RealFeel settings and discussion

Spinelli · Jun 14, 2015

I would like to tune RealFeel in order to make the most out of my wheel. I'm experiencing the same issues with my Fanatec Clubsport V2 as I did with my Thrustmaster T500RS and a friend's TM TX. I can feel resistance, the wheel countersteer due to oversteer, and things like bumps, undulations, etc. BUT what I hardly feel - and it's the most important thing for me when I race - is the actual tyre grip itself (both, lateral and longitudinal) especially the front tyres'.

E.g. I feel almost no difference from having full tyre grip, to just slowing getting some slip, to being in a good slip angle, to pushing slightly too far past that slip angle (just a bit), to pushing really far past the slip angle, to pushing super far past like an understeering lunatic

. Furthermore, the lack of FFB info is obviously therefore present when trying to lessen the slip angle as you're in the re-gaining grip phase - no FFB information.

Therefore I've created this thread for use to start helping each other with RealFeel values in order to get more "pure tyre grip/slip" information from our wheels and just to optimize this absolutely wonderful game with all our different wheels out there

Spinelli · Oct 2, 2015

Spinelli · Oct 2, 2015

Heitor Facuri Cicoti said:
@Niels_at_home, thanks for all this explanation! I really like the way you do things, even more that you share it with who enjoy this little passion.

Having said that, I would like to clear up a doubt, or suggest something. Not demanding or else, just for the sake of science. Haha..

I do not fully agree with your way to ramp up on FFB (or maybe I just didn't understand it properly), but that made me think of a way to diminish deadzones. What if, in stead of modifying only the initial forces, we carry that to the whole range of forces? I made a graph to try to explain it:

View attachment 106769

Let's take a steering wheel with 30% of deadzone (exggerated for viewing purposes). If we make a plot with calculated FFB on the horizontal axis, and FFB sent to the wheel on the vertical, the default FFB would be the blue line and, obviously, it would result in 30% of deadzone.

If we add a threshold to compensate for the wheel's internal friction forces, and maintain clipping at 100%, we would get the orange line and virtually no deadzone. But that doesn't look good for movement inversion, so a third function seems likely.

Correcting the orange curve with a simple linear start, we can reduce the deadzone in this case to 12, less than half (or much less if there's no need for visual clarity ), and no major problems when crossing to the other side, for negatives values.

There are some parameters we can change in order to optimize the deadzone, and leave only the "SetDeadzone" for the user to try with. I really don't know if this is doable or not, but I can't think of why it shouldn't be.

If this is somehow helpful in anyway, tell me and I can try to explain everything I have in mind. Otherwise, just give me a beer for the effort. Haha

Cheers!

You can sort of do what you're saying by the following two things. It won't as perfect as your method though but in the meantime may be better than nothing for some:

A. Add a minimum force setting (like RF2, IR, AC) of 30% (for your example)

B. Make the FFB non-linear. This is easy to do. First of all, most wheels aren't linear anyways but to adjust linearity, all you have to do is increase/decrease your FFB strength settings from your control panel. Some games also include FFB linearity adjustments. For example, rFactor 2 has a setting called Steer Torque Sensitivity (STS) which adjusts the linearity curve (think of the linearity curve when you adjust the sensitivity of your pedals but for the range of FFB forces instead).

I bet one could manipulate the minimum force, FFB control panel strength, and/or the game's FFB sensitivity/linearity (e.g. RF2) settings to get something close to what you're saying. Obviously your method is the most elegant way though and just-about the only way to ensure 100% accuracy.

Diego Balzarini said:
Can you please post the link to the mod you are talking about?

Thanks

Sure, here you go

http://www.xtremefactor.es/en/descargas/viewdownload/97-mods-gsc-2013/1043-formula-1-1967.html

By the way, it tells you to edit your RealFeel file for this mod, however, the instructions have an incorrect title for the car's RealFeel section so what I did was start the game first, go on track for a second so that the RealFeel file automatically generates, exit game, and then go into your RealFeel file and edit it to match the recommended RealFeel settings. I also copy and pasted the exact one (with the wrong title) included with the mod but I'm pretty-sure that one doesn't do anything because it's incorrectly titled but I did so just in-case.

Heitor Facuri Cicoti · Oct 2, 2015

Spinelli said:
You can sort of do what you're saying by the following two things. It won't as perfect as your method though but in the meantime may be better than nothing for some:

A. Add a minimum force setting (like RF2, IR, AC) of 30% (for your example)

B. Make the FFB non-linear. This is easy to do. First of all, most wheels aren't linear anyways but to adjust linearity, all you have to do is increase/decrease your FFB strength settings from your control panel. Some games also include FFB linearity adjustments. For example, rFactor 2 has a setting called Steer Torque Sensitivity (STS) which adjusts the linearity curve (think of the linearity curve when you adjust the sensitivity of your pedals but for the range of FFB forces instead).

I bet one could manipulate the minimum force, FFB control panel strength, and/or the game's FFB sensitivity/linearity (e.g. RF2) settings to get something close to what you're saying. Obviously your method is the most elegant way though and just-about the only way to ensure 100% accuracy.

