[GT5 physics] Internal friction and rolling friction

[psavi]

Yeah, I know. Physics threads are really getting old, but there is something I found out today that I had never realized.

I was driving a BMW Z4 M around Nordschleife today and did a good first lap but decided to try a second one. Accelerating out of the chicane before Flugplatz I missed the gearshift at 158 km/h and decided to give up the lap but continue to drive in neutral.

I couldn't believe that I managed to reach Bergwerk even having to brake a little in Aremberg and a lot in Wehrseifen.

I know that Bergwerk must be in a lower altitude than Flugplatz, but I'm certain that I would never get so far in real life. Therefore there must be something wrong with the internal and/or rolling friction of GT5.

I would like to read your thoughts on the subject.

 

[Exorcet]

All I know is I've taken F1 cars to Daytona and oriented them on the banking such that one wheel is off the ground. If that wheel was spinning, it will never stop spinning until you brake of bring it down to the ground. If you do brake, no matter how fast the wheel is spinning and no matter how light the braking, the wheel stops instantly.

Wheels have no inertia and perfect bearings.

 

[Morgoth_666]

The only real "test" of this I've done in GT5(beyond just observing it to seem rather odd in some situations) was to stop while climbing the hill at Füchsrore(I'm guessing in a stock 2010 Camaro) and then let it roll backwards towards Aremberg and only steering enough to keep it steady. After over 45 minutes of watching TV, the car had still not come to a stop.

I can't remember now exactly how far the car was moving after it reached a state of no more change because this took place quite a while ago. If I had to guess I'd say probably around a hundred yards of travel back and forth steadily for about the last 15-20 minutes of that. I'm not sure if I have a replay to check or not.

I don't believe this "feature" is new with GT5, though it may well have worked at least slightly differently in the earlier games.

 

[VBR]

Cars do seem to roll on in neutral for an unrealisticly long time. Even just coming off the gas they still roll on for far too long. There's a huge lack of friction in neutral, & not enough engine brake with the gas off.

 

[ElementRacer]

And what? You want NFS's infamous under 30mph auto brake?

 

[Samus]

And what? You want NFS's infamous under 30mph auto brake?

Yes because that is the only logical alternative.

 

[Griffith500]

Ha, I did a similar (impromptu, I was actually testing the low speed tyre model and tyre sound) perpetual motion test at Spa, Stavelot to be precise. It is disappointing.

Funny that the wheel stops spinning instantly when you brake it, but it keeps spinning for an age when you lift off the throttle (i.e. during powered wheelspin). Presumably this is because it's in gear, and because the drivetrain inertia is usually too high and very simply modeled.

So, shonky inertia modeling all around, and no respect for the law of entropy.

 

[BRADZZZ11]

In Japan I hear they have these wheel bearings which have 0 friction and a neutral gear which defies the laws of friction and so on!

 

[seanneedscar]

Ah yeah, good thread. This is exactly the type of **** I care about in my racing simulators. How far the cars roll in neutral

I hope you guys complain to Kaz about this like all the other trite things you come up with

 

[BRADZZZ11]

Ah yeah, good thread. This is exactly the type of **** I care about in my racing simulators. How far the cars roll in neutral

I hope you guys complain to Kaz about this like all the other trite things you come up with

People will always complain about things whether its on a game or in real life. You may complain about how early you have to get up for work or for school because it bugs you. Some people will complain about the gimmicky physics in a game they play often. Leave people to complain about what they want.
Also if you notice, you'll realise nobody has actually complained about it. The OP asked others about their thoughts on the point they raised.

 

[nickg07]

I tested this last night on the motor way in my mr2 in real life. At 70 mph you lose little speed very slowly in neutral. Probably less that 5 mph every 10 seconds.

Going up hill obviously the rate of deceleration increases.

 

[KiroKai]

Ah yeah, good thread. This is exactly the type of **** I care about in my racing simulators. How far the cars roll in neutral

I hope you guys complain to Kaz about this like all the other trite things you come up with

The thread is just about an observation and some examples where this can be seen have been posted, why so negative?


There are banked spots on some tracks, where I've used this little bug one or two times to have spinning wheels while I parked to get some weird/cool photos (no motion but the spinning rims, if you have the wheels spinning just a little and hit the right shutter speed it looks like a different wheel with less/more spokes sometimes).

