Automobilista 2 Custom Force Feedback - Overview & Recommendations

Yes I did and even the JCW Mini ffb but everything stays the same, to heavy on the exit pitbox

 

But doesn't the LFB value always have to be zero?
3.1 is better for me.

 

Interesting, will be good to figure out what is better in 3.1 and in 4, so best combination can be made

 

Try halving the values you increased in this last file.

 

I have some first impressions feedback, idk if I should share in PM or here but in any case this might me more informative and more people might agree or disagree with what I say and that might bring a healthy debate over ffb.

First impressions on rfuktor experimental 4:
A lot heavier (between 10 and 15 +gain feel difference to previously)
A lot more weighty feel to the car
More abrupt transition from heavy to light when locking up)

Now I don't push into understeer as hard, but I push the rear tires harder into overheating with the throttle.

v4 is closer to default+, almost feels a bit of a hybrid between v3.1 and default+ regarding feeling of more front vs more rear. but V4 seems at first over corrected and heavier than both v3.1 and default+. with v4 somehow it's harder to push over the limit as push strong slip angle as hard making me more consistent and cleaner with ease of matching laptimes while being a tiny bit slower. it's also harder to correct as fast just because even when correcting things feel heavier.
somehow it gives me the impression that I don't feel the rear as much as in 3.1, but now I can feel better front grip in v4 in a way that I didn't feel as much in 3.1

v4 also requires less gain, clips more often but the ffb telemetry % levels are more spread out across all the bars where before the ffb was spreading less across varying degrees of force.
(but this could be just related to ingame ffb gain % compared to base, etc)


I think it might have been a bit over corrected in a sense because the difference in ffb power is quite big. at the same time this v4 gives a bigger level of "confidence" in the ffb and many ppl will say they feel more connected with the car than with 3.1 because 3.1 did feel a bit more "loose" as an ffb canvas. but somehow after testing I was actually faster in 3.1 and still able to find consistency with it.

I've just tried default+ and it definitely feels quite a bit lighter than the new v4 rfuktor. I was also able to go slightly faster on it but with rFuktor I'm more consistent (both 3.1 and 4).
With the v4 after I set a pbest, I can lap doing variations of that pbest that are 0.010 to 0.090 away from it. TBH with v3.1 it's possible to do similar but a bit harder due to having a bit less weight feel and everything being slightly looser. It's really hard to tune this without changing how "snappy" this 3.1 feels when correcting and how clean the car slip feels.
With 3.1 I still get a slightly faster pace.
My fastest lap came from Default+ but again, it could be a fluke. the diff is too small, and with default+ I struggle to be as consistent.
3.1 feels the least dampened.

It's really hard to judge because the strength difference is quite notorious and affects the overall feel a bit, you could test me and say you changed tire load or something and in the end left it the same and just bumped or compressed values to make everything stronger and I would be a fool.

Idk if it's something more going on in the ffb, or if some things just come into the forces earlier. Due to that, I'm more precise and consistent and less overconfident but somehow I push less the limit and the gain must really be tunned down. In the end idk how far down I would have to go to be at 3.1 force levels but even then idk if it would still be as informative or I would just "grip" the details meaning holding the wheel to tight to compensate for forces and my own hands muting the details.

It really didn't help that we had an update, things changed, I had to re-do all the laps on Default+ and rFuktor 3.1. I like the consistency of the new version but it feels like it might have been overdone or overcorrected, but this is 1 car at 1 track... now I need to do a few more races during the week and try more stuff out.

Sadly one of my best races in LFM "king of the streets" GT4 at Long Beach I was performing quite well with rfuktor 3.1 and then I missed 3 litres of fuel in the last lap otherwise I would have gotten P3 and that race was even livestreamed by LFM own twitch. Really sad that AMS2 miscalculated that fuel.

Gineta G40 GT5 at Brands Indy "memorable" hotlaps of today:
rFuktor 4.0 (50 0 25 0) 0.54.003 (best w this ffb file)
rFuktor 4.0 (49 0 20 0) 0.54.033
rFuktor 3.1 (60 0 20 0) 0.53.943
rFuktor 3.1 (60 0 20 0) 0.53.936
rFuktor 3.1 (60 0 20 0) 0.53.929 (best w this ffb file)
Default+ (55 0 0 0) 0.53.903 (best but very hard to replicate)

(post update default+ feels slightly lighter than before? I ended up bumping from 50 to 55 it's gain)

I have videos of all the laps mentioned above... maybe later I can edit and make a comparison compilation while showcasing my current #3 time trial lap for GT5 at Brands Indy.

