Automobilista 2 Custom Force Feedback - Overview & Recommendations

Thank you Sean,

These experimental files have a very different damping approach, so should def. start at 0 and only go up from there. Later I might combine it with the older way as well, and also the way LFB is done might change completely....

Thanks for sharing your experience and settings! And for the cheering as well!

Best,

-K

 

Thanks for the QUICK updates Karsten!!
I decided to do a race at Spa, just to get something 'useful' done, lol. A number of times the extra feedback helped me save a notable problem coming instead! The slip information helps you tell that you are heading to a probable total loss of the rear end.

One funny thing that SEEMED to happen was while testing/practicing, when I come out of the pits at Spa and the tires are cool (maybe even cold really), if you have a less information FFB (Default or Default+ or when I had FullFFB) you could roar up around Eau Rouge almost flat out and the supposed cool tires hardly 'made' you have to slow down. I got used to being able to go 95% flat out fine.
But with real information coming in, and then the steering goes very light with those cold tires (and it feels like that is what they SHOULD be doing!) you have to go slower then! It really pushes wide then - versus with other FFB it is as if the game (engine?) says "You have good feeling of grip thus you must really have good grip" ! (and vice-versa for FFB giving you lighter, lower grip, feel)
It makes it seems as if the FFB OUTPUT goes back into the game/sim to decide what grip level it has! LOL.
But I very much doubt they would ever do such a thing!

I will have to test that some more.
I am 'sure' that with lighter steering - due to low grip giving lower FFB strength then - I am NOT steering any differently because of that different feel in the wheel. And the way it runs notably wider, with understeer, is like it is the PHYSICS.

I am not sure if it could be the "opposite" of how having more feedback can HELP you drive/lap better..... and in this case having LESS feedback, from lesser information FFB files, helps you NOT understeer in that Eau Rouge case!! LOL.
Whilst I am 'sure' of the results that occurred, I still DOUBT it could even be true!

 

Hi guys, if you want this is base 3.1 with some modifications made by me, it seems not bad to me.
I tested it in Imola on Mclaren GT3, it's my reference bench.

 

@Leen-q

Reg. the front wheel drives, please find this line:

(front_drive_torque_feel 1.7)
and change it to:
(front_drive_torque_feel 0.0)

I think that fixes the issue.
If you find 0.0 loosens up too much on throttle, just give it in the range 0.1-0.4.
But 0.0 is likely right.

If you find the best value, let me know so I will change it in the next file.

 

This level of performance is only achievable for me with the rfuktor experimental v3.1
idk if I'm disregarding a bit some feeling missing there but somehow this level of consistency I can only achieve it with that file.
I can't really understand much of the difference from the v3.1 to v3.2 but soon I shall test it to try to understand what is going on.

(this race was on a cold track, the date was winter cuz the grass looked so pale instead of green, and I had to overtake lots of lapped cars multiple times and avoid contact from people spinning, otherwise I would have had even better and more consistent laps during the race.
P2 did his best lap 0.1s (1tenth) behind me but then by the end of the race I was 20+ secs ahead of him.
(I nearly replicated my fastest lap of the race 5 times! (0,1 of a sec))

 

This is really awesome, hopefully we can marry better feeling with the high performance, that is what 3.2 tries to do at least. However for a race it prob. matters a lot to keep the forces in check at all times to avoid fatique etc., which kind of goes opposite of more realistic sensations in some cases, like higher cornering resistance. I have an idea for how to fix this, if that is the case.

 

I have done a bunch of races with V3.2 and it is very good. I haven't spent any real time 'assessing' specific areas - such as low speed, high speed, grip reduction per slip etc - but all in all it does a great job of providing all that you NEED. Which makes the racing highly reliable/consistent.
You can't quite truly feel the rear end losing grip with a range to tell exactly where you are on the slip curve, but enough to tell you that you had better begin 'preventative' action before it LIKELY escalates into a danger zone or incident! Plus it helps you correct a slide, much better than if you did not have that FFB information.

One thing that remains a BIG issue is that 'base vibration' at standstill. It might have a link to track surface, because when you are in the pits you get it happening to a lowish degree - even with no ignition on, nor engine started. It 'fades' rapidly to zero (it seems) on any forwards motion. Then if you just drive out a bit but then stop again in pit lane, it comes back at a level exactly as per when you were in the garage.
However, after each race I coast to either the Spa turn 1 immediate run off area, which is bitumen, or I stay on the track near that trun 1 outer edge. When you come to a stop in thise places the vibration comes on and it is HUGE! It shakes the crap out of my setup and you have to exit out quick smart! Thus it seems that some 'forces' - gravity level there ( a sidewards pull from the downhill to the right), or surface roughness, or...????..... are a factor in the equation causing it.

