“ are rfactor2 physics broken” video

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

  1. RaceNut

    RaceNut Registered

    Joined:
    May 21, 2013
    Messages:
    659
    Likes Received:
    539
    Grazie! Absolutely adored our time spent in Italy, btw. ;)

    Ah yes, the mighty G25, I had one of those quite a few years ago. I have to say though, rF2-ffb did not favor that wheel. It was when I upgraded to to the Fanatec CSW that I started getting better results. I later moved to the AccuForce DD-wheel, and now I use a SimuCube1 system. A G25 and a DD-wheel are miles apart, especially when it comes to details and overall fidelity.

    Where the more subtle effects exist, a gear-driven wheel does not have the same response & precision to convey some effects as effectively. I suspect that may be why you're not getting good brake-lock cues. Are you using some min-force in rF2 to try to account for that?

    The other thing I've done for the brake effects is to enable & increase the "Brake Effect" parameter (in your custom Controller profile). I don't know what to tell you as far as the effect level but, it provides increased steering resistance as the brakes are applied. You would have to experiment with the setting but, I'm not sure how effective it might be on a G25. It's quite dramatic on a DD-wheel and adds to the intensity / immersion (imo).

    There may be a few tweaks in your custom ffb-profile you can apply specifically for the G25 but, those are not in my "wheel house" these days.
     
  2. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Hey hello ! :D Indeed, it is a fascinating country isn't it ? From its lazy rolling valleys and ancient shrines to wines that can make you taste the sun. Just an amazing place - I am so glad you enjoyed it and hope you will come visit again!

    Right that explains it then - when AC came out the G25 was still King. So they would have optimised the ffb for that. I am sure the newer hardware is more nuanced - and your description of it sounds fascinating. I personally don't know anything about the new wheels and tech. I do hear guys like Ermin sing odes of delight to them, but you know, Ermin only stretches so far before you have to start taking pinches of salt that turn into spoonfuls for everything he claims :D - god bless him.

    Actually it was the same Ermin that also said that his ACC ffb experience improved as he upgraded his equipment (I still have to test this with some salt :D) but it does tally with what we are conjecturing : these games are newer than a G25 and so they probably use newer APIs to generate the ffb. Actually I remember that later version of the G29 have a special version that uses some sort of new API as well.

    What is important to me is basic feel - if I have that I know what I am doing wrong in the car. I want to try F-Vee at some point and that is why the realism aspect is important to me.

    My G25 setup: No min ffb is turned off in rF2 for me - but I will turn it on now to check (the fact that rF2 has no standard manual makes it really hard at times to know what all this stuff does with precision - especially if you have tons on during the day).

    Ok so the brake-effect is in the G25 controller profile somewhere (can't see it in the normal UI but I'll study up on this topic). Yes exactly ! When I apply the brakes in AC the steering stiffs up and can simulate wheel judder (which I tried and checked in my road car - problem there is power steering ...). In rF2 - zippo.

    I am grateful for your advice - thank you :)

    You will never guess what ? I did my first AC race : bloody brilliant ! 60 mins of GT3 immersion - I had turned down the colour saturation and turned the photographic filter on : it looks amazing ... just wanted to ask, do you race online ?
     
    Last edited: Feb 8, 2021
    RaceNut likes this.
  3. David O'Reilly

    David O'Reilly Registered

    Joined:
    Jun 12, 2012
    Messages:
    1,050
    Likes Received:
    756
    ;););):D:D:D:cool::cool::cool:
    What rhymes with "chick bait" and gets viewers.
     
    Jack Dyson likes this.
  4. AMillward

    AMillward Registered

    Joined:
    Apr 4, 2017
    Messages:
    1,879
    Likes Received:
    1,837
    It's like that Austin bloke. Said stuff knowing it got people wound up and clicked on the video.
     
    David O'Reilly and Jack Dyson like this.
  5. RaceNut

    RaceNut Registered

    Joined:
    May 21, 2013
    Messages:
    659
    Likes Received:
    539
    Yes, Italy is amazing, been there twice with the most recent trip just prior to the arrival of Covid.:(

    I don't do online racing very much, it can be great fun though, especially with the right group of players. I have limited time to play so, offline play suits my schedule much better.

