FFB settings with the AF responsive mode. Would like some feedback.

[Paul Loatman]

Blank

 

[Spinelli]

This is a bit off-topic but do you have to adjust the FFB force of your wheel in order to make-up for massively lowering the game's FFB curve/linearity from 1.0 (linear) to 0.275?

I remember with the T500RS, you raised the control panel's FFB to 100% to make-up for lowering the game's FFB linearity (0.275). How do you figure-out what strength to set a Thrustmaster TX/T300RS's or a Fanatec CSW V2's control panel FFB strength to when lowering the game's STS from 1.0 to 0.275?

I ask this because I liked using your settings (T500RS control panel @ 100%, RF2 STS @ 0.275) better than "default" (T500RS control panel @ 60 - 72 %, RF2 STS @ 1.0), in-fact, I found the experience even much better than my current CSW V2 setup.

 

[Marc Collins]

Good question, Spinelli, because by moving STS so far off the default/linear setting, there is bound to be repercussions. Some of which you won't even realize (with certain cars) because you are using an "extreme" setting. I prefer to leave everything at or very close to default and let the car modders figure-out how to do proper simulation at base settings. By using non-linear anything, you are distorting the direct simulation feedback.

If extreme non-linear feels better (and still realistic), then I would suggest there is something wrong somewhere in your set-up. The real-world is linear. Extreme non-linear should only be necessary to mask defective hardware...and that's part of the reason most of these settings are not in the UI. Shouldn't need to play with them.

And this advice is not specific to AF wheels. Perhaps there are driver differences with those wheels that are non-linear in the opposite direction? Or, you just have to turn things down to prevent the problem you are experiencing.

Lastly, I can generate the same unpleasant handling/response issue that you have described, with all default settings, just by tightening the steering ratio so it is too small relative to reality. This exaggerates the forces (compresses them all into a smaller range), which makes the wheel behave badly as you describe). You would assume that the FFB would be the same strength and characteristics, and only the steering ratio would be affected. But no, it changes the characteristics so that the control of the car is affected. If I recall, the change results in different FFB if you change the ratio in the Windows control panel versus using the rF2 setting. Or maybe they just interact in strange ways? Regardless, I think something along these lines is affecting your set-up.

 

[DrR1pper]

Thanks for the post Paul. Your observations would seem to suggest then that software was indeed hindering the accuforce wheel motors performance envelope (regardless of whether it was for the better or not), i.e. by converting the output response curve from linear to non-linear. If the original firmware/software allowed you to stop using (what was a lot of) Steering Torque Sensitivity in rf2, sounds like the prior firmware's/driver's had something similar to an STS applied by default. It's very plausible that that is the reason for the perceived lag/delay in responsiveness of ffb details on the accuforce vs the bodnar and OSW wheels when Barry (of the Sim Racing Garage youtube channel) made his comparison videos with 10 or so other people making the same exact observation.

I'd be really interested to here how the wheels fairs against the bodnar and osw again with the new firmware/driver update installed. It could provide quite a different result now that the ffb output is no longer being filtered by an STS equivalent on the accuforce firmware/driver by default (again, regardless of whether it was for better or worse).

I believe i can see now how/why you feel-again/felt the need to use STS at all. Since using a lower STS value should make the steering wheel behave more tame like and closer to real-life behaviour in general. Still not correct but can be much better in it's general behaviour than without any STS which makes for a very skitish ffb wheel that requires a death-grip to control rather than a light touch/grip. The negative of this approach is that it can massively mute fine ffb details and cause noticeable ffb lag/delay (which manisfested itself when the reviews said the details on the accuforce felt a muted and delayed vs the bodnar and osw which now seems to make all the more sense.)

But still, my reasoning could be wrong and one way to check it would be to see if the wheelcheck.exe step log 2 test result is different from what it was before. If you have the time to spare Paul, i would really like to see a comparison of it's before and after responsive curve please. If not, no probs.

Lastly, really fantastic video you've found and posted in the OP!

 

[DrR1pper]

If extreme non-linear feels better (and still realistic), then I would suggest there is something wrong somewhere in your set-up. The real-world is linear.