So that is something that needs to be found and removed.
I did those many 'block removals' of all effects type stuff which did not alter it at all.... it was definitely from the RACK. And then I did remove SOME blocks from that 'chain' of building the rack up, to no avail. But removing the rack(!) did stop it.
Some of the rack areas are not simple 'blocks' to remove, so I guess it is in there somewhere. Hopefully it is not in something critical to helping produce the good outcome the FFB has now!
Or maybe another 'weird bug' that needs a 'special' line, or variable, etc to have it prevented - like those other places that say they need that.

Oh, as for the 'tiring forces', of bumps, tread shudder etc etc....I have no issue with that as it is REALISTIC! The steering in such a car type is not like driving a road car, lol! And I am not 100% sure on GT3's but those sort of car types are usually very 'driver punishing' with vibrations, steering weight, steering 'retaliations' (from bumps etc). I only drive/race the GT3's (so far) and link them to real world Australian V8 Supercars (driven via track day events that you can pay to drive[race] them), which are different in a number of ways yet a close enough reference in most things. And they are very 'harsh' to drive! But that is even the fun, rewarding, part of it.... the effort and knowing you are 'fighting' quite an aggressive mechanical beast! And taming, controlling it!

I assume the FX slider does just those???? (I didn't test what responses it specifically did). So with FFB Strength and FX strength, you would think there is enough for people to soften it up if they want/need to(?).

 

FullFFB - New Version​

What's new:

• Changes in front tyre grip/tracking/bite feel
• Changes in response to vertical bumps — now pitch-dependent (follows on-screen bumps accurately)
• Changes in car rotation feel

Still to improve/implement:

• Improve small lateral bumps feeling

 

For the V3.2 (or any of them really) I am using Strength 85, LFB 50, FX 50, Damping 1
But the GT3 per car FFB is 90% - so total AMS2 FFB strength is probably 76%.
My Moza R12 set for 55% in its software.
So I guess this means you get 12nM x 76% x 55% = 5nM (????) - which sounds LOW, but in real use it is not.

This all gives quite a strong 'fair effort' level of control/steering - with a lot of bumps/shudder/kicks, and a good 'feelable' range from tires with high grip down to losing grip. Probably a little bit less than realistic levels - maybe Moza with 10% more would be about that. But various GT3 race drivers (James Baldwin etc) have made comments how the GT3 power steering 'absorbs' quite a lot, thus I opted to go down that bit more than I would have expected (or versus real V8 Supercar type, which are quite harsh steering returns!). Otherwise I might have set the Moza for 70% region (7nM area?).

 

I am struggling a little on the lighter feeling on oversteer, since it makes understeer less prominent and also makes drifting not as real. I think it will be a splitting item. Perhaps we should lower the effect to maybe 10-20%, since oversteer is also simply felt by the steering itself turning. Undecided still, but worth considering?

 

Careful about this, if you are talking about FFB intensity. The only thing that limits R12 "power" is the torque limit setting.

 

Well... I re-tested some other FFB's when I thought about something the rFuktor seems to make you 'see'.....
So firstly, rFuktor gives much more notable drop in grip when there is tire slip. eg steering goes light. This makes it FEEL like the car is sliding a lot more - but I think in reality it IS.
So then I watched from Chase View to see the real line it takes... versus what you FEEL. But you not only feel it, you really partly SEE it - in the visuals ahead changing, where you can perceive the car's true direction of travel, which is not necessarily in-line with its body(!), and also the STEERING ANGLE you are using - for what the car is doing.
On some longer medium tight corners, that need HELP to turn, you then use a lot of power to give rear slip and rotation - and then you see how you turn in X amount, but that is nowhere near enough to do the turn but you used the REAR slip to easily turn the car - optimally.

So in the end I decided that what I FEEL is true to what the car is DOING. eg the steering goes LIGHTER because the rear rotating takes load OFF the front tires. The fronts are not doing as much work then..... though optimally ALL tires are doing their maximums.

