rF2 FFB system and philosophy

Discussion in 'General Discussion' started by DrivingFast, Nov 4, 2020.

Tags:
  1. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,639
    Likes Received:
    1,070
    Can you confirm that the rF2 FFB system produces an FFB which is 100% physics engine driven, without any added fake effects ?

    I have noticed that several ACC fanboys question this principle on youtube, and say that there are canned effects.

    I hope they are wrong.
     
  2. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,124
    Likes Received:
    712
    It would be up to them to demonstrate this claim.
    In RF2 all FFB output is result of physic engine.

    Just a story that mean nothing:
    Back in the times of RF1, I was unable to play, despite having a FFB wheel, FFB made no sense to me, I kept crashing everywhere.
    Then REALFEEL plugin came, it promised to replace RF1 FFB with a physic simulated FFB. That was the only thing that allowed me to complete a clean lap in RF1. Then RF2 came, and ISI claimed that his new FFB was physics based, that was insta-love for me, never looked back, never, in 8 years felt a single jolt that made no sense in the 3D world.
    Youtube videos are made to collect visualizations, not to say the truth, truth doesn't win visualizations, being polemic do.
     
  3. Rastas

    Rastas Registered

    Joined:
    Feb 17, 2017
    Messages:
    462
    Likes Received:
    971
    ,They are wrong,canned effects only in rFactor1,and they can be turned off,rFactor2 is pure raw FFB from chassis and tires,but most from chassis\suspecion geometry :)
     
    DrivingFast likes this.
  4. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,639
    Likes Received:
    1,070
    In fact it was not about youtube videos but comments below the video.
     
  5. Mitch9

    Mitch9 Registered

    Joined:
    Jul 26, 2020
    Messages:
    44
    Likes Received:
    38
    The title of this thread is peak DrivingFast.

    And well, I´m sure this could lead to an interesting discussion about whether or not dynamic ffb is always appropiate or there´s a benefit to using canned effects in a sim, kind of like the randomized bumps rf2 makes with the road mesh...

    but honestly, youtube comments? that´s your source??
     
  6. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,639
    Likes Received:
    1,070
    Hey, man, I just asked for confirmation, I didn't say anything.

    For me there is no canned effects, just want confirmation.

    There is no source, just ACC fanboys comments.
     
  7. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    1,769
    Likes Received:
    2,541
    Yeah thats pretty weird thread to come up with in a sim that has been out for years and one of the most unique points about rF2 is probably how everybody thinks that rF2 FFB is really good lol

    No one can claim anything completely, only the people who made the codes, or the people who could examine the codes if they had access and all the required competence to judge it. I suppose that would be quite a task depending on complexity of codes.

    One thought about how people judge simulations, especially these boring uncreative GT3 days. They take few GT3 cars between few sims, and then base whole sim on that. A few cars, especially of the very similar type, manufactured very similarly and most importantly running very similar tires if not the same and also probably made by single man is very narrow take on the whole simulation. A car is just a little package of parameters that a sim works with. A sim is way bigger deal than that. A sim could have all not very well simulated cars, but actually capable of much better, many scenarios are possible. Also it all come down to some chill dude, like me or you, at home who has a final say if that is a great simulation or not lol. And the main philosophy about FFB that devs often seems to use is using sliders for FFB for some DIY simulation for a chill dude, like you and me, who then can share the magic FFB configuration online claiming to have sorted out the sim lol

    There should be no philosophy about FFB in proper simulations, as well as there should be as little as possible sliders, and that could be one of main indicators that the FFB coming is pretty true. The only philosophical part could be perhaps the nominal FFB strength dialed and default setup stuff, mainly - caster and steering lock.

    Most other sims apparently can't get away with as few sliders as rF2 without upsetting some of their fans for how steering/car feels. I put steering vs car, because many simracers can't really tell what is just steering, and what is some legit physical chain effect that influences how car moves and not that much how steering torque is developed. For some reason simracers tend to reduce whole physics to steering torque a.k.a FFB. It is fine. I am working with physics for quite a while now, and I keep discovering what exact effect some actual legit physical parameters make. I slowly get better at pure idea how car works, in rF2 I almost never think about FFB anymore, but about actual physical parameters of car, with exception of overall FFB strength. rFactor2 benefits a lot from physical tire model. All these bristle points, bristle springs, tread deformation, carcass stuff... it makes empirical modeling obsolete. rF2 simulates tire that comes before the empirical figures (things that empirical descriptions are based on), so it is literally ahead of empirical model. Of course there seems to be a problem regarding low pressures, where vacuum tire is perfectly usable, but it seems like it is not some fundamental issue, but rather lacking some limiting factors that would make tire unusable in such circumstances, such as for example - unstable and highly resisting rolling. Chassis flex must be important feature too, if you don't get sub-bodies inertias of rF2 chassis correctly, then it has big influence on how car works, and FFB feel eventually too, of course (something that I have learned recently, and already improving many cars a lot). Separating cars into more and more sub-bodies would probably give even cooler results, but that would probably cause complexity to model the car to go through the roof, while at the same time with highly diminishing results. Chassis geometry is rather straight forward kinematics probably, every proper cars game probably could get it alright, GTS looks like it has some nice kinematics, but it is surely an arcade. Aerodynamics is big force caster for a car, and directly influences stuff like tires friction, yaw moment, pitch moments, maneuverability and even inertia (and rF2 simulates that), so that adds up to car feel a lot.

