with steering while lightly braking to induce oversteer. This pushes the rear tyres momentarily to an extreme and increases the slip angle on the rears quite a bit. The rears are then near the "cliff" in such a maneuver and more often than not I go off the track - the car "bites" me.

    The cars in rF2 is either 1) relatively stable in turns with slicks 2) relatively unstable/undrivable with treaded (compared equivalently to say AC). To compound all that as far as I am concerned - reducing the tyre pressure creates an unusual amount of "drivability" in the car - up to unphysical laptimes - when I do that the rF2 car will no longer "bite" me *as much* no matter what I do on hot tyres (within reason).

    I know all this depends on the car and high speed aero lift on the rears in the historic racers like the BB20 and the spoilers in the Nissan and bumpy surfaces and what have you, and yes I know slick tyres have different profiles and tyre compounds and will give greater latitude to the driver in terms of overdriving; nevertheless I am speaking about relative inaccuracies within the simulator that lead to potentially unphysical behaviour and an unnatural feel compared to real life driving at the limit.

    I don't intend anything about the responsiveness of a setup (rate of change of direction per unit input of steering or throttle say) in my answer. I know setup will change the driving limit and the nervousness of the machine - the tyres will determine the envelope of performance however and this is what I am feeling and relating to.

    RF1 was better in this respect because it had a better calibrated tyre model (simpler - like AC) - I recently read a good article about a Formula Vee champion who used it for analysing/comparing AI driven laptimes against real life ones and found a very good correspondence between the two.

    So : to conclude I am saying that the rF2 tyre is way too "stable" over the limit when depressurized in my opinion.

    Cheers :)
     
    Kelju_K and Kahel like this.
  5. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,926
    Likes Received:
    3,876
    @Jack Dyson I have seen before you wrote about FVee, I must mention that discussion about it has occured once, the important bit to learn is that it turns out that FVee is different in different places, for example in Australia they would have it handling better, in Brazil they have it more challenging. And there was onboard video that proves it. It can be improved with a setup quite a bit though, but still that thing teach things, realistic IMO.

    Here is the link: https://www.actiongatortire.com/evo...Jv6NlKDssrZV1o2YfPH7QJxM7d4SKs_qE6ZHCRbAUfve4 I was basically researching anything I can find about 50s tire durability and related stuff.... I have found a lot of interesting information by doing that recently.

    Thats a correct observation that in rF2 for some reason lowest pressures are always best. At least for short duration stints certainly. I used not to believe that. And there are indeed some cars that are quite difficult before you get pressures down. Just for example Honda NSX or Chevrolet Corvette are not so great before you minimise pressures and increase negative cambers, then they are pretty easy, and sometimes perhaps even too easy.

    I agree with your conclusion. I think the same way, however there are discussions happening sometimes, some people have good arguments of why lowest pressures is not wrong to be fastest. However there are good arguments of how rF2 does lack of limit how low the lowest could be. Technically cars can be made to have reasonably low lowest pressures. But it is possible to make a mod that would have no pressure at all in the tires, which is 0kpa or complete vacuum, and tires will be fast. I myself been working with physics moddign for a couple years intensively and I have a lot of interest in tires and aero as well, at least to how much I need to work with 1950s and 1960s cars mostly. I can make tires that will not be allowed to be depressurised bellow 200kpa for example, and I can adjust their performance to the level which those tires are expected to have at that pressure, and thats a work around.

    I'd say rF2 physics are not handicapped by the fact that you'd probably never have to search for optimum pressures to achieve best laptimes. There is a handicap of experience for not having to search for optimum pressures. However, I guarantee there would be some unhappy people if they would suddenly have to put effort to find out best pressures, instead of mindlessly just minimising them.
     
    Nieubermesch likes this.
  6. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    Concise documentation on that sort of thing isn't really available. But if you look in your rFactor 2 folder, you'll see a Moddev folder - in there go to Vehicles, Skip barber, and open its .hdv file (in a good text editor) and you'll see how the car aero and suspension is defined. Most lines in that file have comments.

    Much of it is unchanged from rF1 so you can search online for specific parameters and usually find some explanation or discussion, but the majority is self explanatory.

    For aero you'll find the sections I mentioned above (front wing, fenders, etc) and the various parameters give a decent indication of what's possible. Although, it's only when you're actually modding a car that you find ways to manipulate the various values to achieve the desired results, and this process also uncovers ways that things can go wrong if you don't consider possible side effects.