Now, you might/could(?) think that maybe the steering should go HEAVIER - maybe because you load the front up more and thus it has more grip, thus more steering weight. And maybe that is TRUE. But the real information gain is about what the REAR is doing. And if the front does slip more (understeer) then it all gets even lighter, but you KNOW that combination/result from experience.
eg You mentally add up the visuals, that you are NOT understeering really, and the RELATIVE lightness - which is a combination of front and rear, but if you were cornering well you won't be getting that front end addition anyway.
So you are sort of losing the front load that you would truly get, in exchange for MORE information about the rear - which is of far more use in sim racing!

I actually thought the rear still did not drop off enough.... so I changed that to 1.5 (up from 1.0), as very rough test idea.... but it felt very good then.
All the other FFB sets have virtually ZERO rear feedback(!), and get heavy into high load turns - like Eau Rouge.

You can't have BOTH..... front load realism AND rear grip loss information!
And I could do Eau Rouge flat out in rFuktor, but not in other FFBs !! It is marginal, but one is 100% with JUST able to do that, and the others are you 'must' lift for a split second at the entry kink. Though the car SHOULD do the same for any cases really, so it is probably a FEELING of the other FFB's that you would need to learn.
But in total, the rFuktor FFB gives way better information that HELPS you lap better and more reliably/consistently. In an intuitive way - probably largely (all?) due to knowing what the rear end is doing.

I am probably biased now... but all the other FFB's just seemed LACKING now - just in that rear end shortfall. In tons of other ways they have all the usual/typical good things. (though they don't have the stationary madness thing! lol)

Also.. I forget now which FFB I used, versus rFuktor.... but I would come out of the pits on cool tires and with rFuktor you 'had' to go up Eau Rouge much slower or slide off easily. But the other FFB it was almost as if the tires were already close to operational temps!
I am meaning to go back to test that stuff, to see WHY.
Like the later 'flat out Eau Rouge' I think it has to be the FEEL that LEADS you into doing something different. The information coming back tells you to 'turn less or you will slide', or whatever, but in reality it would just do the turn faster fine! And probably vice-versa for other cases.

But I will re-test that sometime in the near future as it was a very interesting situation, that would be good to understand.

 

Oh... I have that Torque Limit at 100.....
But if you reduce the Game Force, say to 50%, you then you only get that very weak reduction level - seemingly 50%....
And if you do the Moza 50% and the game 100%, it seems identical result.... still just as weak.
Then at 100% for BOTH.... it is unusable! Way too strong.
Or at 10%.... virtually no FFB, wheel strength, at all....

They would not give the 'game force' if it wasn't going ot do something about THAT input.
I would say that the Torque Limit is how much power IT will use, to satisfy the SPEED required to achieve a move. So the Game/Moza FFB is only the "Continual Strength" the Moza would ever be requested to do (by the game).... eg, cornering you get X nM maintained, maximum. If set for 50% Game Force (or in game) you would get 1/2X nM on that cornering.
But if you hit a severe bump, which needs a sharp kick, the Moza will still not have the base request of game FFB to do it super strong (max), but it will USE whatever its needs to - up to that Torque Limit - to give the fast response required. Because a fast move, even if not backed up by 'ongoing strength' still needs TORQUE to get the wheel etc moving fast.

Setting 50% in game, or in the Game Force, will not alter the RANGE of moves - a kick that makes the wheel turn 30deg wiill still do the exact same 30deg.... it is just not backed up by the force anymore (because the request is now 30deg but with only 50% strength used). Though this does mean a 30deg move might NOT get to 30deg, if the user is opposing that X nM maximum enough.

As far as I can tell, if you set a lower Torque Limit, you will affect EVERYTHING.... quick response AND 'maintained force'.... and it is best left at 100%, because you WANT the 'super fast' capability to produce high frequency fine details adequately.
You would only reduce that 100% if you believed the wheel was providing MORE than you would ever need.... but that is not really the case in any wheel under 20nM I would say.

And I think this is where a "30nM" wheel shows its merit. Not in super strength you cannot even turn! But you set the Game/Game Force to 40% so it is much weaker and realistic at more like 10nM (or whatever you like or think is the real car feel) BUT the wheel's huge Torque can make the moves FAST and track the real signal more accurately. It will USE 30nM on those fast move needs (whatever it needs to acheive that).

THEN there is another problem... that allowing that 30nM maximum means IF the FFB does a strong sustained move, or that super fast high angle move, for whatever reason ( a crash, a bug/glitch), it can be "human damaging"!! LOL
But that is the cost of having it be able to do the very fast fine detail responses.... so you either set it for that, at the risk of super force danger(!).... or you WASTE even having the 30nM mega dollar wheel!!

