Physics discussion thread

IMO the rear of the car is completely predictable and behaves as it should. I've never lost near without knowing what caused it.

The front end in other hand.. It has all the grip in the world and i can just steer the front however i wish regardless of inertia. Then i just keep the car in very controllable and unrealistic front end slide thru the entire corner. This enables very unrealistic rotation so overshooting the corner is impossible.

It feels very rewarding tbh but completely wrong. Atleast i can feed my ego with laptimes which would be alien level in ACC.

This applies mostly to the gt cars but in some extend to all classes, expect some classic low df formula cars.
They feel very nice with a loose rear and front end on rails.
The front maybe has unrealistic amount of grip even in these classes and you can really dance the car on the nose.


Overall i think AMS2 physics are excellent and ahead of many competitors. Only those issues around rotation and front end grip’s disrespect to laws of inertia botches the ensemble.

If Reiza can fix this, i think AMS2 become all time greatest sim to the date.

 

Partially agree, I just don't think fronts "have all the grip the world" on when braking / corner entry, you can get plenty of understeer there if you brake too late/hard (and don't got the trail braking just right) and I definitely feel the inertia in that phase, but mid corner onwards you tend to have a lot of front grip and enough rotation even when on throttle (rotation from rear also adds to that).

I find the braking phase very good in AMS2 for GT cars too, even though the default setups often tend to unload the rear a bit too much and too long in some scenarios (very track/corner/car specific though) and won't allow trailbraking that deep, but a lot of that can be mitigated with setup changes if needed.

 

I love this comment !

 

Solid comment.

I maybe exaggerated a little. Yes i agree you can overshoot and miss the corner entry if you brake too late. And i don't have issues with physics in that regard. The thing is - if you're even half competent - you can brake very late, trailbrake in to corner entry and suddenly you're in this Need For Speed/Mario Kart slide, where the front has "all the grip in the world". So there starts that insane mid corner rotation where we initially agreed.

Taking Flugplatz flat out with cold tyres in BMW M Hybrid V8 feels nice, but not realistic at all.

 

It seems that you were totally right with your comments - since the last update the sudden oversteer in GT3 class has gone (driving Double Gauche at Spa or the frist right hander at Barcelona for example).
I just wanted to say thank you because it's people like you who make devs aware of such issues. I was wrong about the setup thing. You could "conceal" the issue with setup changes but it was - indeed - an underlying issue with AMS 2's physics. You were absolutely right.

 

This discussion had nothing to do with FFB as I also pointed out. It was mainly about the floating feeling or boat feeling

 

I have a question. Firstly I am very glad that this yaw setting, which I am sure few of us had even heard of or thought of before we learnt of this new change, has been altered in a way that makes the cars better - stuff like this helps me even to understand physics in general.

What I am curious about is - what is the reason that the "real" effect of air can't be found or calculated? Why is it possible to get this "wrong" in the first place? Is it due to a necessarily simplified aero model in the game? Or perhaps it's just not really possible to actually calculate easily? I understand that a very large amount of the effects of the physics model (in the Madness engine at least) are essentially emergent - but we know the density of air, the resistance of a wing at a given angle, etc, so why isn't the "correct" yaw effect just emergent as well?

 

Aerodynamics is a very complex system that is dependent on a lot more parameters than air density and the resistance of a wing. If we were able to simulate aerodynamics absolutely accurately, F1 teams wouldn't need a wind tunnel for testing. They could just simulate everything. Even the best CFD tools can't simulate aerodynamics to 100% accuracy. And these tools require a massive amount of computational power and can't simulate in real time. Due to the these computational limitations, models like the one's used in racing sims are very simplified.
Let's break down a little bit why aerodynamics is so challenging to simulate correctly: The downforce produced by a rear wing is not just dependent on the aera and the AOA. The body of the car accelerates air which creates high and low pressure areas around the car. These highly affect aerodynamic components. Also the downforce generated by an aero component doesn't increase linearly with the AOA. Once you get past the critical AOA, the wing stalls and you lose downforce massively. These things are still possible to simulate (though the computational power needed is probably too much for a racing sim already) but when it comes to turbulences, that's where simulations really reach their limit. Turbulences are a common phenomenon in race cars where air interacts chaotically with wings, underbodies, and tires, creating vortices, boundary layers, and wake effects. Aerodynamics involves solving the Navier-Stokes equations, which describe how fluids like air move around objects. And these equations remain mathematically unsolved for turbulent flows.

In conclusion, aerodynamics just bring a massive amount of complexity and the models used in sim racing will always be greatly simplified.

 

In physics, assume every cow is a sphere until it isn't.

 

Limited real data, limited models.

 

Unfortunately, an old "bug" has resurfaced.I'm going to post it here again, because it's completely irrelevant whether it's in the default settings or a custom file.The in-game damping isn't working properly in my opinion.With these new updates, it's gotten worse again.What I mean by that is that the damping increases the vibration behavior.The wheel rocks more when the damping is increased.Additionally, when the damping is increased, a constant grinding noise can be felt in the steering axis, comparable to sand in a gearbox.This is really annoying, because it was much better before the update.I'm sorry if some people don't like it when I write things like this.Ultimately, I only want the best for Ams2, and this isn't meant to be condescending criticism, but rather constructive criticism.