 

[Samus]

Ah yeah, good thread. This is exactly the type of **** I care about in my racing simulators. How far the cars roll in neutral

I hope you guys complain to Kaz about this like all the other trite things you come up with

The issue was demonstrated in that scenario but clearly this is a problem that affects the entire physics engine. As Exorcet says:

Wheels have no inertia and perfect bearings.

Do you really not see that as an issue in the physics engine?

 

[MrWednesday]

I don't see perfect bearings as an issue, and lack of wheel inertia is only an issue if it's not accounted for elsewhere (which it might reasonably be as a scaled drag on the power delivery).

 

[Scaff]

I've acieved the same thing at Laguna Seca, around 2/3rd of the way down the Corkscrew you can get a car rolling back and forth across the track for a good hour (that's how long it did it when I left it before getting very, very feed-up.

As for those who ask why care, well its all related to the inertia and tyre issues that mean you can burn out in a perfectly straight line, why the tyres don't 'catch' as they should from a standing start and some of the lower speed power oversteer being more of a handful than it should be.

In my opinion its something that PD are well aware of and may be the reason why they are so keen to avoid standing starts for just about any race.

 

[Samus]

Pretty sure that is more down to the slow AI.

 

[Griffith500]

I don't see perfect bearings as an issue, and lack of wheel inertia is only an issue if it's not accounted for elsewhere (which it might reasonably be as a scaled drag on the power delivery).

Perpetual motion is an issue, though. Besides that, there are lots of little flaws in GT's physics, little "hacks" and approximations that work OK so long as you stick to certain confined rules. But if you stray outside of that scope, but still well within the scope of what is required of real-world physical objects, it all goes wrong in lots of little, subtle ways that add up to a lot of "uncanny".

You have to tackle this sort of thing properly from the ground up. It's not a question of computational resources (iRacing's physics engine was pretty much mature, in terms of approach and functionality, back in 1998), but of adapting a large database of physical objects that are tuned to function in one universe (GTs of old) in order for them to function in a new one with new rules (where GT needs to be, for realism's sake - assuming that's PD's target).

You can't do that overnight, sure, but I hope they will do it eventually. There's nothing like seeing all the little subtle effects of a proper multi-body physics engine in action (the kind the aforementioned Papyrus games got me used to), and the beauty of it is, once it's working, you can add detail with impunity - more forces to account for more, ever subtler effects - and it all reacts as it should.

A "black box" approach is fine for a game where the driving experience isn't so important - I don't think GT is one of those games.

 

[Scaff]

Just to add I very much doubt that PD would have modeled drivetrain friction in anyway other than in regard to engine to wheel bhp losses and even they could well be simply two separate values rather than a calculated value.

However rolling resistance should be being calculated to some degree.

 

[MrWednesday]

Perpetual motion is an issue, though.

I'm not convinced of that. It seems like it would be a very small drag factor and I'm not exactly clear on how ignoring it has a wide-ranging impact on vehicle behavior.

Besides that, there are lots of little flaws in GT's physics, little "hacks" and approximations that work OK so long as you stick to certain confined rules. But if you stray outside of that scope, but still well within the scope of what is required of real-world physical objects, it all goes wrong in lots of little, subtle ways that add up to a lot of "uncanny".

I don't disagree, but I think the issues arise in different areas than frictionless wheel bearings. I'm also not convinced that wheels that stop instantly when contacting the ground indicate a particular problem with the wheel model (vs. possible problems in other areas), though I could be sold on it.

 

[Griffith500]

I'm not convinced of that. It seems like it would be a very small drag factor and I'm not exactly clear on how ignoring it has a wide-ranging impact on vehicle behavior.

You're not convinced that altering a global friction component in the physics model has wide ranging effects? I'm not saying it's smack you in the face obvious, I said it was subtle. Many such subtle omissions all add up to create what I called the "uncanny" effect; physics that are not convincingly real, but your brain still accepts it because it's a game. If the drag is not being applied to the wheels, how is that in any way a correct simulation?

I'm not convinced of that. It seems like it would be a very small drag factor and I'm not exactly clear on how ignoring it has a wide-ranging impact on vehicle behavior.
I don't disagree, but I think the issues arise in different areas than frictionless wheel bearings. I'm also not convinced that wheels that stop instantly when contacting the ground indicate a particular problem with the wheel model (vs. possible problems in other areas), though I could be sold on it.