    There are many parameters that I personally feel could be expanded on or altered to better facilitate proper reactions across various scenarios, but it's certainly more functional than the car setup pages might suggest.
     
    Jack Dyson likes this.
  7. Kelju_K

    Kelju_K Registered

    Joined:
    Sep 1, 2013
    Messages:
    393
    Likes Received:
    374
    This exactly. This allows these crazy rally techniques, that are completely out of place in road racing cars, especially cars like the rear engined porche cup. These moves (the nudge Jack mentioned, and the crazy rally flick) that are possible now should bite you really hard with rear engined car.

    And some of us would be very happy about it :)

    I love the porche cup car, i think its one of the best sim cars ever made, but when lapping with it just for fun, i find my self shaking my head too often when i recover from mistakes that i shouldn't have. It gives such an arcady feeling and ruins the fun, and takes the smile of my face when im enjoying driving.

    Rf2 competition system is so much behind the curve compared to competition that for that thing to take off properly, you need something extraordinary in the physics side. Not this. Especially when you total the amount of money one has to invest into the content to be able to race online.

    This game made its reputation as a best physics sim, when the cars were harder to drive.
    And for those nitpickers who will pick this and be like: "hard to drive doesn't automatically mean realistic".
    YES i know that. Do you think that the reputation was created solely by racing engineers playing he game?
    NO! it came from ordinary people realizing that hey, I got to learn proper driving technique and setup adjustment to be able to drive this car as fast as i was able to in the more arcady sims.
    THAT was the hook. The hook that made people to switch from the "other game".

    S397 needs that hook, and the switch. Both, as there are limited amount of simracers, that are already spread around.
    Now there a lasso in place of that hook. It aint gonna cut it. Mark my words.

    Now if someone is coming from other sims to try this out, based on the reputation, he is not going to be impressed, and those few that i have been able to get to buy the game were not that impressed either.
    Right now FFB is the only thing clearly above other sims, but that's a thin thred to hang on to.

    And who would like critisism? haven't met anyone in my 48 years of life. But developer ego is #1 killer of games, because it filters out that critisism. And the fan boys do a great job helping in that.

    Thats all from me to this topic. Cheers.
     
    Last edited: Feb 4, 2021
    kados, avenger82 and Jack Dyson like this.
  8. vava74

    vava74 Registered

    Joined:
    Dec 14, 2020
    Messages:
    839
    Likes Received:
    1,100
    This should provide some perspective for the rFactor 2 users who suddenly feel "insecure" after Ermin's lazy and unbalanced video (not questioning that the physics engine needs work):

    This is a "must watch video" and... a "must keep the link available at all times video":

     
  9. Yzangard

    Yzangard Registered

    Joined:
    Dec 6, 2018
    Messages:
    351
    Likes Received:
    173

    To be honest, latest tire model is MUCH closer to what rFactor 2 is right now.

    Oh, and out of curiosity, I tested this "overturning wheel steering techniques thingy" in ACC and iRacing and you can do it in both games. Much easier in ACC than iRacing but still possible...now, you don't get any time doing so, not sure you actually gain some in rFactor 2 but I know it involves a special setup I don't want to use, so...
     
    Last edited: Feb 6, 2021
    Nieubermesch likes this.
  10. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,926
    Likes Received:
    3,876
    Well it doesn't help rF2 guy very much, since partcular rF2 cars are probably unrealistic just as much, but on opposite side of a scale.

    By the way things like 1:33 in that video proves that there is too much peak friction in iR (well obviously much too fast laptimes and need to drive on rails is also an evidence). While these cars has high CG and high grip anyway, I believe they have a possibility to tip over, but such easy topping over with so low kerb and barely starting to turn is only possible because too much friction.

    Also a proof in 20:02 in this video
    Thats Lotus 49 almost flipping over due to grip !!! I think it is fair to say that this is outrageous, not surprised that thing is undrivable and it basically killed interest in iRacing for me.