    I was probably incorrect somewhere.

    I accidently posted stupidly too big post again.
     
  8. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,065
    Likes Received:
    485
    @mantasisg only as a side note, when posting walls of text maybe you can add a TLDR? Cause for me as an example your post is just To Long and Did not Read.
     
    pascom and Remco Majoor like this.
  9. Kahel

    Kahel Registered

    Joined:
    Jun 6, 2020
    Messages:
    98
    Likes Received:
    59
    What the ACC FFB worshippers means by RF2 'canned' FFB effect... is the fact that rf2 FFB doesn't reproduce ONLY the steering rack force of a real car... but instead add force... but not canned force (like rf1) but forces based on the physics engine.

    They have been partially brainwash by the ACC devs whom embraced and glorified the Church of 'only real steering rack force should be allowed'...

    And then of course you have some brainless follwer spreading the holy words around...


    PS: Sarcasm aside... nobody is immune to be a brainless worshipper of some kind... the point should be accepting that but, growing the knowledge base... trying not to just call them stupid fanboy... after all, what they are saying does come from somewhere...
     
  10. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    1,769
    Likes Received:
    2,541
    Of course, for you I can always try to do my best:

    TL;DR to the very long post that is too hard to go through for busy businessmen:

    1. This thread is weird.
    2. No one knows everything except somebody.
    3. People usually only drive few cars and reduce whole simulation software to some parametrization package that cars are.
    4. No FFB sliders is the truth.
    5. rF2 is awesome. I am genius. Physical tire model is literally ahead of empirical model. Chassis flex is cool. Aerodynamics are interesting and matters a lot. GTS is arcade.
    6. I don't have a lot of confidence.
    7. I am sorry for posting long post, I admit a failure.
    Is this short enough ?
     
    Last edited: Nov 4, 2020
  11. Kahel

    Kahel Registered

    Joined:
    Jun 6, 2020
    Messages:
    98
    Likes Received:
    59
    I can get behind most of the 7 points... except the 4th one... Sliders are fine.
     
  12. DrivingFast

    DrivingFast Registered

    Joined:
    Mar 1, 2018
    Messages:
    1,639
    Likes Received:
    1,070
    @Kahel

    What is/are the difference between :

    - Reproduce ONLY the steering rack force of a real car

    - Add force (but not canned force) based on the physics engine.
     
  13. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    1,769
    Likes Received:
    2,541
    Sliders is the worst thing. Steering force is what it is. I am not going to go deep in that and write again something that would require TL;DR. But there are sims that has sliders that makes absolutely no sense. No sense physically. No sense in terms of making simulation more simple and direct to deal with, and certainly not more realistic. Actually the less realistic the sim is the more sliders makes sense, because the changes people will randomly make won't hurt cars anyway as they are not that realistic in the first place, they can always ruin them though, and the community "sorted out" combinations of sliders is also always questionable.

    Also what forces does rF2 add up that are not from steering rack ? I am not quite familiar with that. Could it be by some chance some road noise ? Does that depend on track .gdb parameters ? Anyway noise wouldn't be that much of a force, it is more a vibration, but I am not sure if even that is happening ?
     
  14. Kahel

    Kahel Registered

    Joined:
    Jun 6, 2020
    Messages:
    98
    Likes Received:
    59
    Let's say I make a perfect model that reproduce the force felt on the steering of a Formula 1 1988 McLaren mp4/4.

    Those force are very important but doesn't give, a lot of feedback that the 'real drivers' get... such as g-force for example... Should a sim try to include them in the FFB?

