Guide: Optimal FFB settings for rFactor 2 - The key to being in the "Zone" :D

Discussion in 'General Discussion' started by DrR1pper, Mar 26, 2014.

  1. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    Ah ok, cool thanks. Will come handy when i talk to leo again tomorrow. Any chance you know what it was for rf1 and iracing as well?
     
  2. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,408
    Likes Received:
    928
    drr1pper !

    hiya, your collating a lot of info to walk round with in your head :D

    are you planning on making your own wheel or something?


    p.s I could be wrong but from my very distant memories

    rfactor 1 200hz approx ( could be less but higher than 60hz )

    Iracing 60hz approx ( upscaled in Leo's firmware for servo compatibility )

    assetto corsa is said to be 200hz to I believe (although it dosnt feel like it)
    -as a test I've manipulated the files in assetto corsa to run at 60hz & the result was
    "grainy " ffb , it was almost like I could detect the data steps, no doubt this why Iracing had to be upscaled due to a 60hz output

    ( rf2 @ 400hz is how I remember it too )
     
    Last edited by a moderator: May 22, 2014
  3. TechAde

    TechAde Registered

    Joined:
    Oct 13, 2010
    Messages:
    603
    Likes Received:
    36
    rF1 was either 90Hz or 100Hz, can't quite remember which. iRacing was 60Hz last time I checked, however that was years ago so it could have changed by now.
     
  4. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    lol, would love to for fun but i'd prefer to stick with something already on the market that is a tried and tested, well thought-out product....so long as it's not extortionately priced.

    Cool, thanks for confirming the rf2 rate and giving rf1 and iracing rates.

    I want to ask Leo if he still stands by his thoughts about ffb in Why Force FeedBack in Computer Simulators Does Not Work. I was initially convinced by his ideas but now not so sure. I think whilst his solution is the perfect representation of what physically happens in real-life, i'm not so convinced that it's necessary, i.e. that the current method is good/adequate but a few conditions must be sufficiently met. One of them is the simulator in question being used. I'm not just talking about ffb update rates (though i'm not saying 200hz is or isn't enough) but other factors such as the way different physics engine account for more or less variables that need accounting for for a truly realistic car handling and ffb response. Perhaps I'm wrong about this and at the risk of sounding like an rf2 fanboy (which incidentally I am because i tried many other simulators in the past and non of them captivated me like rf2 did the very first time i tried it back in 2012 and to this day) but it could be a physics related issue leading to that feeling of not being completely realistic regardless of how strong the ffb wheel is (in his write-up).

    I want to ask him if he still feels what he wrote back in 2011 to be true and if he has tried rfactor 2 lately.

    If anyone has a question you'd like him to answer, let me know. I'll be calling him tomorrow or Monday.
     
  5. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,408
    Likes Received:
    928
    I was going to mention that whilst there, but bearing in mind he's released a ffb system I think he's happy to have that article forgotten lol
    +I think it was "de-published" a while ago

    although people argue about whats realistic & whats not should we really care that much?
    I guess opinions are down to a personal viewpoint

    mine tends to be that I like to have as much feedback from the sim (through ffb and to my motion setup ) as low latency as possible (I'm unfortunately very sensitive to lag ) and at a high detail rate as can be ( thus why I've spent out on the gear I have )

    what ever our setup we will always be at a disadvantage to real world however good our software/hardware is, but with good feedback and added effects ( that contain non "canned" ) to tell us what the simcar is doing help in some way make up for this loss ( of real world sensations ), our brains will tune into these messages providing they are constant and make some sense

    Although I'm fussy about these things & moan sometimes I think we're at the stage ( with good software / hardware / settings ) where real world can be replaced with simulation or at least that fulfil that "itch" I sometimes get to do a trackday, this itch I was actually getting recently ( I emailed spa enquiring about their trackdays last week ) that itch has gone now I'm getting into the new formula Renault

    The trackdays I've done was mainly for research, eg focus on the ffb from a the ( real world ) car, focus on the bodys movement within the seat (restricted by the tightness of the seat+race harness )
    but these have been restricted as I and I think generally people doing track days are overwhelmed by the high gforces and also the self restriction of not driving closer to the limit so not to destroy someone else's car or ones self
    ( mainly done Silverstone the cars are similar to the formula 2000 )