Are you saying I can do it right now? Or is it another possibility for Reiza to implement? If the first one, how do I add a minimum force setting?

Also, wouldn't changing FFB strength settings from control panel result in a linear function too? I thought it just multiplied the FFB by a constant, thus maintaining function type, only stretching it a bit. Am I wrong?

Spinelli · Oct 3, 2015

Heitor Facuri Cicoti said:
Are you saying I can do it right now? Or is it another possibility for Reiza to implement? If the first one, how do I add a minimum force setting?

Also, wouldn't changing FFB strength settings from control panel result in a linear function too? I thought it just multiplied the FFB by a constant, thus maintaining function type, only stretching it a bit. Am I wrong?

The FFB linearity changes differently. You have to check your wheel. The easiest way is apparently by running WheelCheck (v1.72 is the latest version as far as I'm aware) and doing the Step 2 Linear test (open program, don't touch anything, click the down-arrow box and select the Step 2 linear test). Once the test finishes, a text file will be sent to your my documents folder. Copy and paste everything in there to a spreadsheet program and make an X/Y line graph with "Force" on the x-axis and "DeltaX" (or the other one with DeltaX in it's name, I don't think it matters) on the y-axis.

The following are examples I made with my Fanatec CSW V2

With regards to your minimum force setting question, no, it's not available in SCE.

Heitor Facuri Cicoti · Oct 3, 2015

Awesome tool! Thanks for the tips!

s2173 · Oct 3, 2015

Hm, I'll try that mod... im curious, because Im coming form GPL... it's probably the rfactor mod conversion tho, and that mod wasnt very highly regarded I think...

Niels_at_home · Oct 3, 2015

I'm not too sure how reliable WheelCheck is.

The problem with deadzone is, if the wheel firmware multiplies whatever comes in times zero at 0 degrees steering lock, we can't compensate for it.

It would be nice to have an 'input vs output' curve that you can edit in an ini file for 'pro' users to compensate for your wheels possible non linearity...

Goes to far for what we're doing at this moment I think, but it will be piled on the 'good idea' list..

Spinelli · Oct 3, 2015

The x-axis is what the game/software (WheelCheck) is telling the wheel to output, and the y-axis is what the wheel is actually outputting.

Minimum force is not perfectly accurate by those graphs because the initial forces are done while the wheel is at rest, which means more friction (initial extra force required to get the object to overcome static friction and start moving or whatever). Once the wheel is moving then the minimum force should be better. I read that halving the minimum force according to that graph is a better idea, or maybe even 1/3. Just some general amounts I've read.

The following was also posted in the rFactor 2 forums regarding WheelCheck not being the final word

I have seen in many post the linearity graph obtained by that external software that compares input vs output. I don't think that dynamic effects are included in that test so it is not useful to check if under fast variations of input forces the dynamic response is being the expected one. It does not show either the delay wrt to the input which takes to the wheel to provide the commanded force.

source: http://isiforums.net/f/showthread.p...the-ffb-in-rf2?p=389384&viewfull=1#post389384

I don't know if that statement is accurate but I thought I'd post it anyways.

s2173 said:
it's probably the rfactor mod conversion tho, and that mod wasnt very highly regarded I think...

It's a great example of how the FFB makes lateral as-well as longitudinal grip loss very apparent. It reminds me of when I used to do a lot of controller.ini tuning to really be able to feel lateral and longitudinal grip loss/gain/recovery. You can totally feel the tyres on the verge of lockup, the instant they lock hard, the instant they unlock and re-grip, etc. As well as when you combine that with steering. It's so front grip focused and maybe even slightly "in your face". It's great and makes driving much, much, MUCH more intuitive by feel rather than memorizing and guessing/assuming where the grip will be, when the grip will recover during a lockup or understeer, all the tiny in-between moments, etc.

Niels_at_home · Oct 5, 2015

Realfeel is physics based, and while we've identified some things to improve the realism, these things would make the forces drop more *before* you've hit the optimum slip angle, as that is more realistic. Currently the FFB peaks before optimum slip angles but doesn't drop as sharply as it should.

Longitudinal force is also physics based in the force feedback. The more slip ratio you have (a locked tire, or wheelspin under acceleration), the more lateral force is eaten away. FFB is lateral force * pneumatic trail, so if longitudinal slip eats lateral force, it also eats FFB.

Of course the rear tires aren't ''used'' in FFB calculations as an input. But if rear tires start to slide, they tend to be attached to the car which means the car will change its yaw angle for example, which will affect the slip angle of the front tires and thus the FFB.

Spinelli · Oct 5, 2015

There is a setting in the controller.ini that adjusts how late/early into slip angle the force drops off. There is also a setting which adjusts how much/little the force drops off once you get slip. Or something like that haha I'm no expert but ya it's something like that.

FFB steer force grip function="x" // Range 0.0 to 1.0 (previous hardcoded value was 1.0) - lower values will make steering force decrease LATER as front tire grip is lost - try 0.3

FFB steer force grip weight="x" // Range 0.0 to 1.0, recommended: 0.4 to 0.9. How much weight is given to tire grip when calculating steering force.