I haven't tried cars on the "opposite end" of the spectrum meaning LMDh or that kind of car where the v4 file due to it's weight might be more helpful in avoiding mistakes from excessive inputs.

Right now it's odd, the ffb that makes me feel the least confident in my driving skills (rFuktor 3.1) is exactly the one that better allows me to consistently lap in the 0.53.900s

Again sorry for the huge post if you have to chew through all this but I've tried to be as informative and as useful as possible.

 

@carloscepinha @Marnus Zwarts @Leen-q @spyke70

Thank you for the feedback! Great to have from multiple cars/tracks/drivers, will try to adjust from those angles and hopefully it leads to something better than both 3.1 and 4
Found a cpl additional areas of possible improvements now bc of your posts as well, so this will be interesting!

Best,

-K

 

New I imagine that the latest update of Ams2 must call into question at least some of the custom ffb?

 

In which way do you mean?

 

"rFuktor ? experimental 4.1"

* Combine some aspects from exp. 3.1 and exp. 4.
* Adjusted tear effect on rear tires.
* Adjusted feeling on rear spinning out.
* Faster response to downforce changes.

 

Thanks Karsten I will test it later today and let you know, have notice by yourself the force iwith the Mini on exit pitbox?

 

Is there somewhere that the rFuktor information is kept? Like what the game sliderrs do. Or recomendations for wheel bases etc.
It should be either in the FFB file, or another text file with both then in a zip.... so the information is alwyas there with the FFB file (in it, or with it).

 

I tried/tested the rFuktor V4.1..... hmmm, terrible!
Then tested V3.1....... "very good", but has a few issues - which no one seems able to get everything right, and possibly that is even impossible (due to game stuff).

V4.1 was "terrible" because it gave pretty well zero front, and rear, grip information.
V3.1 gives very good front end grip information, but SOME...poor... rear end information.

To test these.....

Turn off TC.... do a big straight burn out, which will likely kick out the rear end anyway, then MAKE the rear end kick out with just a fractional quick sterring turn. This is so that you don't use the fromt tires at all really and then all feedback (steering wheel FFB pressure) is from the rear end slip. It SHOULD make the steering go quite light - but it doesn't. Though it DOES go some amoutn lighter - just not enough. So this shows up in racing use when you can't really feel the rear end slip/slide almost at all. It needed the severe test case to even show some up.

For the front end.... just race into any corner - longish preferably - too fast, and understeer through it, probably running wide off it. Spa Pouhon is good for this as it has a bitumen run off area still also, so once you run off you can also do some other 'actions' to test some other responses further. You will see that V3.1 gives way better loss of steering force than V4.1 (V4.1 gives any????). And again, when you race you can feel that loss.

But when you race, often there is grip loss (of some degree) at BOTH ends, and then how can you tell which is which? This is why you have to do specific tests to check EACH END on its own.
If a FFB system DID give you good/correct feedback for both ends, then YOU will have to learn how to assign how much was at each end - it can never be 'true', it can only be what YOU assign according to what you have learnt from racing around and have to tie those FFB feelings it to what you SEE in rotation etc, plus what unfolded. Once you learn all that well, you can make use of the FFB drop to guide you on what to do - front and/or rear end motion change (via your inputs).

I have not really checked anything other than the front/rear grip loss, FFB 'drop', as they are pretty wel the number ONE aspect to have in FFB. Though lots of other things are great and also of notable use to have also! In general V3.1 seemed to cover the range pretty good.
WAY better than the Reiza Default, which is really TERRIBLE!!! The Reiza Default+ is tons better (than Default) but rFuktor V3.1 is notably better than Default+.

FullFFB, which I used to use and considered the best in total..... that went to 'crap' on the new V1.6.5 update. Though I should have spent a bit more time with its sliders - but it seemed to lose out on the front/rear end FFB drop almost totally, so that was a very bad sign right away. Making me think that would be 'unfixible' by any sliders really.

rFuktor V4.1 also has that weird 'in pits' oscillation - but V3.1 does not have that (plus did most other things better anyway!)
So V3.1 just needs some 'increased' information about rear end grip loss/reduction. For one thing at least.

 

After a bit more 'real' racing I decided that the rFuktor V3.1 FFB loses TOO MUCH force on FRONT END grip loss/reduction. You are getting too light of a feeling when cornering at an optimum 'slight' slip angle where it is likely that the steering force feeling would not even reduce at all - or at most a small amount.
Meanwhile the Rear End provides almost nothing - no reduction in steering wheel force (slide around, or out of turns, intentionally to produce that resultant - but it doesn't turn up.....)