    Like you I wanted to know everything about the technical side ( especially when I was building motion prototypes etc ) and use to think about simulated ffb and motion options/effects etc and found I was thinking about this stuff
    ( most of which Ive now forgotten ) more than I was actually enjoying it which is what I try to do now

    mentioning ffb/servo's the other day, I brought up the topic with one of my engineering pals in costa coffee which resulted in him telling me the vast difference about motors vs servos for about 3 hours most of which my mind was a 1000 miles away lol ( I caught some of it lol )

    re.ffb wheels unless your building your own ( something I considered until leo officially announced his ...I was in his office the day he announced in stock ) just buy the best wheel you can afford -you get what you pay for (the models available at the moment), don't get caught up in hype like "this wheel is comparable to the best ffb out ther blah blah blah" ...if your with me on what I'm referring too
     
    Last edited by a moderator: May 22, 2014
  6. smartin13

    smartin13 Registered

    Joined:
    Dec 23, 2011
    Messages:
    85
    Likes Received:
    0
    assetto corsa is 333 hz
     
  7. hmaia

    hmaia Registered

    Joined:
    May 10, 2014
    Messages:
    16
    Likes Received:
    2
    It was a typo. I am actually running STM=0.04000, and the wheel is a G27.
     
  8. hmaia

    hmaia Registered

    Joined:
    May 10, 2014
    Messages:
    16
    Likes Received:
    2

    Hello Petros,

    Please do not take this as criticism, I posted it just to see if other users were experiencing the same reactions. I love the 787B mod and really enjoyed the ISF3. I will look forward for the updates when they are available, and thank you for the hard and superb work.
     
  9. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,408
    Likes Received:
    928
    smartin

    333 ? gosh are you sure, I'm so surprised that a lot of people really like AC's ffb, it feels half empty with in-fills of canned effects to me, the data resolution dosnt even seem like 200hz,
    with contradicting info out there I wouldn't be surprised that anyone knows (outside of kunos offices) not disputing the claim though

    p.s pleeeeeease no one flame me for my opinion of AC's ffb :eek:
     
  10. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    well, i was wondering about that myself until i checked the date for his article which was released after he released the bodnar wheel, so i'm not sure so that he minds (or at least he did not mind at the time of publishing the article).

    It's not about whether we should or should not, it's a choice and if discussing/arguing about topics pertaining to areas of your interest end up being (possibly) beneficial to your interests (such as the countless discussions on the forum over the last 5 months about how ffb works which lead to my humungous lap time improvements from understanding what ffb clipping and initial ffb deadzones were really doing to driving performance) then it's a no brainer for me, it's an easy "yes". :p
     
    Last edited by a moderator: May 22, 2014
  11. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    No one should ever flame someone for holding an opinion (unless they're racists viewpoints or in-sighting violence or something of the sort) but if someone disagrees with you they should be allowed to present their reasons for why they disagree.

    Personally, i only tried the AC public alpha when it was first released and did not enjoy it. Much that was before i really understood ffb clipping and AC has had many updates so it's only fair i try it again. Whilst it's certainly possible that another company could produce a software on par or supersedes what ISI has made with rf2, i hold personal doubts/bias that that is unlikely to happen.
     
    Last edited by a moderator: May 22, 2014
  12. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,408
    Likes Received:
    928
    what people should or shouldn't do is one thing ...what happened in reality is another thing completely take a look at some of the closed forums lol


    p.s that article was originally released about 4 years ago ( the original taken down ) someone must have re-released it online
     
  13. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    Ah, well the reasons for those threads closing were more often than not usually due to lapse of common sense from certain member(s) posts. ;)

    Interesting and fair enough. I still wish to pick his brain though to know if he still stands by it and if so to discuss it. I'm not so convinced anymore but maybe by discussing he can show me what i'm missing or perhaps vice versa. It's a very interesting topic to me, made all the more interesting by the fact that the answer to such topics are matters of absolute and demonstrable truths. But they're not always so easily obvious to us but still always attainable.

    As an example, Leo used (loosely in context of the article) an example of how gyroscopic couple is (mostly if not completely) why motorcycles remain stable at higher speeds. But if Leo did in fact make this 4 years ago, he would have been right in his thinking amongst the scientific community but as of recently (no more than a few years ago) it has been scientifically demonstrated to be untrue.

    My point being, discussions are good for all parties involved as it allow us to question and check each others ideas, to everyones benefit. No pride or ego involved, just the pursuit of truths.
     
    Last edited by a moderator: May 22, 2014
  14. Adrianstealth

    Adrianstealth Registered

    Joined:
    May 28, 2012
    Messages:
    4,408
    Likes Received:
    928
    I didn't ask the question to really not invite the obvious answer " we've made our force feedback system to simply address a lot of the negative points we made in the article about using ffb in simulation"

    completely true or a sales pitch answer ...that's the obvious answer I'd expect ( & give myself ), in no way am I meaning any disrespect to leo & his team in this post.