    Do you know why they put such tires on it ? I remember an answer from forums. Turns out people found it difficult to drive with lesser grip tires (supposedly realsitic grip, of which I just guess). So they put totally unsuitable high frequency high grip hardcore bite tires on relatively low frequency car that should perform quite a bit of sliding and totally ruined it lol. It is so bad it is funny to me. It is so bad that I get that much sad that it becomes absurdly funny. While S397 70Mclaren F1 drives like rally car on snow with fat slicks and downforce, iRacing Lotus 49 with no downforce drives like on instant super glue. These two extremes are also funny. Meanwhile there are quite some properly driving similar cars in rF2. But they are not "the ones to use"... lol
     
    Last edited: Feb 5, 2021
    Jack Dyson likes this.
  11. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    Would like to know what Ken Block think about RF2 tires ;)
     
  12. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    @mantasisg

    Hey my apologies for this late reply - I am so bogged down with all sorts of things here I just could not find any time to answer you properly!

    Yes indeed I have done some research on the differences in F-Vee a few months back and I knew that they are different in different countries. I have improved the Reiza one via setup quite a bit as you say and doing a mega lap I can almost get within 0.5 seconds of the AI on Silverstone National. The completely open coast is very hard to manage - but it is very good car control training mind you. I love F-Vee cars and was hoping to get a chance to drive them one day. That's why I got rF2 in the beginning.

    Thanks a lot for the link ! Fascinating to see that "trial and error" played such a big part in the development of tyres.

    I do agree about the Honda NSX - it is not an easy car to drive at the limit. Yes every car I have driven is transformed upon using this trick - I just did a race with the Nissan and it gains masses of grip in the corners when the tyres are deflated. In fact it handles like a kart :D because you can jump over curbs and land on the other side and it will not bite you.

    Now I understand what you are saying about pressures - in the real world low pressure tyres give more contact patch (mountain bikes on rough terrain) but the moment the tyre pressure reduces so the compliance (sidewall rigidity) increases. That destroys the handling and the temperature characteristics of the tyre. However it does give you back a lot of grip in the corners. Technically it should also increase tyre wear. Thus an overly compliant tyre is not recommended in real life (this is treated in for example "Tune to Win" by Carroll Smith). There is an obvious problem with deflating tyres on aero cars which then go faster :D



    I think that people are really welcome to their arguments and rF2 as well - we can always choose what is best for us. There should be no tribalism involved. Informed criticism is always a good thing however.

    I think you are right - limiting the setup is the way to go in mods : this is what AC does as well - for example you can't lower the front ride height on the AC Nissan BTCC lower than 45 mm. Actually I am blown away that you do mods, it does sound so interesting and I would like to know how you manage to understand all the tyre parameters for rF2! - can you send me a link to your page ?

    I think no product is perfect, it never will be. However borrowing from physics, we should always try to keep the models as simple and as reliable as possible. This would have benefits for the modding community and the community that drives those cars and track. A simpler model is easier to maintain in the long run as AC and rF1 have shown.

    For sure rF2 physics are not handicapped - as you point out they can be very valid, but they are managed and calibrated wrongly; I suspect this is the case because the model is not a simple direct linear feedback system : from what I have seen there are neural networks involved and that stuff can be tricky. If I have understood this correctly : the approach is brilliant and novel - however it is also difficult to understand for various reasons. I have seen Marcel admit this in an interview himself. He also said that the tyre data they had to calibrate their tyres was unreliable and conflicting. Personally, I think S397 took a risk by spending 4 years on the UI which most people today don't like and is less reliable and has removed certain features previously available. It might have been the way to go from a software engineering point of view, and I can understand that, but from a user's point of view, it was better to serve their main forte/selling point first : the much vaunted vehicle dynamics mode "397"l. What has happened now is that they have the worst UI in the industry, a vehicle dynamics that is incomplete, no real documentation, graphics that are being beaten by the competition and modders alike. They are, unfortunately sinking their own game (and I think you said this in an earlier post you made about the UI). If you read the updates they post : they are proceeding as if they believe their own hype. That is very bad for a business for many reasons. Almost all of us (that I know) who race rF2 are exasperated to put it lightly.

    Ayrton Senna never made any judgement about a cars performance without knowing the tyre temperatures: in real life the pressures have to be adjusted to give you the best overall contact patch and grip around a course and if people want to drive an accurate sim that is just the way it has to be. Of course this does not restrict their choice or opinion.

    I will just finish by saying: if a sim car exceeds real world laptimes around any given track by a "significant margin" then the model needs recalibration at the limit.

    Most of all I am really glad to have got to know more about you during these chats - would love to see more of what you are doing and thank you for the article again!

    best :)
     
    Last edited: Feb 7, 2021
  13. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    XD I know I know .... That thing is based on the Papyrus-Kaemmer GPL engine which was famous for this sort of thing. Strange they never fixed it.