 

Good information, thank you!

It's surprising, since I did add a few things that should have resultet in more clear loss of ffb on rear end going lose in 4.1, so will have to revisit that, since it seems like it's not doing the job well enough. 4.1 does not lose force as much on hard cornering as 3.1 did, but more than 4 did. Will revisit again.
4.1 has less damping and less LFB, which should give more precise input at the low end. The experiment in this regard was to see if damping set that low simply is too low and should be higher, looking for the "right spot".
(edit: Your 2nd post observations match some of the exact aspects 4.1 tried to fix)

Thanks again!

 

rFuktor:
(rear_grip_loss_feel 1.3) #How much rear grip loss decreases feel of grip [was 0.0]

Seemed to be the parameter I was looking for.
But it hardly made a difference. I THINK it did.... lol
Did some 'skid pan' burnout testing......

 

rFuktor:
(rear_grip_loss_feel 3.3)

Still seemed to do little, or nothing...

I looked at its use:
(grip_l (max 0 (- grip_fl (* rear_grip_loss_feel (- 1 grip_rl)))))
(grip_r (max 0 (- grip_fr (* rear_grip_loss_feel (- 1 grip_rr)))))

Could not make out what their scripting format is or does.....
So then I looked for the use of grip_l, grip_r and they do not seem to ever be used anyway!(???)

I thought, expected, they might be used in:
Final Rack & Effects Blending
But no.....

 

Well spotted!
That is the bug that I just have fixed here.I will post it in a few mins just so you can try those values.
I need to just tune some rear tear first, so it's not crazy bad at the same time here, as I made a few additional adjustments.

 

"rFuktor ? experimental 4.2"

* Fix for rear slide.
* Adjusted damping slightly up towards 3.1 level.
* Adjusted rear tear slightly.
* Changed rear slide calculation.

---

@PeterV If you try to adjust rear slide feel, then these are the ones to look for now:

Added to 4.2, related to the in-slide calculation:
(lat_weight 2)
(long_weight 1)

Those will impact when the rear tear effect is felt, so these 2 then will need adjustments:

(rear_tear_threshold 0.2)
and
(rear_tear_scale 0.50)

 

@Karsten Hvidberg
I've been using your files for months, but never got a chance to create an account and comment on your progress. I appreciate you taking the time and sharing your art with us. Without your dedication and artwork, I wouldn't be playing AMS2 so thank you for all you do. You never stop improving the game experience and that's golden.

Regarding the recent files:
Speed 2 was very muted and the Damper effect was extreme. It caused the car spin on high speed corners. Please note, I've only tried the recent files on my T300 (Setup B) as my DD setup is going through some changes.

100% damper as you advised has never ever worked on a belt system for me on any files. There's that moment the car gives up and starts spinning. Anything before 4.1 had a lose rear and you were always on edge of your seat with tight muscles fighting to save the car.

4.1 has been fantastic so far with zero damper on the Thrustmaster software. I feel way more confident taking corners and I feel like (for some reason) any McLaren absolutely love that 4.1 version.

I haven't gotten the chance to try the file on my DD, but T300 doesn't like too much damper. I just noticed you released 4.2. I'll try that today.

Low speed everything is fine, I believe AMS2 still struggles when it comes to high speed corners. Also, I feel like 4.1 has fixed the spongy... left/right confused movement after coming out of the corner (not 100%, but manageable). That has to do with the physics of the game I believe so there's so much you can do to fix it

My setting for T300 over 4.1:

Everything 100% except Damper and Spring on Thrustmaster software.

In game:
60 to 70% FFB (depending on the car) and zero everything except 20 for road effects.
I feel like I'm much faster on lower FFB setting than 80% that was suggested.

Thanks again for all you do to keep this fantastic title alive!

Sean

 

Karsten - I am about to go for a drive to test that new version.

I was wondering, why does the car 'vibrate' (wheel, FFB) before the engine is even started?
I noted, by turning off the engine after having been sitting there with it started, that the vibration/pattern of it, is identical. So it seems that the vibration is created by the 'engine running' process - but in the game it is not running, weird.
I thought that after I saw that in V4.1, and went to V3.1, that it wasn't threre - or was at least a lot lower - in V3.1. But on re-checking it really was much the same for ALL versions (V3.1, V4.1, V4.2).
Surely it is some 'error' in the FFB construction?