Automobilista 2 Custom Force Feedback - Overview & Recommendations

hello Karsten.

I am starting to do some back to back comparison between Reiza default and your variant of the FFB by curiosity.

I posted in the Stock FFB thread but didn't get traction, are LFB/FX tweaking influencing the same thing in the Reiza current definition compared to yours? Do we know?

thanks so much

 

Hi Nolive, I hope all is well with you!

It's almost the same. I've had Silver Raw use a different curve for a while & that worked well at low-ish lfb values, but not so well when going higher, since it then hurt the high end force details too much, so I went back to more normal LFB recently. Also it was hard for people to switch between Default and custom when there was this difference.

Now the only difference is that in Silver Raw v47 it is using "dual lfb", one LFB per front tyre. This gives a more distinct feeling of each front tyre. The strength of the LFB should be appr. the same.

Best,

-K

 

I’m probably sitting in the back of the class by myself, but I’m not finding .47 to my liking . ive tried and tried to find settings in game and in file that gives me a steady and consistent feel that feels as good as .45.

it feels strange or vague around centre at times while feeling tight others , and has some unnatural feelings mid corner and exit , that don’t seem to match what I’m expecting or seeing. , like it’s either delayed or not reduced fast enough. im not getting the sense of relative grip that I do in .45
Mid corner adjustments don’t correlate as well and aren’t as smooth and quick.

I just feel like there’s to many force’s acting and they aren’t in harmony. .

Since .46 it’s been much harder to switch between cars using the same settings in game, with .45 I can pretty much set and forget gain and lfb , The smoothness and speed of the wheel is consistent, and the different physics of each car can be felt without the doubt of if I’m optimised in terms of FFb.

I just can’t find the balance of usability and sensory information with .47 unfortunately.

 

yes doing well.
weekend time to test FFB you know will try your latest PC2 one in my evening starting soon over here

perfect explanation as usual, I am clear now thanks again for the quick reply

 

It seems likely this is an effect of running high LFB values.
LFB exaggerates the low forces, which is normally just fine. If you run very high LFB it still can be fine, depending on what you like.
Maybe try and run higher gain & lower the LFB and see if that helps.

(The dual LFB probably further exaggerates this as well, I will add a fix for that part in v48. It still cannot fix the overall issue of running very high LFB, so if you are sensitive to the skewing that brings with it you have to go the higher gain/lower lfb route.)

 

I feel like v48 wip lets me reduce LFB MUCH more and also feels cleaner and generally better as far as torque balance goes so 48 should be a good one!

 

Apologies did not really have time this week to turn a wheel so only a quick feedback session


Ver 47 - only had 5 minutes and as a tantalizing comment, I would suggest that @mikk-schleifer and @Michael-enright have a go as I am seeing convergence with 31. The lightness on cornering and the vague feeling around the centre are now gone. I upped the gain slightly and fx but got to say its the best after 31 I have felt. Need more track time but the fact I didn't delete straight off is going in the right direction for me

just noticed the suspension comment so maybe this will cure the need to up fx?

@panos-schoino many thanks for taking the time to create a bespoke file, although i do have to report it wasn't as good as 31 and after trying 47 i am not sure feedback unless you have a specific request is worthwhile.

Combo - as i am focussing on a car that i am driving in a league in the interest of fairness and to provide wider feedback what combo @karsten-hvidberg what combo would you like me to test

If people want my current settings they are here Fanatec DD2 Settings - Share yours

if I get chance I aim to use both 31 and 47 in a two race league event tomorrow for more comparison so will let you know how i get one

thanks again for all your efforts and dedication to the ffb cause

 

No worries mate, it was an interim attempt until 0.47 was ready to be posted. I am happy that you find 0.47 good so far. I.r.o the combo i am not sure if Karsten has a preference, i would say GT1s, GT3s should be ok. Also, for .47 please try various settings for lfb,damping and fx. Any values that feel right to you should be fair game. For reference, with my SC2 pro i am using gain mostly in the area of 30-40, lfb 5-20, damping 5-20 and fx 50-60.

 

i’ve been trying all sorts.
settings similar to panos don’t feel too bad and generally sort out the corner exit weirdness.
but i’m finding turn in a bit lazy and lacking nuance and accuracy.
and in mid corner on fast bends i’m getting this odd knocking feeling similar to scrub but slower , it doesn’t seem to reduce if i turn scrub scale down.
i’m not sure if it’s to do with hotfix, but i’m significantly slower with this set up .

 

I think I've read multiple people being slower bc of recent game physics changes, so might not be related to changes in the ffb file or your ffb settings.
The feeling you get in corners is actually a new effect mostly present on cornering. I can tune it down for v48.

