1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Physic discussion thread

Discussion in 'Automobilista 2 - General Discussion' started by Avoletta1977, Jan 3, 2021.

  1. Avoletta1977

    Avoletta1977 Well-Known Member Reiza Backer

    Joined:
    Mar 12, 2016
    Messages:
    569
    Likes Received:
    276
    We have a long topic in beta but a similar one may be interesting for other contributors unable to access the specific forum.
    As the community (hopefully) spreads, I think it would be beneficial for all to stay in the same place (at least for everything it's working in the same way on the main release).
     
    • Like Like x 2
    • Agree Agree x 1
  2. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    @oez

    Everything, you have mentioned is basically discussed and concluded already in the beta section of the forum and it definetly matches with all experiences, i have made (also when comparing turning with AMS1, for example).

    Reduce diff preload to 0 Nm and then compare to just 10Nm and maybe you experience a surprise, how much the impression of cars in AMS2 varies and how close it comes to your assessment.^^
     
    • Like Like x 2
  3. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    The clutch LSdiffs aren't working properly, because the reaction to ramp angle and preload adjustments is showing some serious oddities and unreasonable results (or none). There is enough prove for it already by recorded telemetry and Reiza is aware of it.
    Of course you are faster with higher lock on a clutch LSD, but this isn't the point of the discussion. ;)
     
    • Like Like x 1
  4. oez

    oez Mayor of Long Beach Staff Member Reiza Backer

    Joined:
    Aug 27, 2020
    Messages:
    630
    Likes Received:
    540
    I just wanted to point out for confused readers that this and its surrounding comments is roughly where this diff discussion stems from: Automobilista 2 V1.1.0.0 RELEASED - Now Updated to V1.1.0.1

    I have to try this. Sadly not a globally user applicable fix since many cars have fixed clutch LSD values in the setup. But will surely take what I can for now :).

    I'm glad to see more discussion about this out in the public, because it felt like nobody cared or saw this. Of course that wasn't the case.

    The discussion in public is mostly just valid descriptions of what different players are feeling. Like the "no liftoff oversteer" notion from GamerMuscle. It's annoying when some people - and I want to be clear that I don't think GM ever said this - jump to the conclusion that Madness itself is somehow doomed to fail like this.
     
    • Like Like x 1
  5. Avoletta1977

    Avoletta1977 Well-Known Member Reiza Backer

    Joined:
    Mar 12, 2016
    Messages:
    569
    Likes Received:
    276
    upload_2021-1-3_16-26-31.png
    In the above graph the wheel are perfectly differentiated for the 620R in 4th gear at 180(!) km/h...


    upload_2021-1-3_16-43-49.png
    Also here it seems to work quite fine at bus stop exit in second gear.

    For the rest of the lap, no matter the abuse the speeds where exactly the same (even starting with one wheel on the grass and the other on asphalt).
     
    • Like Like x 1
  6. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    Yes, because often simracers don't really care for this stuff and every car, that makes vroom is immediately "amazing". Or it has to be this way in their eyes, because it's "fast" and "immersive".

    No offense to the guys, enjoying the title. But the arguments, i sometimes see, are giving me a pretty good headscratch.
    Even GM (he is not the usual techie guy) noticed, that changing coast ramp angles did basically nothing. :D


    @Avoletta1977 Yes, Bus Stop shows some differentiation, for many cars, that's all. (I now whait for the argument "but it has to stay locked the rest of the time, because it's faster".)
    Let's running circles again. :D (Btw. this bit of differentiation looks still pretty "reluctant")
     
  7. Avoletta1977

    Avoletta1977 Well-Known Member Reiza Backer

    Joined:
    Mar 12, 2016
    Messages:
    569
    Likes Received:
    276
    Umh, apart from bus stop I had a possible reason why it's working in that way, but the slippage there ruined my theory... :D
     
    • Informative Informative x 1
  8. Maser V6

    Maser V6 Assume nothing. Verify everything. Have Fun Reiza Backer

    Joined:
    Mar 27, 2020
    Messages:
    2,126
    Likes Received:
    1,145
    Hi @oez
    As @CrimsonEminence and @Avoletta1977 have indicated the Diff discussion goes back in Beta as well as main
    not sure if you have seen any but here is just 1 link Formula-Retro drivability
     
    Last edited: Jan 3, 2021
  9. Raceracerace

    Raceracerace Active Member

    Joined:
    Jan 2, 2021
    Messages:
    336
    Likes Received:
    70
    So whats the go? What needs to be done?

    Seem to remember patch notes about this in a dev diary thing. Seemed it was put on the backburner for a time. A choice was made etc. to get perhaps repeatable results, consistency rather than nuance.

    Madness engine been around a long time, many iterations, its something that can probably be seen to now cars tracks, tires is well advanced along
     
  10. Maser V6

    Maser V6 Assume nothing. Verify everything. Have Fun Reiza Backer

    Joined:
    Mar 27, 2020
    Messages:
    2,126
    Likes Received:
    1,145
    Its already been suggested by above but elswhere
    Devs are aware and I for 1 remain optimistic and confident that the "further polishing" and constructive feedback will yield results but like above its a little frustration time wise. Brabham bt44 in particular for me, as a favorite car. However the Gt1 merc and procar will keep me ocuupied for a while.
     
  11. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    The best at status quo right now would be, to implement another (maybe simpler) method for clutch LSDs, if the culprit doesn't show itself. The open diff also behaved like a spool before a major change and cars with open diffs are now totally different, than before this fix. (they are way more believeable and enjoyable to drive)

    This issue is a dangerous opportunity to mask other oddities with cars, that possibly receive fixes, that lead to not fixing the true origin and can result in further problems after it. (It already starts with reports about bad turn in and snappy behaviour in slow corners)
    When the clutch LSD then gets fixed eventually, cars could behave strange again after it. That's why it should've a very high priority, to tackle this issue ASAP, to not further play shenanigans with content in the game and it could even improve the polish-work of all physics dependent things in general.
     