For the wheels to stop immediately, either inertia is zero, or tyre grip is infinite - so it's pretty obvious what's happening. Modeling the inertia in wheels creates all sorts of subtle handling (and braking...) effects, especially through the force feedback (if it's being done properly).

Going back to this 1998 Papy game, if you should find yourself on the roll-hoop (it happens...) with somehow only the one wheel spinning, braking causes the whole car to rock slightly in reaction to that transfer of energy (the collision geometry was low-res, so you had this very low friction fulcrum when you flipped the car). It's those sorts of subtle behaviours that get very important when you're trying to home in on the last few percent of realism.

They're even more important for motorbikes, due to the relative mass of the wheels to the bike, and all the gyroscopic etc. effects that change the handling. All of these things are important in cars, too, they're just there at a lower apparent, or perhaps "perceptible", level. They are nevertheless still important for feel and for wiping away that uncanny look.

 

[Exorcet]

I'm not convinced of that. It seems like it would be a very small drag factor and I'm not exactly clear on how ignoring it has a wide-ranging impact on vehicle behavior.

Sure it's small, but it would still be nice to have. Also if they are lumping it in elsewhere, would there be any other effects? For example what if it acts only on the centerline of the car and not individually on all four wheels? What if you want to model damage to a higher degree, but you've got no choice but to lump rolling resistance to one value at a single location?

Then there are those effects that you may not notice without back to back comparison between them being on and off. For example wheel inertia coming into play when the wheels leave the ground then land again. Rather than just instantly spinning back up or down to proper speed they're going to slip a little.

I'm also not convinced that wheels that stop instantly when contacting the ground indicate a particular problem with the wheel model (vs. possible problems in other areas), though I could be sold on it.

The wheel thing bothers me. A high fidelity sim should model them properly, they're fairly important to the whole car. In GT we also have the ability to alter car (changing wheels) though for whatever reason this has no performance impact currently. Calculating fairly accurate Moments of Inertia for different wheel designs shouldn't be too difficult. It would allow accurate performance from aftermarket wheels and the possibility of user created wheels.

 

[Tornado]

I think the physics engine not measuring wheel inertia might explain why the aftermarket wheels don't have any effect on anything; though I'm still very curious why the drag forum seems so convinced that they do. Placebo effect?

Ah yeah, good thread. This is exactly the type of **** I care about in my racing simulators. How far the cars roll in neutral

It's good to know that physics issues command your upmost attention.

 

[Scaff]

I think the physics engine not measuring wheel inertia might explain why the aftermarket wheels don't have any effect on anything; though I'm still very curious why the drag forum seems so convinced that they do. Placebo effect?

Given the generally very small gains they are saying they get I would say its within a margin of error anyway.

 

[DGXR]

It is impressive how high people's expectations are for GT5. I'm amazed how far people will search to find an error or imperfection in this game. While it is a driving simulation, it is also a game, a piece of software, and to my knowledge there is no perfect software program -- they all have bugs or omissions or some other inconsistency that people will notice and call out as a "physics error." Does anyone realize the processing power that would be needed to create a driving simulator that reproduces every tiny detail in perfect accuracy? The PS3 doesn't have that kind of power.

It just tells me that PD did a fantastic job doing everything right that really matters, if this is what people are griping about. So yes, in GT5 the wheel bearings are perfect, there is no driveline friction in neutral, and there's no air resistance at low speeds. Someone please explain how this takes away from the enjoyment of the game. Or am I completely missing the point, and this merely an observation about an incredibly microscopic GT5 factoid?

 

[Exorcet]

Does anyone realize the processing power that would be needed to create a driving simulator that reproduces every tiny detail in perfect accuracy? The PS3 doesn't have that kind of power.

Can you provide some numbers? How much would it tax the PS3 to have wheels with inertia?

Why does this plane have a suspension and tire model along with subsonic and supersonic flight model? PD could probably replace the latter two with non zero mass wheels.

Someone please explain how this takes away from the enjoyment of the game.

GT is supposed to be a simulator, they have quite a few things that are obviously wrong and/or curiously absent, and some of these things wouldn't be very taxing to reproduce.