    I think it's fair to say that a DD-wheel will improve the FFB in any title. The results of the increased dynamics are simply undeniable, imo. Without even considering the increased torque aspect, the precise detail and smoothness they can deliver certainly put's them on a whole different level. Keep in mind though, I fit into the "more is better" group regarding ffb.;)

    The Fanatec CSW 2.5 is a good alternative (lower cost / belt-drive) option or as a stepping stone to Direct-Drive.

    The value of better hardware really comes down to personal preference but, a DD-wheel made a massive improvement for me. I may be more sensitive to the fine nuances in ffb effects but, the benefits in terms of immersion make the additional cost very worth it. In addition, my overall consistency in lap-times and racing improved. Better pedals (Especially a load-cell brake) can really help with braking consistency and precision, easily improving lap-times and consistency more than a better wheel.

    Purely from a competition perspective, spending on a DD-wheel may not be justified but, for those that value immersive FFB, they are extremely useful, imo. It really depends on what one's priorities are within the context of Sim-racing, and how serious one is about sustaining the hobby over time.

    If you are familiar with Automobilista, I would rate its FFB as being closest to rF2. The Formula-Vee does a good job of bringing that same ffb into rF2 (I think so, anyway).

    I'm glad to see that you are enjoying racing in AC, it really is a great platform, with so much to explore. It certainly is a title deserving of a long-term place in our playlist. :)
     
    Jack Dyson likes this.
  6. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,346
    Likes Received:
    6,572
    Back to rF2 physics for a moment, I'm sort of pulling from all of the recent threads here but this thread's still relatively active and by now people should know to ignore it if they want, so...

    I'm seeing some discussions and statements regarding the physics and tyres that seem to be missing the point, I was thinking of listing some of those and saying what I think but then that's really too focused on my own views and no one cares that much. So instead I'll just stick to what I think the core issue is (or might be).

    The tyres appear to hold on too well, and are too forgiving, at relatively high angles. The Formula E behaviour (despite the recent update) highlights this, with rally style sliding around hairpins that I think everyone agrees isn't realistic (and if it were beneficial in real life, someone would at least be doing it in qualifying).

    Having said that, the steering-wheel-overuse videos may* illustrate that, but don't really show any performance hack. I don't think the actual cornering ability is increased by extreme steering, but it may help in other ways (flatter grip response with changing angles, temperature effects - I think I may have said this about 20 pages ago in this same thread...).

    *depending on the tyres, the peak slip angle might require a bit more steering than is expected - which is why I think the "I'm turning this corner, now I'm turning the wheel more and the car isn't going straight!" videos/examples aren't necessarily valid - along with the "going straight" part which is overly simplistic... but I digress. This being the case, the logs showing the actual slip curve at play may also not be that relevant, because the issue may not lie there at all.


    I don't think we can look at one particular aspect of the tyres as being the root cause here, there are probably several areas that contribute. But I think given their nature we can start to hypothesize about the likelihood of their relative influences:
    • slip curve - the logical choice for the "angle vs grip" equation, but also quite easy to check for in ttool tests and very likely real data exists to compare to. Perhaps the dropoff amount is an issue, but sometimes I think the peak is barely reached too - and some tyres will actually have quite a small amount of dropoff. Niels fans would have heard him say the same, and I think this is a bit of a red herring in the causes list.
    • temperature curve - tyres start sliding, rubber heats up, grip is lost. How much? Not easy to find a good reference. I think this is an area where rF2 could be more flexible from a modding standpoint, even if the temperature curve mechanism in the .tgm reflects some sort of rubber properties data it would perhaps be better to allow a lookup table rather than a 3-point curve, to allow balancing extreme temperatures and the temperatures immediately surrounding optimal. I feel that trying to tweak this to give more dropoff with moderate sliding will either lead to excessive loss of grip at very high temperatures, or a lack of the same dropoff. (the 'over temperature' point brought closer to optimal, and no ability to have more dropoff higher up)
    • simple static vs sliding grip - I think mantasisg might have mentioned this one, and increasingly I'm also looking at it as a potential trouble spot. It provides a simple mechanism for balancing sliding vs non-sliding grip, and because of the modelling doesn't act as an absolute and nasty switch. Small changes can have relatively big effects too - anyone who's modified a track's grip to 99 or 98% of original to mimic a dirty or sandy surface will know that those small changes can actually have quite an impact on handling. Looking at (only some) of the S397 tyre database tyres, the sliding coefficient tends to be around 70% of the static, but the ones I've looked at weren't just 'around' a figure - the ratios tend to be pretty well rounded (69.99xxx%, 70.99xxx%, etc). To me this stands out as a slightly arbitrary ratio, and will have a definite effect on the feeling as you move from grip to less grip (and back). If you wanted to get fancy you could adjust the sliding curves, but the static/sliding ratio might be a good starting point for a broad effect. I think adjusting this ratio by 2-3% (probably raising static slightly to compensate for the overall lower grip) might have good consequences on the handling without even needing to remodel the tyres entirely.