I have used 20nM (21nM Sim Magic?) and found that was immeasureable benefit in fine detail to me. Irrelevant to driving optimally. Thus after testing I decided that the 12nM region was perfectly adequate for FFB in sim racing. Plus as mentioned, it only needs about 70% region (game and wheel totalled) FFB level to be realistic.
It is a LOT like "HiFi Ears".... where the supposed expert audiophile can HEAR such detail that they need a $20K stereo system - and possibly even DO to some degree (but nowhere near the real need or value for that!).... and some people think that extra (but truly mostly irrelevant) FFB feel/detail is worth that high cost too.

 

I did but nothing changed on the exit pitbox, other fwd has it also with less power force, maybe is it a mini thing. On track however everything feel real good till you go on throttle after a sharp turn, the understeer force is way to strong.

 

With 100% on Torque Limit you can get 12Nm on R12, as long as the signal arrives at is maximum - we call it "100%" signal and it can be achieve in many ways (see next paragraph).
To limit the wheel to just 6Nm (why would you do that? ), you set Torque Limit to 50%.
This may not make a lot of sense for the R12 but imagine you just want 12Nm from a Moza R21?

We have several layers of signal processing. All of them, except the last (Torque Limit on Moza wheelbases), are intended to be adjusted to reach 100% of their limits for optimal dynamic range (however, in practice, the resolution of the calculations makes this less important than it might seem).
So, a multitude of combinations are possible. The problem is avoiding signal clipping in one of those stages.
See this post to get an ideia how it works, but keep in mind custom files can use sliders on a completely different way (i.e., I can make a file that uses FX slider to increase just rear axle forces).

Also, keep in mind that many rF*cktor use GAIN (in-game settings) to also compress part of the signal, not just to multiply signal intensity. For instance, on those files, 40% GAIN does just multiply the signal by 0.4, with "0" compression, but with 100% GAIN it does a x1.0 multiplication AND does a "1" compression factor (see graph below).

 

Please post your ingame and base settings, I would like to make sure I undstand this right.

 

Do you have the best setting in the pithouse and in game for your new fullffb. I have a r9 v3. Thank you

 

"rFuktor ? experimental 3.3"

* Bug fix for weird behavior at standstill.
* Front and rear grip loss feel calculations advanced slightly.

Modify rear/front grip loss impact in the 2 lines:
(rear_grip_loss_feel 0.5)
(front_grip_loss_feel 1.0)

 

Default Moza settings (they are neutral) and 65-0-20-10 in-game.

 

Sorry, I couldn't help myself on this little additional experiment...

"rFuktor ? experimental 3.4"


* LFB slider as advanced compression control.


LFB slider always acts like a compressor ofcourse, but most times it focuses on the low forces, as the name suggests.
The idea behind this LFB control change is that compression on multiple things, not only the low forces, can be valuable for many reasons, and at the same time the degree of how much cornering force, downforce, yadayada we each want to feel is different. A competitive racer might require compression on a high end DD wheel base, simply because it can still be excellent feedback, while not fatiguing when driving for longer periods. A racer on a budget will require the high compression for other reasons. Somebody who wants it to feel as real as possible, or just slightly more varied, might want more final dynamic range on the wheelbase, not as compressed.

So the LFB slider, now, instead of doing low force boost on all or some specific forces, changes 5 variables that each define how much to compress a specific area, like low force boost(but to a far lesser degree), cornering strength and downforce scale. The 5 dimensions are not all making radical changes going from 0 to 100, more subtle changes rather, but they compress multiple parts of the final signal, so gives a detailed control of how the signal is compressed.

The LFB slider is no more "correct" or "real" at 0 than at 100. It simply is different compression. The whole range should work for most bases, IF you adjust the gain accordingly. Lower end bases will want to use higher compression most always, I think.

So, IMPORTANT: LFB in this file CAN and SHOULD be used anywhere between 0 to 100%, whatever compression works best:

1) Start at LFB 50%.
2) Adjust gain to fit.
3) Now adjust LFB up/down depending if you want less compression or more compression.
Lower compression will give you bigger difference from driving slowly and straight to cornering at high speed.
Higher compression will give more equal strength across the board.

You likely will not like it close to 0 or close to 100% - but there's always the odd one out, ofc.