    To my understanding from reading along the years... rf2 does give you more info than just the bare steering rack force... But since I'm far from the best one to give you very specifics details... let's see if someone more knoledgeable jump up in the disccussion to give us a refresh or call out some BS my own church is spreading.
     
    avenger82 likes this.
  15. vittorio

    vittorio Registered

    Joined:
    Jan 11, 2012
    Messages:
    1,065
    Likes Received:
    485
    Thats a very good point! Is rF2 FFB exaggerated to compensate missing forces like G-forces? And if, how? Probably only S397 can tell!?
     
    Pawel44 likes this.
  16. Comante

    Comante Registered

    Joined:
    Nov 20, 2013
    Messages:
    1,124
    Likes Received:
    712
    No, RF2 FFB is everything that has a reflect on the steering mechanism, the "problem" is that much more stuff that usually one think has a reflect in the steering: from rear suspension , to car body movements. is really much much simpler to let the physic settle himself than to artificially inject this or that.

    G forces influence the carbody dinamic, thus the steering.
     
  17. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    8,202
    Likes Received:
    3,589
    FFB output matches steering column torque. Is steering column torque anything other than the force coming through the suspension from the tyres?

    (That's a rhetorical question, hoping that's sort of obvious)
     
    Emery, DrivingFast and mantasisg like this.
  18. mantasisg

    mantasisg Registered

    Joined:
    Aug 17, 2016
    Messages:
    1,769
    Likes Received:
    2,541
    The idea that FFB should compensate for gforces feel is cancer of simracing. Then you get proactive FFB that reacts to yaw moments earlier than it should, and you end up driving a car nothing like real thing. Just letting steering wheel rotate itself everytime there is little yaw rotation. You basically say goodbye to such dynamics like possible neutral steer, four wheel drifting throttle steering, you also refuse to be as active yourself as you would be otherwise. And visual cues are enough for everything to begin with, so there is nothing really to compensate...

    I would go as usually.... - controversial, and say that you actually get more and more real seat feel with a force feedback that is as it should be - forces from steering components. It is so because if rear of a car starts wonder around a little, do some "sideswiming" with little or no cue in steering force for particular reasons, drivers brain should pick up it more because of difference in feedback. In short, you should be detecting a car changing yaw rotation more than intended, or unintended. And it should provide a seat feel. Not exactly only a feel of tiny acceleration changes by your body is responsible for a seat feel. Seat feel has become kind of legendary and mythical term. Should be easy to check, close eyes for one turn at speed in nordshleife in sim, and then do the same IRL lol How is this for FFB philosophy ?

    I have many times expressed how I think AC FFB is proactive, always very early and eager to engage with yaw rotations. In rF2 however it feels like it depends, driver and car is more like a 50/50 team, where driver action is as important as cars reaction. They made it so much more natural in ACC in early release 1, but perhaps it was little too difficult and also people that were used to specifics of AC and heavily believed in that didn't buy different idea on "FFB", which is probably more of a physics job, since Stefano has clearly said then that FFB code was exactly same, but the change was in physics... Here we are years later, and more and more it is FFB, and less and less it is physics in the topics... Is it a good direction ?

    I watched Ermins video recently, where he called rF2 FFB proactive, it left me quite a bit confused honestly. Maybe he meant that from drivers perspective. If driver is proactive, then steering force isn't and vice versa. Anyway that depends a lot on physics too. If car is going to have fast steering ratio, a lot of grip, hard front toe-in, large caster it is going to have more proactive steering than otherwise, or in other words steering will likely do stuff before driver. That also depends on how fast or a wizard the driver i.....

    Ah damn.... long post again. Better stop now, otherwise I'll have to write TL;DR for Vittorio.
     
    Emery and Binny like this.
  19. JimmyT

    JimmyT Registered

    Joined:
    Jan 16, 2012
    Messages:
    825
    Likes Received:
    515
    Too funny. Keep up the good work @mantasisg :cool:
     
  20. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    8,202
    Likes Received:
    3,589
    My own wall of text incoming, not sure to what degree.

    I think you can draw a comparison with judging different game philosophies to creationism vs evolution. (uh oh!)

    Whenever I find myself 'arguing' with someone who doesn't believe evolution is possible, I find the main disconnect arises from their entrenched beliefs and how they would apply to evolution - which doesn't work. For example, evolution very (very!) broadly says that animals evolve to better suit and take advantage of their environment. Someone not understanding evolution will tend to say "well that's silly; how can a fish decide to grow legs?", and of course that is silly.