    And completely disregarding the tyres, I think some of the recent DLC cars also have some aero issues that others have checked into (perhaps more on discord than on the forum), so that's another issue in itself. And for me it's another area where the available parameters perhaps lack some flexibility, particularly in front wing height sensitivity and interaction between the various aero components with dynamic body orientation and missing parts.
     
  7. Remco Majoor

    Remco Majoor Registered

    Joined:
    Apr 22, 2014
    Messages:
    995
    Likes Received:
    1,068
    This has to be the most useful post for the devs. I'll try to get this done attention by the devs (as I suspect they stopped reading this)
     
    Jack Dyson likes this.
  8. David O'Reilly

    David O'Reilly Registered

    Joined:
    Jun 12, 2012
    Messages:
    1,050
    Likes Received:
    756
    Does that make him the "Tucker Carlson" of Rfactor2 reviewers?
    Or just a plain old Shock Jock?
     
    Jack Dyson likes this.
  9. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Morning !

    Yeah I can just about remember December 2019 :(, life changed overnight for all of us here too: no more restaurants above the sea, no more beach time ... whereabouts did you guys visit ?

    Thanks a lot for all the details about the newer wheels as well as the precious recommendations based on experience. That can save time and money - especially with so many unbiased "quality" reviewers about, lol.

    I don't have a lot of room for a full rig and the G25 is so well made, compact and portable. Plus, I wanted a solid wheel for competition - if it gives me 80-90% of the feel, I am happy - the savings can go to track days or whatever. That said I am pretty serious about sim-racing long term. I am going to get a cockpit thing and VR down the line and move all this into the garage. As for speed I seem to have done ok so far whatever I have been able to get my hands on at the time - while (to make a balanced argument) undoubtedly greater resolution and immersion helps focus (my main problem actually) and that does seem to mean better equipment with the newer titles.

    Yes I have the rF2 F-Vee and it is good but quite hard to drive (and that is good). I use also use the AC Russel-Alexis FF car too - that is also very highly recommended by some real FF racers. I know that Reiza did a good job with AMS 1 (it is based on the rF1 ISI engine ?).

    I completely understand about the time commitment for online racing - if you do decide to have a go, consider coming and doing a few with us at RD. All gentlemen I assure you.

    Hope to see you here again soon,

    Ciaooooo :)
     
  10. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Thanks for the post @Lazza : I think this sums it up nicely - I wanted to point out a couple of things that might be of use:

    1) It might not be just the tyre model alone, but the coupling of the mechanics to the ground itself? : ie the real road might have a role to play in this - it might need to be calibrated better ? When training on a green track (suppose) and then going online - the lap times can differ by a couple of seconds.

    2) One might get a good idea about effective grip (dynamic or sliding) levels by transforming any track into a circle of radius L/2*pi where L is the racing line length for a category. From here we can get the average grip level the simulator uses around the circuit for a particular hotlap. Assuming the same tarmac and temperature, one could compare that to a real life hotlap on the same (circular) track (transformation) and see if there is a substantial deviation between the two?

    I was thinking a lot about your post last night and these observations popped into my head.