The world may be linear (with respects to how laws of physics behave consistently in reality) but the current ffb solutions employed (on all sims) to simulate reality are patently flawed though (but it's not the game developers fault either) and with all due respect most people simply can't/don't see how and why because they lack suficient understanding of physics or time spent analysing the problem to understand the problem or to even realise/notice that there is a problem. I've been over this before in the past so i don't wish to go there again but just know that what paul is describing is a fundamental problem of ffb and that using tools like STS can help to alleviate certain areas of it's inaccuracies. The solution is still far from perfect though and some people will in fact still prefer keeping the bulk of those inaccuracies in order to maintain the lively feel of the force feedback coming through their ffb wheel (but in the end, having more of one means having less of the other unfortunately). You're free to disagree with me ofc but i personally won't ellaborate on the subject further. I know that sounds like i'm rudely shutting down the/a conversation (with me) but plese believe me when i say it is not meant that way. For example, I'm not trying to shut you down because i believe i'm some authority on the matter telling you that you must believe me. On the contrary, disagree and discuss away if you're understanbly unconvinced.

 

[DrR1pper]

Like i described in the last post, it changes where you feel the peaks, it will do this regardless of the car, track, corner, or car setup.

The way you describe how lowering the STS feels on the ffb output again Paul is as we've covered in the past, is caused by nothing more than the result of a non-linear ffb output. And how you feel it's affect on the ffb is consistent/matching with how STS affects the ffb linearity.

Also, people have already measured the linearity of the AF, and it's pretty much dead straight, up until it reaches the soft coded speed limit. It's not an issue with the linearity of my wheel.

I'd be curious to see if that linearity graph for your wheel is different now with 'responsive' mode. If i had to guess, the part where the curve plateaus might have been made smaller as a result of this new mode. Or maybe not. Only one way to know, test and find out (pretty please ).

 

[DrR1pper]

By peaks, i don't mean the peaks of the actual FFB. I was referring to the peaks as described in that video i posted, the peaks of the mu-slip curve, those peaks are dependent on what the car is doing and would probably never be the same as the actual FFB peak torque (nominal max torque).

Ah, my mistake.

In a real car, when you turn into a corner, you can tell when you've started to turn too much because the torque will start to drop off, that's when you've past the peak and slip angles will be at a high point, it's exactly as he's describing in that video. With the default linearity in rF2, you don't get that drop off until you've clearly turned far too much, so you never both looking for it at all and that's just not realistic.

I know what you mean and can see how STS will reduce/eliminate that problem.

Also, i haven't tried this with the "high" mode, so it's possible it makes a similar change with that. I'll test it later today with that mode and see how it feels.

Here's a link to 2 step tests, one at a step time of 300 and the other at 150:

https://drive.google.com/folderview?id=0B9wekbxGpXYacnZTVm9ycFRZeFU&usp=sharing

Do you know if "high" mode has been left the same as before? (i.e. a software filter acting similar to how STS affects ffb output in rf2).

 

[Joe]

One other thing i've done is to use a steering resistance type of 0.5, which somehow combines the two types, or goes between them or something and provides a bit more balance for subtle slip angles. This was actually something that Comante thought of and suggested to me last year when i was still using the T500. I didn't really notice it too much with the T500, but with the AF it can be felt and works pretty well. So thanks for that idea, Comante. All of the other settings are pretty basic and necessary for the AF to work properly. I'm not sure if the STS at 0.275 is ideal for the AF, but it feels better than the default to me.

The only downsides to the FFB that i feel are the lack of a realistic resistance, which is the same in either case. It doesn't necessarily mute a lot of the smaller details either, i can't say that it's not lowering the smaller details, but what it does with them feels more realistic.

The standard HID Effect Reports do offer those:
Spring
Damper
Friction
Inertia

rFactor 2 somehow did not offer the Inertia effect. But they only give either Damper or Friction, not both. This is understandable since Damper and Friction is really close same on the the wheel, not for other HW device.
On setting:

"Steering resistance type":0.5,
"Steering resistance type#":"0=use damping, 1=use friction"

the 0.5 maybe invilad since they did not give your such range. It either 0 or 1, though.

I think you effectively disable the in-game resistance (friction or damping).

I saw your setting on AF wheel, which set those Damper, Friction, Inertia pretty high.
I would not use AF for those, those from AF SW (art effect), not from the rF2 game outputs.
Just a thought. My PC needs to be serviced so I cannot run on my AF wheel now.