 

Spent a bit of time today comparing 47 with 31 (as 31 seems popular with a few users)and default.
31 feels totally connected to the track at all times which is good, although lacks in variation between high/low forces even with low lfb.
47 feels generally good, but goes vague at times exiting some corners and in the compression at brands, where i'd imagine it should really load up.

I still much prefer these files over default, but the one area the default excels at is the variation of forces especially in corners- the ffb really loads up nicely when accelerating through and out of them.

Could be completely wrong but maybe some settings in these files are working against one and another (acceleration?), cancelling out certain forces that don't affect dd wheels.
Any of this make sense?

 

Thanks Alegunner!

The load up should be better in v48 again. It is the limit on high downforce load that is too strict in earlier versions.
I am not 100% sure about the the going light coming out of corner issue, what Lfb are you running at?
Is there a good car/corner combo that showcases the issue well?

If you want more load in corners you can set power_steering down to 0.0, btw and raise the gain.

 

would be good if the new scales in file like power steering , front slide and lat edge had a description next to them in .48
can’t tell if they need adjusting or not .

 

Hi Karsten, it's during the exit as I start to straighten the wheel the ffb seems to drop off too quickly, almost the same feeling as going over the brow of a hill or locking the fronts.
Running low lfb to get as much force variation as I can, tried at Brands gp in the f309 and porsche gt3.
Seems to happen most in the fast mid field corners.
If you can, try 31 and 47 back to back and you may see what I'm trying to explain.

 

I'm still rocking 38 or 39 cant remember which now ,,slightly tweaked the engine vibe & road detail and its the best ffb i've had yet ,,,obviously some cars excel more than others due to physics but i couldn't be happier ,,,,

Hey Karsten have you thought about putting a small section on the first page explaining how to tweak maybe 2 or 3 key features in the ini file for folks who are not quite getting the feeling they were hoping for , just so some folks who are struggling can grab one of your files and have a go at adjust it more to their ffb taste as i it must be hard trying to please all the folks with different wheels and different opinions of what the perfect ffb is .. just a thought.

I don't think i'd be playing this game half as much if it wasn't for your custom file,,,, keep up the excellent work dude .

 

thats probably the acceleration effect to lighten the ffb slightly when tyres unload, it was to simulate unloading during hard acceleration, especially rear engine cars. not sure what wheel base you have but on DD1 the value in 47 was fairly low already, you can reduce or turn off with "Acel2" in the file.

 

Cheers Shadak, I suspected it may have something to do with acceleration, I'll try what you suggested.
On CSW2.5 btw.

 

Sorry it doesnt show signature on mobile page,
Had a look at the file, its this one, try it at 0.0
acc2_scale 1.0

 

That's the effect i mentioned previously that has got better for me at least in ver 47

 

Last week I did a test race at Simracingonline in AMS2 and a general remark is that FFB was not good. I already heard about and started reading parts of this forum and tried some of the silver-raw files, to see what the effect would be on my Thrustmaster T150 entry level racing wheel. I found that v45 wiht GT1 Porsche and Kyalami new version had good feedback on curves but feedback on especially on oversteer in a slip state to catch the car was not quite there yet. I also tried same settings on a few formula retro cars Brabham BT44 on 1975 silverstone layout but I do not get any sense of feedback in that combination.

I also had a look at the setings file and some questions / ideas came to my mind with my background as electronics design with backgrounds in software and control systems.

I had read somewhere that there is no documentation about the way the files work, but I have a few questions about the content of the file:
1. What is the program language used or how are the functions calculated
2. I assume that there is a standard sample rate that the new output is calculated, is the sample rate known?
3. If this is treated as software and calculated can you do stuff like S2=S1; S1=S0; S0=Input.

If so it is possible to make discrete filters where you can make a sort of equalizer where you boost or supress frequencies to improve the feel using IIR filters. Perhaps there is already software available that can inject noise or swept sines and determine the transfer response of steering wheels which can be used determine a transfer function which can compensate for the e.g. lack of feedback over curbs.

I did not have time to dive into the "code", but I assume there are variables which are input from car model which are calculated in this file before being output to the ffb of the wheel or something like that. What are these values in the code.

I have no exact knowledge of all the physics used in the custom ffb file but reading this thread I notice that some people have excellent results and some have trouble finding correct feedback behaviour...

Maybe someone with a lot of inside information could provide me some information to see if my ideas are feasible to improve the feedback in the wheel or that the way this file works with the reiza code is completely different than I hope it does. It would be highly appreciated to get feedback or some people in this thread to elaborate on this idea.

Kind regards Gert