    Cheers
     
    Last edited: Feb 9, 2021
    Remco Majoor likes this.
  11. David O'Reilly

    David O'Reilly Registered

    Joined:
    Jun 12, 2012
    Messages:
    1,050
    Likes Received:
    756
    Just want to support your points about DD wheel.
    I went from Fanatec CSR to Fanatec Clubsport to DD (20nm Simplicity self build kit).
    While the Clubsport (or later the CSL) lift in performance and realism was big over the basic CSR it was as big again to DD.
    It wasn't about the power, it was the fine detail of high frequency FFB that you get from gripping the motor shaft.
    At first it was a sensory overload for me. I was slower. I got tired from the amount of feedback rather than the force (in many cars I'm at 65% car specific ffb). Then I started to mentally filter it better.
    But the IMMERSION! Another level.
    Very happy I got that DD wheel.
    Would I prioritise VR (which I dont have).over it ? No.
     
    RaceNut and Jack Dyson like this.
  12. lagg

    lagg Registered

    Joined:
    Oct 1, 2012
    Messages:
    3,043
    Likes Received:
    1,958
    I have a DD too and VR.
    Both things are a big step in simracing.
    Could i return to my old T500? Yes. But i'm sure that i won't race without VR.
     
    David O'Reilly likes this.
  13. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    Does this (and other from other) belong to this thread? No.
    Unbelievable..
     
  14. Jack Dyson

    Jack Dyson Registered

    Joined:
    Jan 21, 2021
    Messages:
    23
    Likes Received:
    12
    Relax ... take a pill :)
     
  15. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    2,926
    Likes Received:
    3,876
    Yes exactly, these are very simple mechanics, but they are also the most fundamental.


    TL;DR.... just further thought on same stuff:

    I personally find base sliding friction parameter to be about right being about 55-65% of static base friction parameter. Having static bit higher and sliding slight lower than it maybe would be standard in sirmacing these days, except iRacing :D

    But thats just based on feel and how cars perform comparing to what I can observe in real videos and even pictures (oh yes pictures too, as they are particularly good to see body roll and pitch and attitude of a car).

    For people who are unfamiliar - TGM documentation from July 2013 refers to these most basic tire parameters like this:

    ***
    StaticBaseCoefficient = Base grip coefficient for static friction. Keep in mind this is essentially an absolute peak, perfect conditions, values over 3 are not uncommon for modern racing compounds. Static covers bristles that are static relative to the track surface. Keep in mind that during normal cornering, even toward the peak slip angle, various parts of the tyre contact patch are likely to be static. In other words, toward maximum speed, in a straight line, the majority of the contact patch will actually be static (but not necessarily entirely). Toward the other extreme in a large angle drift, it's probable that none of the contact patch is static (on the rear tyres), everything condition between will be a blend.
    SlidingBaseCoefficient = Base grip coefficient for sliding friction, i.e. Something of a peak coefficient when the rubber is in a sliding state. See WLFParameters and Sliding(Adhesion/Micro/Macro)Curve entries below to vary the friction coefficient over different sliding speeds. It is important to remember that the nodes and resultant LD table primarily define at what point a bristle transitions between static and sliding, while the temporary bristle springs also have a say in this.


    ***

    Few things to note is that those are "absolute peak values". This is a bit unclear to me, because I don't know what are circumstances of peaking. For static I guess it is kind of understandable. But when the peaking friction of sliding happens ? The initiate moment tire begins to slide ? When sliding tire heats up to optimum temperature ? When sliding speed is near 0 ?

    Lookup Data and bristle parameters also has significant meanign, as well as sliding speeds curves. I personally find abrasion curve to affect slidign behaviour quite a lot too, sudden extreme spike of abrasive wear. Haven't done specific tests to realyl prove that. But sudden extreme spike of abrasion curve will cause significant differecne in sliding behaviour and also will bring the flatspot significantly quicker.

    The static curve parameter seems to only influence static friction if I understand this right:

    StaticCurve = The static friction grip fraction relative to temperature curve in the format of (cold temperature, static grip multiplier @ aforementioned temperature, peak temperature, grip multiplier @ peak temperature, hot temperature and grip multiplier for said temperature). The values are blended together to form a wide smooth curve.