 

[Spinelli]

He actually posted that in a thread about a Blancpain GT driver who drove on a motion sim with an AF wheel in Assetto Corsa (i think). I posted about this somewhere else recently, but basically the driver says that the feedback from the wheel is exaggerated and unrealistic. His actual car would have power steering, so maybe that's the issue, but then you would think he would simply say that it's heavier, not unrealistic. Most racing drivers can tell if they have power steering or not, and having it or not doesn't significantly change the behavior of the feedback as long as the steering still has a mechanical connection to the front wheels and isn't some steer-by-wire nonsense.

What did you expect him to say, that a videogame is a perfect recreation of reality? I'm not surprised by his comments at all, in-fact I was half expecting a real driver to say that. All videogames have different physics engines as-well as different programming with regards to FFB. Saying the FFB is identical to real-life would be like saying physics are identical to real-life, or audio, or graphics, etc. It simply isn't and I don't expect anything to be so imperceptibly close for at-least 20 - 50 years.

Even the iRacing cars when being run under 1:1 FFB torque settings clearly display moments which are overly violent compared to real-life. It's no different than physics audio, graphics, etc. relative to real life. Every game will be different from one-another due to different physics engines as-well as FFB programming, let alone from real life. Real-life, in a way, is just another physics engine as-well as type of "FFB programming" just like our games are, it just so happens to be the "physics engine" that every game is trying to replicate.

 

[Spinelli]

Even putting aside the physics engine and FFB "engine" (FFB programming) themselves from each game, there is also a countless amount of hardware FFB settings (from the wheel's control panel), as-well as in-game FFB settings, and also built-in electronics/firmware which can make a difference.

That's a whole ton of things which can make countless, endless, different experiences with the FFB. Basically it leaves you with millions and millions of variables to potentially make the FFB amazingly realistic or terribly unrealistic therefore, to me personally, his comments aren't representative of anything concrete other than regarding the exact combination of game (and build version), in-game FFB settings, wheel (and firmware), and wheel control panel settings he happened to use.

There's a humongous amount of variables at play to be able to come to any conclusion based on that experience.

 

[Joe]

I just tried using the in-game resistance values and they don't do anything....

Did you try with non-zero values below:

"Steering resistance coefficient":0,
"Steering resistance coefficient#":"Coefficient to use for steering resistance. Range: -1.0 to 1.0",
"Steering resistance saturation":0,
"Steering resistance saturation#":"Saturation value to use for steering resistance. Range: 0 - 1.0",

 

[Spinelli]

I definitely agree with all that.

Do you think that STS at 0.275 (essentially lowering RF2's FFB output linearity from fully linear [100%] to just 27.5%) is the best way to run FFB with almost any, if not all, wheels?

I have to say it again: running my T500RS with your FFB settings of a very un-linear and boosted control panel strength setting of 100% + RF2's STS @ only 0.275 felt better than having the wheel at it's most linear setting of 60-72 % + RF2's STS @ a fully linear 1.000.

Do you think it applies to all wheels (maybe not the control panel strength part but the STS @ 0.275 part)?

 

[Spinelli]

Ya. I'm going to experiment with my CSW V2, and my friend's TX, @ 0.275 (STS). Thanks for the help man.

Would I be correct in guessing that the AF feels better with STS set to 0.275 (as opposed to 1.000) regardless of whether you have it in 13 Nm or 16 Nm mode? Regardless of whether you add a bit of control panel friction? Control panel damping? Etc? Basically, does RF2's STS @ 0.275 as opposed to 1.000 seem to, more-or-less, always feel better?

 

[Joe]

Yes, i tried coefficient at 1 and -1, and saturation at 1, and it doesn't do anything. Without using the dampening, friction, and inertia settings from the AF itself the wheel oscillates quite a lot and feels very "loose".

Did you try this with following non-zero value:

"Steering torque minimum":0,
"Steering torque minimum#":"Minimum torque to apply in either direction to overcome steering wheel's 'FFB deadzone' caused by friction"

and different following values:

Steering torque zero-speed mult":1,
"Steering torque zero-speed mult#":"Multiplier at zero speed to reduce unwanted oscillation from strong static aligning torque"

If I were you, I would turn off AF profile and tune all AF setting down to zero. So make sure you have pure "physics raw feeding here from game". Make AF gain 100%. Start treating with AF as if a pure "linear amplifier" (analogy to an expensive audio power amp without a cheap equalizer or preamp connected to color the sound), then adjust in-game settings only, systematically, one at time.

 

[Euskotracks]

Too many parties messing with FFB settings at the same time (ISI, wheel manufacturers and users) can bring no good results.