    I do agree that Ermin's video was a bit OTT and I did try to let him know that what he was saying could not be rigorously verified (e.g what does he mean by transmission physics simulation ?! - are these the internal inertial losses of the block and transmission ? )

    He did, in general have a point however - and yes the video is a bit on the "lazy" side.
     
    Last edited: Feb 5, 2021
  14. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    @Lazza
    Thank you so much ! I think I understand better what is going on now - and so fascinating too. I'll check it out in detail of course.
    best!
     
  15. Yzangard

    Yzangard Registered

    Joined:
    Dec 6, 2018
    Messages:
    351
    Likes Received:
    173
    You feel more in AC than in rF2, seriously ? Your rF2 settings are extremely wrong then, clearly.
     
  16. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    clearly rF2 FF is inferior :D and not to my sophisticated taste ...
     
    Last edited: Feb 6, 2021
  17. RaceNut

    RaceNut Registered

    Joined:
    May 21, 2013
    Messages:
    659
    Likes Received:
    539
    I like AC-ffb and the fact that it's pretty versatile in how it can be tuned depending on one's hardware. I miss that aspect in ACC-ffb.

    rF2-ffb excels at conveying a more intense sense of cornering forces (& front grip-loss) unlike any other title (imo of course). It manages to do so while preserving surface bump detail very nicely. rF2 sets the bar in providing immersive ffb for me, and does so in the most visceral way.

    There are certainly other titles I enjoy (for various reasons) but, none of them can match rF2's ffb. That being said, it is largely car-dependent. Some of the older content using the threshold fall-off effect, loses some of the intensity.
     
    Jack Dyson likes this.
  18. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Yeah I agree, probably it is a car thing and I have not got the latest rF2 GT DLCs (already spent loads on it so I'm not going to get anything else) so perhaps they have a different situation for that. But racing loads of different tracks and cars (Ladas, Cortinas, FVees, Nissan BTCC, Lotus 49, Brabham, F3 '66) in rF2 has tended not to impress me on the ffb side - I can feel the bumps and front and rear slip angles ok as you pointed out, but brake locking and stuff like that eludes me completely. In fact I can't tell in rF2 when the brakes lock if I don't drive an open wheel machine. In fact I have noticed that lots of people do swear by the ffb for rF2. With AC I can tell what the car is doing at every point (slip, braking, traction, yaw, bumps, and so on) for some reason (across saloons, formula ford, coupes and now GT3). It is amazing how down to individual taste this is actually and it is good that we have a choice of good sims as it is quite subjective.
     
    Havner likes this.
  19. RaceNut

    RaceNut Registered

    Joined:
    May 21, 2013
    Messages:
    659
    Likes Received:
    539
    I agree that AC gives very good all-around information via ffb and is among the most natural feeling titles to drive. I just find rF2 ffb is on another level but, I didn't always feel that way either, somewhat due to my hardware.

    Perhaps it's a combination of factors; the cars I use, the hardware, my custom ffb-settings that I've tweaked over time but, whatever the case, it just works on another level. I actually get more brake feel than in other titles, probably because I utilize the brake-effect and lock-up is very evident. My biggest problem with rF2 is every time I play it, I stay up too late when I should be sleeping. :)

    I also think there is much to be said to acclimating well to any one title. There are some cues that take a bit of time to come to grips with, some extremely subtle ones. Then consider that those cues involve different senses in somewhat different ways. If I spend a couple weeks or more playing just one title, most others tend to feel a bit strange when I play them again. It doesn't take long to get reacquainted but, I need to have spent a significant amount of time to learn it well, perhaps even to have a "bonding experience" with it.:eek::D
     
    Jack Dyson likes this.
  20. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    :D Yes indeed I do agree with much of what you say and I am really glad about the most important thing: that driving rF2 gives you so much pleasure, which is what it's all about in the end isn't it? - I use a G25 Logitech which I find to be very good.

    In counter, there is just something about AC that keeps me up too, it is such a complete racing sim, so much style and thanks to the modders (aka "your racing simulator") so much emotion - have you tried the wet weather mods by any chance? I want to get those.

    If you would be so kind could you perhaps show me how you adjust your rF2 FFB settings ? I don't any brake-effect settings in the UI (I have the car force multiplier, MST and something else .. ah smoothness) that I can find - perhaps you are adjusting via file parameters?

    Anyway thanks a lot for your reply!

    Buona domenica (as we say in these parts)
     

Share This Page