Physic discussion thread

I tried several cars formula and tintops.
In neural the throttle is sensitive in the lower revs, seems partly caused by the extra pressure I need to get the throttle pedal moving, I have more control lowering the revs than when i am gaining revs
But I never drive in neutral,
As soon as i engage gears the response is good linear without any jumps or leaps, I never have the feeling lacking throttle control-
So I experience no throttle problems.

 

Even lower end wheels and pedals are not that bad.
My SRW-S1 combined wheel/"pedals" have a resolution of 1800/1023 steps - as can be seen here.

 

General
At Baseline (default) 50
ie/eg
1. .
At idle the pedal be it pot or l/cell has a Ohm /resistance of 'A'. Sim responds in software.
Engine idles.
With the most gradual increase in throttle even with the lightest of touch. the next resistance is registered. 'B'
2
The 1st change in the Ohm/resistance 'B' takes place. Sim responds in software engine RPM will increase to a level that is incorrect. (High)
ie in some cars the RPM may be as high as 4000 rpm
3. After maybe 'C' 'D' Ohm all is fine


Some other /points Observed
The resolution of the pot/cell seems to make little difference. The 'jump hole' as I call it (explained why earlier #383) will remain to a great or lesser degree.

1. The 'B' 1st change resistance may not be the only 'next' change of Ohm to effect unrealistic 'jump hole' 'C' D? unknown atm.
2. At sensitivity for example 18-35 and for most majority of vehicles . The transition from 'A' 'B' 'C' 'D' etc The jump hole is diminshed or gone Somewhere between the 2 numbers/settings 18-35 is what I would call The Baseline/default. That would enable player to modulate throttle as in any real vehicle. and gives a more natural/correct increase in RPM as in RL.
This 1 difference in some cases may well explain the many times we here the question why do we all feel different things in ams 2 even with the same gear.
ie
Player A "The car tries to kill you" ---- Player B "I have no problem with it"

Player 'A' is on setting baseline /default 50 , Wild Stallion with issues getting started /turns etc
Player B on 35 or less Thuroughbreed tip top
Theres more but stuff to catch up on

 

Well... tell that to Logitech . I phrased it a bit too generic, because it depends on the manufacturer and bundle. But practically speaking there are a lot of 8-bit pedals in use out there. I didn't mean that it's bad. Just a fact.

 

And the game is designed for the cars to work properly in gear. It isn't a 1:1 simulator of every mechanical element of a vehicle.

There are other compromises that are also titled toward realism when racing/in racing conditions, as opposed to ensuring the cars behave perfectly in a car park scenario.

I agree with various comments above that the throttle mapping (of the car engine) has been modelled to try to simulate the behaviour of the car while racing. This may result in some bending of the curve for certain WIP cars that are otherwise too sluggish with pure real-world data. I am sure later, when lots of refinements are completed, we will have realistic idling and throttle response while in the pit garage. In the meantime, I will accept some kludges while Reiza works out the kinks in a sim model that is far more complex than AMS was. It isn't just one simple throttle response/engine power delivery curve. It's a bunch of curves that interact with each other in some known and some not so well known ways.

The linearity of the controller performance is verifiable in the telemetry HUD screen. What we don't know for sure is whether there are any hidden or unintended dead zone or controller input overrides buried in there. For example, is there a small dead zone preventing initial response even when we have set the dead zone to 0. It wouldn't be the first sim to be victim of a hard coded minimum dead zone...

 

Everything here is conjecture. My conjecture is that maladjusting the controller sensitivity (i.e., moving it off linear unless you have low res hardware) to compensate for a modelling issue is counter-productive and a giant waste of everyone's time and effort. My conjecture is that if a car behaves well in gear and while driving/racing, like the Lotus 23, but it has the tiniest bobble between idle/1,000 rpm and 1,500 rpm in its throttle control while in neutral, this is not proof or even evidence related to a problem with some cars where there appears to be a completely missing rev range between idle and some higher range.

You couldn't get a smoother car to drive than the Lotus 23 in any sim. It is a complete pussy cat. Therefore, it is possible for any car in the sim to be programmed as a pussy cat. Those that are currently programmed like tigers may be over-the-top...but they are supposed to be much more aggressive and challenging to drive than a Lotus 23 given their engineering characteristics, so I am not shocked by this.

If a dead zone issue is exacerbating torque delivery, which makes the tigers into rabid tigers, then for sure we want it fixed. But artificially manipulating the input curves and turning the tigers into tamer cats will only result in a bunch of other more modest vehicles being tamed too much. The devs apply the curves for power delivery and responsiveness as best they can to accurately represent the vehicles. I am not going to apply a distortion factor (non-linear controller inputs) on top of their curves because all that will do is mask potential issues with the sim and guarantee that some/all of the vehicles will not be responding as intended.

I also refuse to distort professional music recordings put together with love and care by performers and sound engineers with a $29 equalizer. I prefer to listen to the music the way the creator's intended it to be heard.