 

[Scaff]

It is impressive how high people's expectations are for GT5. I'm amazed how far people will search to find an error or imperfection in this game. While it is a driving simulation, it is also a game, a piece of software, and to my knowledge there is no perfect software program -- they all have bugs or omissions or some other inconsistency that people will notice and call out as a "physics error." Does anyone realize the processing power that would be needed to create a driving simulator that reproduces every tiny detail in perfect accuracy? The PS3 doesn't have that kind of power.

No it doesn't have the power to do everything, and I'm not claiming it should, however some of the issues with the physics engine (of which this is just one) do not require more than the PS3 is capable of and as such PD's missing them is an issue for some of us.

It just tells me that PD did a fantastic job doing everything right that really matters, if this is what people are griping about. So yes, in GT5 the wheel bearings are perfect, there is no driveline friction in neutral, and there's no air resistance at low speeds. Someone please explain how this takes away from the enjoyment of the game. Or am I completely missing the point, and this merely an observation about an incredibly microscopic GT5 factoid?

The problem is that they didn't do "a fantastic job doing everything right that really matters", the tyre model (of which this discussion is part of) is very, very basic and the suspension model needs a lot of work still (and both have done better this gen).

Oh and its not air resistance at low speed, its rolling resistance which is a rather crucial part of the tyre/road interface physics, and as such is rather a big deal. One way you can see this issue is by taking a RWD car and doing a standing start burnout, don't touch the steering and just launch the car. It will not twitch sideways one bit (and it should - quite a lot in a car with a lot of torque) and the tyres will not gain grip in anything close to a realistic manner. So yes what we are talking about does take away from the enjoyment of the game if what you are looking for is a halfway decent physics engine.

Keep in mind that titles were managing some of this back on the PS2:



PD have only just managed to get close to that on the PS3 and Enthusia still has a better suspension and tyre model.

 

[DGXR]

Can you provide some numbers? How much would it tax the PS3 to have wheels with inertia?

No I cannot, sorry. My point was, even if PD did include proper inertia and friction models, someone on here would come up with something else to gripe about, as if it's reasonable to expect a computer game to have no imperfections of any kind. Pretty simple statment, I'm just saying.

No it doesn't have the power to do everything, and I'm not claiming it should, however some of the issues with the physics engine (of which this is just one) do not require more than the PS3 is capable of and as such PD's missing them is an issue for some of us.

The problem is that they didn't do "a fantastic job doing everything right that really matters", the tyre model (of which this discussion is part of) is very, very basic and the suspension model needs a lot of work still (and both have done better this gen).

Yes the tire and suspension models are imperfect. Points well taken, thanks. 👍

For the most part, I understand how to tune and properly drive a car in a simulation game. My problem is that I care more about just having fun with the game and I should not engage myself in deep discussions about the accuracy of physics models.

 

[Exorcet]

No I cannot, sorry. My point was, even if PD did include proper inertia and friction models, someone on here would come up with something else to gripe about, as if it's reasonable to expect a computer game to have no imperfections of any kind. Pretty simple statment, I'm just saying.

Are you sure? This is not even happening here. In my own posts you'll notice I've made a much bigger deal about the wheel inertia than the perfect bearings. I could live with the latter, but the former seems like a big omission from a simulator.

I'd hardly call anything here griping. Wheels and tires are pretty important to cars right? So shouldn't a simulator take the time to simulate them properly? Obviously PD can't model a virtual universe down to quarks, but the shortcuts they are taking are sometimes very obvious, as is the shortcut taken here. The inertia of the wheels is what differentiates a car from a sled. If you don't model the wheels properly, you're driving a sled, not a car.

For the most part, I understand how to tune and properly drive a car in a simulation game. My problem is that I care more about just having fun with the game and I should not engage myself in deep discussions about the accuracy of physics models.

And the physics has an impact on the fun for some people. Hence why some driving games are sims and some are arcade. GT has glaring problems that are easily spotted. As long as those remain, and it claims to be a sim, threads like this one are easily justified.

 

[KinLM]

Not trying to deny that this is a pretty major problem, but PD have much bigger problems that they need to take care of first.

 

[Griffith500]

Not trying to deny that this is a pretty major problem, but PD have much bigger problems that they need to take care of first.

There is no "first". It'll get done in parallel with everything else, if they plan to do it, that is.