FFB steer force grip factor="x" // Range 0.0 to 1.0, recommended: 0.2 to 0.6 (previously hardcoded to 0.4). How much of a factor the front wheel grip is on the steering weight.

Those settings can make some very large differences to all this. It's not being dynamically calculated by the steering and such so it's technically "canned". It's still reacting/"connected" to the physics though (obviously).

Spinelli · Oct 5, 2015

To anyone feeling like experimenting for more front traction coming fhrough the FFB...

I disabled RealFeel and tried the Brazilian V8 @ Caruaru, default setup except brake force @ 100%, and brake bias @ 72% front. FFB @ high or full. Wheel is Fanatec Clubsport V2.

The front grip loss feeling is much more evident. I can feel the FFB drop off at the smallest hint of the slightest amount of lockup. I can feel grip before I feel it by visual or audio cues. I can feel the edge of grip much more sharply. It's not subtle at all.

The FFB is unrefined though as I haven't tuned a thing yet. The rate at which it drops off is too abrupt relative to the slip angle if that makes sense. The FFB also goes to 100% steering force/resistance even just a few degrees past centre. It also feels less natural/realistic than RealFeel, and less alive in-terms of bumps, road-surface irregularities, etc.

I would adjust the 3 controller settings above, and mix in a high percentage of RealFeel (75%, 95%, around there).

Finally, from what I remember, RealFeel provided more of a smooth steering and FFB while Leo's FFB provided more of a traction loss FFB. So maybe try mixing in Leo's FFB with around 80-90% ofRealFeel. You have to download and then set-up Leo's FFB though. There are many forums around with info on how to do that.

Personally, I'm tired of all this editing but the1967 F1 mod and it's recommended RealFeel settings are a prime example of what RealFeel is capable of without having to touch the controller file.

Niels_at_home · Oct 6, 2015

There are many ways to bypass logical physics based feedback and mix / mess about with a combination of 3 ways of doing FFB..

You're soon getting into a situation with too many variables that you really can't keep an overview off if you're realistic..

But most of all, Realfeel is physics based, mostly fine, possibly improved by us in a future... I don't see why you wouldn't just feel decent physics based FFB and live with it.

But the beauty of ini files is you can tweak as much as you want..

Julio Barros · Oct 6, 2015

Niels_at_home said:
You're soon getting into a situation with too many variables that you really can't keep an overview off if you're realistic..

I agree with you, Niels. In fact, I call this the Project CARS effect, as it has so many variables to tweak that in the end you have no clue what you've done is better/worst than before.
I strongly believe that FFB should be physics based and should have little parameters to tweak as possible. But that is my opinion...

yusupov · Oct 6, 2015

then you get iracing ffb and no ones happy...

Julio Barros · Oct 6, 2015

yusupov said:
then you get iracing ffb and no ones happy..

Little, not none!!

wikke1984 · Oct 7, 2015

Mikaela Andersson said:
Wheel Fanatec CSW Base V1.5 + BMW rim + Formula rim | Pedals CSP V2 | Shifter Clubsport Shifter SQ V1.5 | Cockpit CSL Seat | Handbrake Clubsport Handbrake |

Fanatec Wheel Tuning Menu | SEn: OFF | FF: 100 | SHo: 100 | AbS: 100 | LIn: OFF | dEA: OFF | drI: 1 | For: 100 | SPr: OFF | dPr: OFF |

Fanatec Driver Menu | Wheel angle: 900° | Dampening strength: 0 % |

Fanatec Driver version: 219 | Firmware version: 095 |

Diego Balzarini · Oct 8, 2015

I've been doing some laps with Retro following @Niels_at_home advice.
With real feel at 90% rather than low the wheel is lighter but I feel I have more control on the car, it's easy to correct and prevent a spin.
The forces are light but I just got used to them and I can drive with more confidence.
Driving with more vanilla FFB I feel a stiffer wheel but I can only notice the force change once it has happened.
The only problem I found is in slow corners, where the FFB is so light I don't know what's going on.. apart from the tyres screeching..
Anything I could do to change that?
This is an opinion from an inexperienced simracer

Thanks everyone for sharing your experience and knowledge

Kenneth Jurmann · Oct 8, 2015

We need chip-tuning. Haha. Just like in the car industry. Mapping an engine (like the one(s) in our wheels).

Then we would have some 11500 (?) single paramteters that we could tune to get the forces (at the wheel? Maybe?) correct. Perfectly linear.

I'm also one of those who love to fidle with the ini's. Most of the times just to get a new expirience when I get bored. Haha.

Right now I use full dampening on my G25 (driver-side and ini-side). The wheel loses a lot of it's dynamic but I get the benefit of a more believable weight when turning the wheel into a corner. (A lot more weight, but realfeel ffb stays the same without clipping)

Would love to have a dampening that is not linear but static, which could be set to a higher or lower value.

See... I never get bored. Haha.