No one else has to do what I do or agree with this philosophy...I am just stating my mildly educated opinion about this issue.

 

Excellent graphics

And if there was a car that was not responsive enough to our liking, we could choose 80 or 90 and have the curve flipped the other way. Then, the cars that are already a bit too jumpy would be even less controllable.

 

Yeah you can adjust sensitivity to suit your taste and ergonomics. After all it can also sometimes be difficult to modulate that very short initial throw that's the important throttle range we're talking about. That's why the sensitivity slider is there in the first place IMO.

Here focus is on that weird gap in throttle input that seems to indeed exist for some reason. That's what Reiza will hopefully review and get a full picture of what's going on. That way we're not "forced" to use a lower sensitivity value to remove that gap.

 

Really, all this conjecture could be avoided if everyone used the Logitech DFGT's pedals. They're almost binary. Problem solved.

 

I don't have a gap, and if there is a gap it is probably only in neutral, so it doesn't influence the driving,
No problem for me.
A nice thing I noticed in all this testing, was the slightly increased rpm when the clutch is used in neutral, probably the drive train friction is reduced by pressing the clutch
Perhaps the supposed gap in neutral is also a partly finished or not adjusted drivetrain friction simulation thingy.
Another thing, lots of cars have a huge torque on low rpm, it is therefore difficult to have the engine chocked, the car just starts rolling in first gear.

 

I could not let this stand. You have a better chance of coming close to the way the professionals wanted to hear the recording with that 29$ equalizer than without. I use DSP and I come way way closer to the way the producers wanted me to hear the music or how it sounds when it was recorded. Your room acoustics make a big difference as well as your speakers and timing of these speakers ,etc.... And then we have crosstalk and psychoacoustics. So you are way off the mark there. Sorry could not let this go

 

I bet the sense of the conversation is that from racing a car point of view, some let's say "more analog" cars that require a more "drifty" driving style will benefit from a more dense resolution of gas pedal positions in the low end. What this entire exercise shows is that depending on many factors (including hardware model and settings) in AMS2 could be more difficult to achieve the required sensitivity in the low end to drive those cars with that style.
If you move to modern cars, likely, there won't be any appreciable improvement because you are not balancing them with the throttle by drifting them out of the corners (e.g. GT3s, DPIs, relatively modern F1 cars etc.)
It is also partially a user subjective (based on their general driving style and gas pedal use) matter as well.
So long story short, some users may find it more relevant for the use of the sim some may not get any additional benefit.
Still an interesting subject for Reiza to look at I guess...

 

Just thinking outloud, maybe or not relevant to you.

Controller resolution have little importance except to the wheel and in some degree to the brakes.
A 8bit controller generates 2^8 steps or 2x2x2x2x2x2x2x2=256 steps of resolution,
10bit leads to 1024 steps, 12bit and you get 4096, 14bit = 16384, 16bit = 65536 positions.

Assuming for the easy of calculus, 32mm as a common displacement of a gas/clutch pedal.
Using a 8bit controller you have 8 steps for each mm of pedal displacement, more than enough
for a good input control. What a sensitive foot/muscle/brain one must have to take advantage
of a pedal with 16bit controller with 2048 steps/mm!

Changing ingame resolution from linearity (supposed to be at 50%) will result in a slope
over or under the standard straight line. The math function of the slope is defined by
Reiza and we can't change it but (if desired) can be estimated using telemetry data.
In AMS2, reducing sensitivity will result a curve under the linear BUT the extremes of
0% and 100% remains the same. This way, reducing ingame sensitivity will not result in
decreased resolution but will affect revs response, less responsive at lower input (0-50%) and more abrupt at highs (50-100%).

And things grows in complexity considering the frequency (Hz or cycles per second)
your controller and the game exchange data.

Another point is that race cars have the lighter flywheel (don't know if this term is correct)
they can in order to avoid overweight, resulting in "irregular" rpm at neutral and low revs.
This is specially noted in old cars. So I think Reiza got the right medicine but MAYBE at
a higher dosis.

 

Hey Newton just curious - but is it possible to interpreet my SRW resolution above like the pedals use 10bit(1023 positions) and the wheel +-1800 = 3600 positions(=11bit-ish ) ?

 

Never heard about existence of an odd-bit controller, I suppose it could be technically inadequate/impossible, sorry but I don't know.
The wheel (normally) is linked to the highest resolution available. Don't know about SRW but 16bit is frequent for wheels.
Wheels and pedals normally have individual controllers even if the pedals are connected to the base.

Edit: I'm wrong, The old Logitech Momo was 15bit (32768 steps)

 

Thrustmaster T-LCM Pedals connected to wheelbase via RJ cable = 12 bit.
Connected direct via USB = 16 bit.

EDIT:
Thrustmaster service said,if you are a beginner,you wouldn't notice the difference,but if you are a"profi" you will.