Automobilista 2 Custom Force Feedback - Overview & Recommendations

I've been using R3E since it was free in SIM FEST a few days ago, them I bought it!
The FFB with my TM TS-PC (with ACELITH formula wheel) is superb, on par with rF2 (if not better).

Well, that leaved AMS2 a little behind... until I also tested @Danielkart latest release.
I do like it a lot, much better than my tweaked version of 4W variants.
I changed a few constants to tame down some FX, because I believe the result is different in ANY other wheel. Let alone subjectivity - we tend to adhere to information and sensations that already make sense to us. That's a (normal) cognitive bias (confirmation bias), no problem whatsoever. Our brains needs to train themselves to identify different stimulus as normal - it's called learning.

I'll continue to test but, for now, AMS2 is not behind, per si, just different.
And AMS2 has an advantage... FFB can be whatever we like! Anytime!

 

-

 

As usual, I attach two Atmos files (with and without YAW) ... and as usual the last original Karsten's file v.5.0.1.4 ....

Name: rFuktor 5.0.1.4 Original Karsten Files

Name: MOD ATMOS 4.1 V.1.3.3 FIX by Stakanov (with & without "Yaw") and based on 5.0.1.4

Note 1: The file ATMOS is the same of the original 1.3.3, with same harmony and same value, but now it is fixed from crash when you load the game and going to Bathurst with Super V8 car. I hope!!

Note 2: Being identical to the previous one, if you have changed some values, you can safely overwrite them without problems.

Note 3: We have found or rather you have found the solution to the steering shake when the car is starting on the track and you leave the clutch by accelerating to the maximum, here on the low end it is not a huge problem (then for now a leave the original value) but for those who want to avoid this it is enough to lower it a lot this voice (the advise is in alternative value):
(slide_factor 0.85) ----> set to 0.50 or down 0.45-0.40 to taste.
By changing down to around 0.50 to 0.45, shake is just right and completely disappears when set lower then 0.45-0.40. If you change this i think that is necessary retune some items because the attitude of the wheels/tires changes a lot in relation to slip and speed, I think that in this item (under indicate) you could start not from values in the file but from Karsten's standard values and raise/reduce them as you like or at yor taste:

(velocity_res_scale 0.40)
(velocity_res_offset 0.60)
(vel_rot_scale 1.00)
(vel_rot_limit_coef 1.00)

This will probably be tuned next time in the v 1.3.4 or 1.4 but nothing is stopping you from doing it to your taste however you like!!!

I will thanks all of you who have directly or indirectly helped me on the forum (both for the shake, for the fix and for others), I don't quote all of you so as not to forget anyone and for your privacy ... thanks again to all.
I hope you like and if you have any advice or doubts you are welcome, if you dont like use direct other I live peacefully, Instead for everything else I remain silent, really disgusting (personal taste of course)!!!

 

Thx for the file, for the fix (I did only one test and go ok ) and for all other @Stakanov, in the end there was a very specific cause for the crash??

 

It was a pleasure to help you test and it was a pleasure to talk to you (available to everything) I have not had or have problems saying/doing it publicly but I appreciate your silence and kindness ... thanks for the file, I will only replace the values that I had changed before with those of now, right??

 

Sorry for long text.... and for my english
No, there was not a specific error in the code of the 1.3.3 file and if there had been this should not have happened ... from how the defect manifested itself I am convinced that the game in the Bursthust-SuperV8 combination has a different behavior in loading of the game itself and in loading the FFB file, among other things I would tell Reiza to investigate, not for my mod (not important), but because it should not happen that the presence of a custom file in the documents folder sends the game in error and above all even if you use the Default/Default+ profile shouldn't happen that an FFB file crashes the game completely.
The use of the file should be tight and only affect the outpout sensations on the wheel and be read only on the data acquisition side (not change the game performance), if the file is bad, the FFB should either not work (flat output) or at most work out of order (inconsistent output) ... in this case the problem has traced back and invalidated both, the game (with crash when use combination car/track) and the use of Default/Default+ profiles (then again the game fail) symptom of something wrong.

Trying to keep it short, I isolated the things that are different in my MOD than the other files and Karsten's original, which are:
- The length of the code (everything is almost double);
- Some lines remodeled and not the same as Karsten's original;
- Blank spaces and added lines for info;
Then I analyzed the fact that the problem occurred only using the following files v 1.2 and 1.3.-1.3.1-1.3.2-1.3.3 and I then identified the 1.1 file as fully functional and then I identified the different code (curb and suspension) and remodeled it in 1.3.3 using the same shape as the original and 1.1 (without the precious indications of the users I would have looked for serch defects for at least a week ... many thanks).