    • Agree Agree x 4
  12. Raceracerace

    Raceracerace Active Member

    Joined:
    Jan 2, 2021
    Messages:
    336
    Likes Received:
    70
    Thanks for the post. Just so happens amazingly I wrote this at the same time in the other thread - that it sounded like at the time it was a deliberate work flow thing which now you mention that would (we'd have to go back and read) accomplished a few things, so they went with that, in order to like you say hit on other things.

    My quote:

    "The madness engine has seen many iterations. They will get to it. BMW challenge on madness engine probably had super naff differentials haha

    I spent a good deal of money on this setup - almost 2 thousand dollars - and I do not feel jipped by this game or the others like Rfactor or ACC. So I totally understand you when you say about differential and being close to the actual thing etc. I often adjust the dif settings, depending.

    I am pretty sure they will get to it. What I am seeing is a workflow thing, not a 'brain drain' thing. The differential needed to be a repeatable results thing - the guy fronting the project said that basically (iirc) a few months back.

    This helps them out the diff thing, its allowed the game to do other things in that time. They will get to it."

    __

    The old formula problem/solution - new numbers, new results. Would appear to be easier than ever maybe to implement this. Though the implementation is 4/5ths the cure in this.

    Oh well. The more one drives the more they come across these things. Guess it will be ironed out.

    I am not a mechanical engineer so I can't help with that - but what I do know is that now most of the cars are in the game, from a project standpoint, now they give consistent results and are complete - they can probably change something as crucial/central as this going forward.
     
    Last edited: Jan 3, 2021
  13. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    This is the thing:
    The issue is likely holding back in "workflow" dependent manner. Because it can influence finding solutions for non-diff dependent stuff, that is masked behind the problem.

    We are talking about it since months now, like mentioned. They will hopefully tackle it.

    I personally can't agree, that AMS2 is always particularly enjoyable/believeable on the physics side (which makes the fun for me personally). It's one of the things, that makes me always go back to AMS1 and other titles.
     
    • Agree Agree x 2
  14. Raceracerace

    Raceracerace Active Member

    Joined:
    Jan 2, 2021
    Messages:
    336
    Likes Received:
    70
    Played AMS1 yesterday again.

    Indeed, as you say and some cars more than others; and yes this is a sticking point that will make myself turn away too. More real cars is definitely a notion that is easy to grasp and would make myself like the title even more. Probably most people. The recreation of that certainly is the draw of the title. At the start of the title I could not believe (in a good way) the way Bathurst had been recreated - that kind of excellence.

    I might say the analogy for driving believability can break down, or in general I often say where does the analogy break down (like in modern physics from the huge [planets etc] to the very small - atoms - the analogy or the theory breaks down.

    Thus this could well be a case for less complexity or more. Maybe a simpler model for those without it, giving two models.

    So you guys certainly have it pinpointed. Do you think they have a second computer where an early model of such a thing is being worked on outside the main branch of code?
     
  15. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    Like we saw in a Development Blog in the last months, they do definety research on games foundation based areas, so who knows. It's likely, that a dev has a development build for deeper research, would be strange if not. :D
     
  16. Raceracerace

    Raceracerace Active Member

    Joined:
    Jan 2, 2021
    Messages:
    336
    Likes Received:
    70
    This was particularly noticeable round julyish - and was frustrating to be honest.

    I gave everything 85 ramp and yeah, those days are gone now thankfully - off throttle turning seemed to not make a difference, and thus people myself included would throw things into the turns; probably much like the comment a few above from someone.
     
  17. Avoletta1977

    Avoletta1977 Well-Known Member Reiza Backer

    Joined:
    Mar 12, 2016
    Messages:
    569
    Likes Received:
    276
    Ginetta GT4 in a kart track.

    130 Nm Preload 89-89 ramps 2 clutches
    upload_2021-1-3_17-38-9.png

    30Nm preload 89-89 ramps 2 clutches
    upload_2021-1-3_17-47-1.png

    Cat 620R default setup

    upload_2021-1-3_18-4-2.png

    So, in some cars the Clutch LSD quite works apart from the extreme influence of preload.
    In some don't... like at all.

    Driving the cars on such a narrow track is quite revealing anyway (the 620R is simply terrible, while the Ginetta is nice... even the lift off oversteer is working when opening the diff in coasting and using low preload).
     

    Attached Files:

    • Informative Informative x 1
  18. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    They aren't totally, unfortunately. There are many cars, still don't showing enough changes. It is definetly showing something, it still needs so extreme changes on the ramps (and taking away the preload spring at all, basically) to receive something of a truly noticeable difference.(good or bad doesn't matter)
     
  19. Raceracerace

    Raceracerace Active Member

    Joined:
    Jan 2, 2021
    Messages:
    336
    Likes Received:
    70
    ok well that would be because I do not drive all the cars to such an extent, and am particular about which cars I drive on what tracks.

    eg strangely I like rfactor 2's 650, thus I do not drive it in AMS2.

    I consider the gen 1 classic f1 to be in such a category and the more powerful caretehams. Unwieldly.
     
  20. CrimsonEminence

    CrimsonEminence Administrator Staff Member Reiza Backer

    Joined:
    Nov 4, 2018
    Messages:
    6,847
    Likes Received:
    4,463
    You could also test it with non-89° ramp angles, at this point for a reasonable picture in "practical" use. I also think that 130 Nm of preload shouldn't cause the diff to lock most of the time on a tight track already. The car is quite torquey to begin with...i just would expect...more.
     

Share This Page