    At the core of that statement, however, are two things they can't let go of: intelligent design, and a short timescale. If you can't abandon the idea that everything is a conscious decision (by someone/something), the fish has to decide to grow legs. Now obviously a fish can't literally grow legs, so it has to decide to make its offspring grow legs. Preposterous! And it is. So you then explain that what actually happens is an occasional mutation, and some of those mutations will be a slight advantage, and slightly more often animals with that mutation will survive and make that mutation more common in the population. Further mutations can then occur. There's no instantaneous jump from 'fish' to 'fish with legs', there are many many small steps between where a particular environment rewards those small changes. But again this isn't intelligent design; for each successful mutation there are probably thousands if not millions of other mutations that weren't successful (some of which could have led to better outcomes long-term, but didn't work at the time) - it's not a linear progression to better and better (culminating in man). And this is where timescale comes in - a diehard creationist sees the world as something less than 10000 years old. If you take that fish evolutionary process and stretch it over 150 million years, those occasional mutational changes sure do start to add up. (simple magnitude is also at play here. It's hard to really visualise how big a big number is)


    Back to driving games: because of computation limitations many physical models have been very simple. rF1 style sliptables are data driven at their core. Completely bypass the tyre dynamics, plug the real-life (or approximated) output in one specific set of circumstances into the game. Add some parameters to try and compensate for different scenarios. Allow a few sliptables for different directions as per rF1 and derivatives (slip angle, braking, acceleration). Add in some parameters for optimum camber etc.

    Now you can derive FFB directly from the forces the tyre puts out, run it through the simple suspension system. In the old days that led to very weak (generally) FFB on the clunky wheels of the time, didn't have cool effects like engine vibration, so no one did it. rF1 had the pure steering arm force calculated but built other FFB effects over the top by default. As simracers and their equipment evolved RealFeel just took the steering arm force and left it at that, and many liked it. (it also forced modders to use a realistic suspension layout to avoid nonsense output, an added benefit)

    Still, many won't feel that's enough FFB. So it's not surprising many mainsteam games don't take that raw approach and instead fluff up the FFB with extra effects or a little extra help for players.

    Now, I can't actually comment on how dynamic the raw SAF in rF1 is. I don't know to what extent the chassis dynamics plays a part. We know in rF2 the tyre itself is a physical model initially, which then generates a multitude of lookup tables (ha! the critics cry! rF2 uses lookup tables instead of calculating everything live! [as they play their game with a couple of sliptables and some parameters...]), and the entire suspension and chassis is a live beast updated at 2400Hz. If a rear tyre hits a bump, that comes through the (flexible) chassis and will have some realistic effect, discernible or not, on the steering rack. Not because a programmer said hey, the rear wheels should have some sort of effect on FFB, let's make it do <this> if <that> happens, or if <this> happens then ... etc etc, but because the physical model does what it does. (and breaks how it breaks, if the suspension isn't done right, leading to rods pointing in the wrong direction and wheels flat on the road, despite no damage)

    In the early rF2 days I myself thought there were extra "helps" in the FFB, because when the rear starts to slide the FFB steers into the slide very quickly (*or so it seemed). That was me from rF1 looking at the sliptable and saying, "but if I'm already near the limit, FFB will be dropping; if the whole car starts to turn, the tyres go even further across the slip curve and FFB drops more, not rises."

    That's just a misconception and an attempted explanation based on my established thinking. When you consider how many factors actually go into the force calculated at the steering column, and the ways any and all of them might be affected by a loss of traction at the rear and the resulting changes in force, load, roll, etc, you realise such simple statements are misguided.

    People who like one game (team, political party, coffee flavour, pizza, ...) more than another can, without any in-depth knowledge of the workings of either and with no regard at all for their own biases, come up with all sorts of theories to support their view. No opinion is invalid, we can only correct inaccurate "facts".

    @mantasisg I would avoid mixing wheel FFB and 'seat of the pants' feeling. I mentioned earlier that FFB seemed to react quickly to oversteer, but there are different types of quickly. Even in my boring FWD mass produced underpowered passenger car, there have been times when a bump in the road has meant for a split second I'm turning a corner with the front wheels while the rears leave the road and are following only Newton's law of inertia. I'm only talking a very short time, no loss of control, no skidding on landing. From outside you wouldn't see anything out of line. The steering wheel didn't move and didn't convey anything. But even with a standard seat and belt, I could feel that movement through the seat.

    You could try to capture that with wheel FFB, but I think it would either be lost behind the other FFB or end up masking it (if amplified). I think there's definitely some feel to be gained by adding extra feedback to mimic that seat feeling, away from the wheel. (seat slider, motion rig, something - I wouldn't try to elect the best method)

    TL;DR: not missing much really, just a simplified explanation of evolution and misunderstandings, parallel concepts in driving games, and people don't know what they don't know.
     

Share This Page