Physic discussion thread

Curiosity: did you try and play with the "look towards the apex" settings of the camera?

 

Will try once again (I like to have the "car attached" to my seat position which is... well, static )
Will also try again what I felt was a more correct feeling - to put the seat position way backwards (of course, with other incovenients).
Is almost like the car rotation pivot is in driver's head and not in the Center of Mass (or the CoG).
Our brain has enough plasticity to make it "real", but when compared, AMS2 still feels a little odd - and I don't think is wrong physics, just wrong/offset visual movement.

 

I found that related to how the "look towards the apex" setting was. I ended up zeroing it up

 

for me, having played RF2(lot in recent weeks) and RRE(bit since the All free content popped up last week), what I am struggling with in AMS2 is a better sense of car weight.
overall it feels still too much light and floaty compared to the above mentioned Sims where vehicles feel really planted on the track and also transmit better this sense of weight transfer through the FFB

 

Well, I just adjusted "SPEED SENSITIVITY" to 50 (was 0) and some lack of precision is long gone (I must have set it to zero to have, well... no speed effect! - perhaps values should be more self explanatory or well documented like car setup settings). Generally, I guess, for certain sliders "50" means "0" and less imposes an exponential curve and more imposes a logarithm one.

Yesterday I spent all day tuning my own custom FFB based on "rFuktor 5.0.X.X - By Karsten with Shadak & Panos" version and got it next to what I want - which is something like rF2/RRE. Those are a reference but AMS2 has the potential to adjust to everyone's taste!!
I'm using Thrustmaster TS-PC with ACELITH formula wheel.

Those changes made screen oscillations be felt on the wheel and hence less floaty/light feeling. The Default+ is good but needs a lot more road/suspension oscillations passed on to the wheel.

 

Another thing about light/floaty feeling - I believe (and it just that) some came from the "display damping internal settings" - these competition cars oscillate a lot more that my road car, which is not what we expect, right? The car should shake under vertical road forces, but should land hard and dry, not bouncing around. RRE and rF2 do this better and AMS2 will also do, I have faith in Reiza .
Even if AMS2's physics are correct, our eyes have a stablization subsystem governed by our internal ear and body "sensors" - we do not experience car oscillations like that.
Try read a phone in someone's else hand with and without touching his/her hand - our eyes/body immediatly compensates for the hand motion and reading becames much easier.

Reiza should add that feeling, considering some found the current model to be good.

 

Thanks Joaquim for sharing your views.

As we have the same wheel base (TSPC Racer owner here), do you mind posting your Custom FFB file and also both TM control panel and INGAME settings?

What cars&tracks combo you are using for your testing would also help if you can share this.

thanks a lot

 

Warning: It's just a work in progress, not something I want people to download as a good product.

I also changed steering speed sensitivity to 65-75 (instead of default 50, "linear") because it seems to reduce some more sloppiness around the center (imposes a logarithm curve or just turns it into a real "linear" curve, I believe).
Been testing only with 911 GT3 at Hockenheim (because I wanted a back to back comparision with RRE, a very good reference).

All my TM driver settings are at 100%, but the max force - I have it at 70% since I feel more details using less overall forces (or perhaps avoiding clipping?)
My in-game settings are GAIN:70, LFB:15, FX: 50, DAMPING: 30

Search "#JP" lines to find the ones I've been changing and their original value.
Today I reverted back a lot those values with little overall significance.
This is really a matter of what's important to you. Mess with the values.

If someone can help, here's what I'm looking for: a more gradual sense of front wheels load under turning - it's almost like it depends more on steering angle and less on effective wheel loads.
Example: try to turn at variable speeds and it seems the force is constant and not speed/g-forces dependent until the car stops and we feel a sudden force drop - in a real car, that force should be coupled with speed.

 

thanks, I appreciate you sharing. Custom FFB is for experimenting anyway so I am not going to blame you if I dont like it
Actually I want to like AMS2 more because the base package is good, I am not giving up on it right now because of these car physics lacking, for me that is

 

Niels Heusinkveld isn't on AMS2 because he disagrees with the new tire model. I do wonder why that sim physics genius disagrees with a supposedly more "realistic" tire model especially when AMS1 felt more realistic?

This is beginning to feel like a marketing business model to make money rather than make an accurate simulator in fact i felt this way when rfactor 2 came out as a new game (selling for almost 60 bucks!!) rather than just a graphics patch for rfactor 1....

With a patch you can't justify such massive prices but with a "new" game you can and thus profit and make shareholders happy and realistic representation of real world cars comes second.

 

Niels may be smart but even he knows his limitations. He also admitted in the past that his approach to physics building weren't taking into account aerodynamic models, to which (at the time) he also gave credit that Renato was the better person in that approach I am sure he has learned a lot since. Niels is also still credited in AMS2 so while not an active employee, he has probably been consulted enough times to warrant being credited.

Niels simply prefers an empirical approach, which, while simpler to implement, still carries its limitations in what regards a very dynamic engine and approach to physics. A physical tyre model is not a scam, but it is a very complex thing to work on and the reason the tyres keep getting overhauled is because the team's knowledge over such a model keeps getting deeper and as such, they are able to extract more out of said model.

If this is grounds for you to consider it a marketing scam... so be it xD
I'll just keep enjoying the improvements.

Oh and AMS2 is an independent studio made of passionate people seeking to produce the best game they can so they are free to evolve the game as they see fit. No shareholders to please here (except for us users) but money still needs to be brought in to put food on the table and to pay for licenses.

 

I think Niels just says that maybe adding more complexity isn't the way to go, rather the opposite. That's why he prefers the empirical model. There's some beauty in simplicity, if you can find a simpler code without having to compromise the fundamental physics aspects, that could also work quite well.

Sometimes less is more and maybe SETA is a bit too complex. ACC e.g. took tyre pressures to a whole new level, but do you really want to deal with things like .1 or .2 psi difference decide over 'grip' or 'no grip' in the next corner? That's not even what happens IRL.

Didn't Einstein say 'make everything as simple as possible but not simpler'?

To be clear: I think Renato and the team are doing a fantastic job, but maybe they could find a way to simplify things and still have realistic tyre behaviour and vehicle dynamics.

 

....

 

Simplification was never the way forward in Simulations.

 

...

 

When you have no arguments to reply back with;
When you spit out RD's "Top Minds" talking points;
When you don't want to concede that your so called opinions are easily shattered,
Hit them back with

"..."

That'll show them.

 

I can only agree with you. Sometimes less really is more. I changed a lot in ffb and sometimes removed things because I felt something was wrong. The shouting was always big and dismissive, but in the end they suddenly found it better. Reiza is doing great work and making progress with the physics of the cars too. But in my opinion there is a lot to do because of the ffb. You should deal with it more intensively, especially with the standard ffb. What good is this excellent physics of the cars if the standard ffb is so bad. I hope Reiza will take care of it in the future

 

Agreed. If people want to play with the same rFactor 1 tyre models for another decade, AMS1 will always be there. I'd rather the tyre models move forward even if progress sometimes comes with some hick-ups.

 

Personally I play the standard FFB (no default + and no custom) and I prefer it over Rfuktor etc. FFB is very subjective and largely depends on your expectations and experience driving sports and race cars if any.
I don't get why people consider their own personal choice as a universal rule especially on such a subjective matter. Even different real world cars have different feeling at the wheel.