    Thats a bit weird as I suppose heat would influence rubber sliding. But it seems like heat influence for sliding comes from "master curve" which is mentioned in WLFParameters description, which is connected to sliding speed curves. Personally I like sliding speed curves as they are good way to alter slow vs high speeds corners stability and feel. But the WLF stuff is much less intuitive, you really must know hat you are doing there precisely lol

    Aerodynamics in the mix complicates stuff further as they with first order influence to grip by altering loads has similar effects, also being speed relative like sliding, more speed more aero gains, but also more potential loss. In general it seems like very high speeds should be truly scary, as static friction will peak with highest downforce levels, and at the same time sliding friction will become more edgy, which should maybe then bring car close to base static friction, but the sliding friction will be near its lowest. At low speeds static friction wont be as high (with downforce car, opposite with lift car), but sliding friction will be closer to its peak, so it makes sense if at low speeds peak slip angles would be more safe, as transitioning from static to sliding wouldn't be as rough, but still we see can see some cars by aliens being thrown WRC style at slow turns, that is definitely a sign of sliding friction being overly high or/and static being too low.
     
  16. David O'Reilly

    David O'Reilly Registered

    Joined:
    Jun 12, 2012
    Messages:
    1,050
    Likes Received:
    756
    I'm sorry buddy if my responses to this thread don't please you.
    I'll try to avoid missing the point and will concentrate in only offering high quality and on topic posts in future, even if the whole premise of such a thread is a click bait video of very dubious value.
     
    vava74, AMillward, Dave^ and 3 others like this.
  17. Havner

    Havner Registered

    Joined:
    Jul 14, 2020
    Messages:
    359
    Likes Received:
    356
    And obviously the fact that it created lots of constructive discussion doesn't really matter, cause you don't like it.
     
    Pawel44, avenger82 and Jack Dyson like this.
  18. David O'Reilly

    David O'Reilly Registered

    Joined:
    Jun 12, 2012
    Messages:
    1,050
    Likes Received:
    756
    I happen to think it did generate some good discussion (although it-the video was clickbait )and that does matter.
    So it's not as obvious as you may think.
    I just get a bit sarcastic sometimes when people decide to become the thought police on such a thread which has wound itself a very winding path of conversation.
     
    Jack Dyson likes this.
  19. Yzangard

    Yzangard Registered

    Joined:
    Dec 6, 2018
    Messages:
    351
    Likes Received:
    173
    There is something that rF2 (or other sims actually, maybe better in ACC but not sure) don't simulate : tire stress under very low pressure. It seems to me that tires can't unseat from the rim in rF2, which should happen when driving with low pressures as we do. I'm not convinced there is a problem with tire model per see, in fact, the more I look for technical information on tires (in this case Pirelli P Zero DHD) the more I realize that what we think we know is shaped by what we see in reality, except that in reality it is impossible to drive as we see our "aliens" do for the good and simple reason that the tires would either be destroyed or stripped off. The forces at work are such that these tires would not hold, in reality the tire is not solidary with the rim, it is constrained but not solidary. In rF2, it is solidary, it can't unseat, so it "resists" those forces.

    When I set tire pressure a bit higher (to fit Pirelli's recommendations), this steering behavior is much harder and tire is sliding way more. It seems to me that the problem may not be directly with tire model but more the consequences of low pressures that are possible to use in rF2. I'm pretty sure real counterparts would allow same driving style if it wouldn't mean tire destruction or unseat.

    The curves we get from telemetry are not directly exploitable because they depend on so many factors that everything is possible and credible. Whether we accept it or not, Niels Heusinkveld was right on this point, the "ideal" curves obtained with the Pacejka model are only theoretical results in a perfectly controlled environment, which is not the case of a car on a circuit, not to mention the fact that the ideal temperatures will also vary according to the friction frequency, which is not modelled by any laboratory model to my knowledge. Therefore, judging the curves obtained is more than presumptuous.
     
    Bruno Gil, RaceNut and Jack Dyson like this.
  20. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,694
    Likes Received:
    1,218
    Yes it is something that has been brought up several times before, it is nice to see someone else realizing that probably when you reach the limit of the scope of the simulation there are other factors that are simply not ...factored in. In the end the result is the same, but one thing is scream around that something is broken, one thing is to realize that in real life other things can happen and this possibility is enough to change a human behaviour. In RF2 tires, and mechanics CAN'T get damaged from contact with the track surface, this is obviously false in the real world.
     

Share This Page