    I keep meaning to do a review of this servo wheel, but I'd have to re-beef my brain up again on motor vs servo which is really the big selling point and the reason for the price point
     
  15. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    I'm a bit confused by what you mean here but in either case, nothing you've said comes across as being disrespectful.
     
    Last edited by a moderator: May 22, 2014
  16. Jamie Shorting

    Jamie Shorting Registered

    Joined:
    Sep 11, 2013
    Messages:
    2,628
    Likes Received:
    1

    Still the same.
     
  17. smartin13

    smartin13 Registered

    Joined:
    Dec 23, 2011
    Messages:
    85
    Likes Received:
    0
    the ini.file shows 333.
     
  18. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    Ok, just spoke with Leo and he confirmed that he still believes what he wrote in the article to be true. He explained/discussed it in a lot more detail and it makes complete sense to me now. I asked him if the issue was still present in rfactor 2 to which he replied, yes.

    Here is the article he wrote back in 2011: http://dl.dropboxusercontent.com/u/17548791/FFBdontwork.pdf

    In short, the issue is as follows:

    The current method involves sending a force output to our ffb wheels and very shortly after (e.g. 2ms for a 500hz update rate) returns the wheels rotational position back to the sim. However there is a fundamental flaw with this process, the way in which our ffb wheels are designed to work also happens to directly affect the accuracy and outcome of the physics model (thus car handling/feeling) within racing sims. To give an example (that leo used), imagine your in a real car and you suddenly let go of the steering wheel, the car will take it's own path (be it a stable or unstable one). Now imagine the exact same scenario again but this time without any steering wheel attached to the steering shaft at all. The car would take, for all intensive purposes, the exact same path. The reason for this is simple, the steering wheel on a real car has basically no affect on the outcome of the cars handling. The moment of inertia of the steering wheel of a real car is insignificant compared to all the forces that make up the net forces acting on the steering arms, connected to the steering wheel. In short, the steering wheel will change position based on the net forces acting on it without putting up a fight.

    The opposite is true of our ffb wheels. Firstly, they have moments of inertias that are no longer insignificant and that can vary based on how hard we squeeze the ffb steering wheel. The other issue is that because the sim only sends out a force output to our ffb wheels, it is easy for the wheel to overshoot a wheel positional that it would not have otherwise been overshot in real life. This is most visibly obvious when for example we setup a car (in real life) so that it would make the steering dynamically stable if you were to let go of the steering wheel and induce a sudden disturbance, however with our ffb wheels it either becomes dynamically unstable, a self-sustaining oscillation or the least worst of all, takes longer to return reach the stable state. The quality of the motors used alone is not sufficient to overcome this problem.

    What leo is saying is that even with a high torque, fast responding motor such as in the bodnar wheel, the affect can still be felt.

    He suggests the system be flipped (which would ironically make it to work exactly how it does in real life), so that the correct wheel position (based on the net forces in the physics dictating a change in position) at each time interval is always met and never overshot. In order to do so, requires locking the position of the wheel and measuring the force/torque applied by the user. The force/torque input is measured by a force/torque transducer and returned to the sim. The sim calculates the net forces (this time including those of the drivers input) and determines in which direction and by how much speed and acceleration the wheel changes to it's new position (again, based on the net forces including the drivers input forces). The process is repeated with each refresh of the physics cycle. Whenever the net forces reach equilibrium, the motor is instantly locked into position (be it possibly by mechanical means or by a servo type function that prohibits overshooting) rather than having to first overshoot and then reverse direction to reach the equilibrium steering position (as used in the current method/system).

    Tim, someone once said on the forum that ISI should get involved in making a software implementation so that should Leo make such a device it could be used with rf2, to which you responded that Leo can always contact ISI via email to get something possibly rolling. After speaking with Leo and John (his business partner) they are very much interested in making such a device only that they require a sim company to assist in making the software available so that they know it's worth pursuing. It seems like (correct me if i'm mistaken) that both parties are in fact interested. This would be a first for the ffb wheel market and as discussed, at least in the beginning there is a clear and easy market at the professional level that use rfactor pro. It's an exciting possibility that could change the level of realism in sim racing and cascade down to consumer level wheels as well. What say you ISI? Interested in playing a pivotal role in direction of the ffb wheel evolution and market? :)

    Bed time for me now, very late in hk.
     
  19. Rik

    Rik Registered

    Joined:
    Oct 5, 2010
    Messages:
    1,172
    Likes Received:
    9
    Now with JSON i have, by default, "steering torque zero-speed mult":0.3
     
  20. DrR1pper

    DrR1pper Registered

    Joined:
    Apr 29, 2012
    Messages:
    3,286
    Likes Received:
    22
    Why are you telling us that?

    You could do that via the controller.ini file already, no?
     

Share This Page