 

Thanks

If it helps, I recommended a little damping again on June 18th because I noticed that everything was back to normal with the in-game damping. Unfortunately, I can't tell you if this change came with the latest updates or earlier.

In this context, I'd also like to point out that something isn't working correctly with the in-game LFB value.Yes, I know, the Low Force Boost is only relevant for lower wheelbases, but many people still have those wheelbases.It's also irrelevant whether it concerns the default or a custom setup.

LFB should only work in the lower force range to increase the subtle forces and to maintain a firmer center and reduce the dead zone on weak bases. Unfortunately, increasing LFB increases vibrations, even at high speeds (also visible in the FFB graph). It's a wavy curve from left to right that increases with increasing speed and can be felt at the steering axis. Additionally, the grinding sensation is also increased. I don't think it has anything to do with the damping system, because it should work independently of it (but that's just my guess based on my testing). Perhaps Reiza should look into this as well.
Thanks

 

I think sims can (or already do) use multidimensional LUTs to get pre-calculated data. That alleviates real time calculations a lot.

I also noticed that a few weeks ago, but now I’m using damping_slider to change other parameter, and that side effect seems to be gone on my system.

About tyre grip and float feeling
Yesterday, I ran an experiment: I took an Alpine A424 and went to Barcelona.
What I tried: front ARB at max, rear ARB at min (objective: reduce front grip and increase rear grip).
What I got: a car with behavior that I consider correct. For instance, I can still spin (loose the rear of) the car at Turn 4 with TC=6, but it keeps its composure on the rest of the track much more easily (I don't have to constantly fight the car).

I also looked at cockpit view in LMU (Alpine A424, Fuji) and noticed how the pitch view is much more downward compared to AMS2 (for reference, I feel AMS2 is more correct, more horizontal, but stay with...).
What I tried: replicating the same view in AMS2.
What I got: a much lower sense of floatiness.

In the past, I had noticed that moving the seat forward seemed to reduce the floatiness effect. Now I think I know why: the more of the track that is visible closer to the car, the more the floatiness is reduced. What remained was just the normal car sliding from overdriving, but now it feels natural to me.
Again, the conclusion is that this is a 3D world rendering ‘problem,’ not really related to car physics. It’s as if the rotation pivot for the car and the rest of the 3D world are not the same, or as if the car is rotating around its center of mass instead of around the wheel with more grip (or the resultant forces).

Could it also be that my rig setup favors this view?
I don’t take these findings as hard evidence - just personal notes to help understand the root causes of the most criticized aspect of AMS2.

LMU view


AMS2 (changed view)


AMS2 (my original view)

 

I have a different opinion.I don't see any problems with the 3D world or 3D modeling.Sure, you'll have a different view if you move your seat further forward.Your field of vision will shrink, and your peripheral perception will diminish, but the floating or boating feeling will remain. I can recommend another test.Go to the track and switch your camera view.Go to the follow camera, i.e., behind the car, or use the camera view in front of the car.Now, in these views, drive with the steering sensitivity set to 100% and then with 1% steering sensitivity.The difference is like night and day. Perhaps now you'll understand what I mean by less slipping and less boating.These are extreme values, of course, but to get a better feel, I recommend using these values.I also believe it can be improved with car setups or with better camera movements.But the bigger effect I can achieve with FFB (masking) or changing and reducing the steering sensitivity

 

I can't remember personally getting this "rotating from center" feeling in AMS2 personally. I drive on both VR but also on ultrawide monitor where I don't go for mathematically correct FOV, which to me feels and looks wrong, so I have it always clearly higher and/or try to get the monitor as close to my eyes as possible to avoid using some really mathematically correct low FOV values.

BUT I do remember rotation feeling "different" and "rotating from center" on the LMP2 Oreca 07 on rF2 that I have used there a lot (it's been a while) and always only on monitor, but it didn't ever really bother me and always thought it more of a visual thing than anything else.

If it's a more of a visual problem for some, it's probably less of a issue if the car is understeery .. but in AMS2, you tend to have cars that rotate well, sometimes by flipping the rear around, so you tend to notice it even more than in sims where cars tend to be more understeery in nature.

ps. never heard from VR users that they would experience this "pivoting from center" feeling

Honestly, I think you two are talking about two different things that people are bothered with. For some it's the "pivoting from center" issue, for some it's some other floatiness and for some it's both

Also clearly for some it's been the car behavior in certain situations, because suddenly with rear snapping out in certain situations (yaw sensitivity change for higher downforce cars), a lot of people get a lot improved feeling (I think you can also get some more understeer tendencies with this yaw sensitivity change).

 

Anyway, I'm going to leave this thread now that I've shared my feelings and feedback. Maybe it helped, maybe not

 

sorry but did the gt3 cars seems very slippery when driving?

 

Warm up your tires between 80 and 90 degrees.