The FFB is calculated by the physics engine. I don't understand why so many FFB settings if there are no settings for the physics of a car. It makes no sense*.

The wheel should be built in order to deliver those forces as reallistically as possible. Nothing else should need to be done to alter it. However they do offer different settings to adjust.

End users should do nothing but drive. However they do try to improve FFB by today touching those parameters that both the sim and wheel manufacturers provide.

With this approach it is difficult to get anywhere.


Enviado desde mi SM-G130HN mediante Tapatalk

 

[Euskotracks]

The FFB is indeed calculated by the physics engine, however, anyone who uses a direct drive wheel with rF2 would find that there are all sorts of issues with the way FFB systems in racing sims work currently. I know of many of the problems, but won't bother going into the details. Just know that, the system for the FFB currently in use isn't anywhere near perfect and has it's flaws, this isn't just an rF2 problem either, it's a problem with the hardware itself and not so much the software. You would need a completely different hardware architecture to fix these problems.

So you are basically saying that wheels cannot accurately provide the FFB that the sim commands. I would expect so in a cheap wheel but not with an industrial 2000$ DD wheel.

Since it is not a torque capability problem with those DD wheels, I understand that you are referring to error in magnitude and/or delay in time.

This inability to accurately provide a commanded force while the driver is turning the wheel, as you say, should be assessed by the electronics controlling the servo motors.

The fact that they cannot provide a solution for this, doesn't say much good about these wheel manufacturers. The PID tuning should be done by them. They know the input and the output. They have all the information required. If they cannot get such a default setting that directly works plug and play, it would seem to me that they don't know what they are doing. They seem to me to be playing to be engineers.

The end user plays a guessing game with no data, hence with high probability of messing it up even further. A static linearity test is crap to do this. The problem for sure is dynamic and not static.



Enviado desde mi GT-I9505 mediante Tapatalk

 

[cookie]

Hi Paul I gave your settings a try, I could feel the fine detail but I didn't like the light feel in the center.I use Foundation FFB and front suspension bumps only on default mode , have you given Foundation FFB a decent try? apparently it uses the physics engine only and no canned effects, the only thing I don't like is the lack of downforce effect.

 

[Joe]

You can't turn those settings off in the AF and also avoid oscillations. This FFB system isn't perfect and requires a lot of fudging to get it to feel like an actual car….


]Yes, i tried coefficient at 1 and -1, and saturation at 1, and it doesn't do anything. Without using the dampening, friction, and inertia settings from the AF itself the wheel oscillates quite a lot and feels very "loose".

I turned OFF ALL AF wheel settings, only using Game FFB set to 100% gain. Works “perfectly”! No oscillation and “loose” at all. I tried Corvette GT2, Nissan GTR GT, and BT20 on Silverston.

I am using defaults on Jason file:
"Steering resistance coefficient":0.1,
"Steering resistance coefficient#":"Coefficient to use for steering resistance. Range: -1.0 to 1.0",
"Steering resistance saturation":0.2,
"Steering resistance saturation#":"Saturation value to use for steering resistance. Range: 0 - 1.0",
"Steering resistance type":0,
"Steering resistance type#":"0=use damping, 1=use friction",


Some ting is not right here. Did you upgrade latest Sim Commander 4 SW?

On
"Steering torque sensitivity":0.275,
"Steering torque sensitivity#":"Sensitivity curve applied to representable torques: 0.0=low 1.0=linear 2.0=high",

I tried too. Not realistic to me at all. I could not imagine a GT racer steering at that low sensitive wheel (extremely non-linear). He may crash right away. I watched several onboard videos for real-life Corvette GT and Nissan GTR drivers. You can see yourself. For such low sensitive steering, maybe for drifting, I will think.

Here is video of GR-T GT steering sensitivity, life video:

I believe GT-R GT factory steering max angle is 540. At that low sensitivity (0.27), one may not be able to make a U-turn(?)

 

[DrR1pper]

STS has no affect on maximum steering angle. STS only affects the torque output from your ffb wheel.

 

[Joe]

I feel effect on delta(front wheel angle)/delta(steering angle), as STS changed. Sure, the max steering angle did not changed, but max front wheels angle will (?)
Say before I need to change steering angle x to move front wheels change y angle. Now after set that small STS, I need to go >>x steering angle in order to move same y angle of the front wheels.
YOu can tell right away by comparing STS 0.27 and 1.