Then I noticed/modify that:
- code of curb suspension modify with form of 1.1 otherwise the game crashed again (only in that combination of Car/Track) ;
- I don't know why but as soon as I added the smooth in the variables FL.FR.RL.RR_start_load in the suspensions the game crashed again (only in that combination of Car/Track) and then I eliminated the smooth;
- I don't know why but the bumps, the details, the kerbs and all the effects should NOT have been divided between a front code and a rear code (and then unified) but be accounted for directly in a single entity otherwise the game crashed again (only in that combination of Car/Track);
- Finally, I don't know why but I deleted all the spaces and the unused code for reduce the instruction and the lines.

Luckily, the game did not show more rubbish, but I repeat I have not made any corrections only reduced the amount of code and the aggregative form.

Can the shape and quantity of lines of code alter the stability of the game? No and in fact the defect occurred only in that combination of Car/Track ...

 

Yes and Thx for the tests!!!

 

Just a note. For me, you Atmos 1.2 did not crash for me, only 1.3 and up. I'll try your latest and see how that goes. Like you said, it's pretty bad that a file, just sitting in a folder, can crash a game, whether it is actually being used or not. Thanks for sorting it out.

 

Hi Daniel. I finally got around to trying your v36, and don't mind it. But, I still get the constant vibrations. I have turned micro_bumps_scale down, but still does it. Any suggestions please? Thank you.

 

Maybe I'm a lucky guy or maybe even the luckiest one but I haven't get any CTD of any kind since about one year ago.
Tried Single race+Test Day/SuperV8/Bathurst using Stakanov's 1.3.3

So I'm not tempted to upgrade to 1.3.3fix am I right?

 

yes Michael i now, but with another user the logic and code of curbs fron 1.2 cause CTD not in the Load of race (that is ok) but during exit or diuring restart race (same car/track).
Then i go with 1.1 but the file is the same of 1.3.3 only this little modify are from this file (2 Lines of code) … you can swap values from your personal old 1.3.3 without problem … thx again for all …

 

Good to now, you are not the only, many user are ok and whithout problem, but the file is the same, you can simple swap your value from old 1.3.3 in this new 1.3.3Fix without difference and we have another test that all is ok … and if you have problem replace the old!!!

 

But WHY????? Some systems crash some don't. This is indeed a strange one. Maybe to do with system specs. I don't know. I notice that the two of you have Intel CPU's and NVidia GPU's. I have an AMD Ryzen CPU and an NVidia GPU. Who knows.

 

Also I have never crashed the game with all Atmos and Others, I have the same steering wheel as Stakanov and I also use an nVidia video card, I think that only Reiza will be able to answer this question because, barring new discoveries (for example, we must not exceed the 1000 lines of code), the problem could be anywhere from game to file, but with no guidance and or rules on motivation I don't think we'll ever know anything and we were lucky to narrow down the problem.

Perhaps in a future update we will read in the changelog the description "improved performance in the presence of large FFB files", or "resolved a conflict that caused the game to crash in the presence of a custom file in the documents folder", or "corrected the loading of the game in the presence of some combinations of cars / tracks "... or maybe the words "all custom files will have to insert the following line of code inside them" .... we will see.

For now the important thing is to have solved the emergency and to have been able to avoid regenerating the problem and this is above all thanks to all of you who have collaborated. Thx

 

Can I ask what CPU you have?

 

Sure Intel i7 10700k

 

So far, those that have Intel CPU's do not have the issue, but I do, and have an AMD Ryzen CPU, We all have NVidia GPU's. So, who knows? A possibility.

 

Hey guys maybe we have some news, I am waiting for the tests of some users to be sure and if confirmed we were able to get the exact 1.3.3 OLD file back without crashing the game in the SuperV8/Burstust combination (without change of curb, suspensions and others) only different form and with smooth of start_load ... if no problem I will only replace the files in the post above!!!

 

Hi @Stakanov your latest file feels great on my csw 2.5, the load on corners feels especially nice which I think has been missing for a while.
I've made a couple of changes to add load while braking and reduce being thrown across the road on some high kerbs.
Thank you.

 

Interesting I am going to try some values .... I send you the file I am testing privately, it is identical to that of the post above only a different aggregation of bumps and details in addition to the return of the previous suspension without the crush in the combination SuperV8/Burstust ... let me know it should be more consistent but little stuff