Bad tire physics could actually be a weight and inertia issue

[Zer0]

Steering wheel can go lock to lock pretty quick, watch the cool video. It will surprise you how big angle can still be controlled, I have never seen anyone did this in a game/sim.

Can this be done in GT5 using steering wheel/pedal ? Is it also possible on FM4 ? If both games can replicate what's in the video - especially the last run, without the car looking funny/unrealistic, I am happy camper

I don't know GT5 but looks pretty accurate in the demo.

 

[Ridox2JZGTE]

I don't know GT5 but looks pretty accurate in the demo.

That's me driving, trying to do it, the demo is very limited in options, SH tires is too grippy and 370z didn't have enough power. I also used stick controller

 

[Zer0]

That's me driving, trying to do it, the demo is very limited in options, SH tires is too grippy and 370z didn't have enough power. I also used stick controller

Haha ok, the physics are there and your input is smooth should be enough to replicate the same with a wheel and the apropiate tyres. Hope in the final game.

 

[I'm Motarded]

My point is that regardless of input type, a real car will obey the laws of physics. Equally, regardless of input type a sim car should obey the laws of physics. Therefore, when analising potential inertia issues, input methods are not relevant (input yes, but not method). Instant lock to lock for example is impossible in real life, but if in a sim the physics are correct, the sim car should not display anything against the laws of physics even if that extreme example was enacted.

Yes, I get that physics are physics. But if we have never seen what would happen in RL with 100% lock applied instantly how can we compare? Input is relevant, in fact it's 100% necessary to compare the results as it does affect the outcome. As you can see above the method in which the input was applied with a controller directly affected the outcome making the inertia look more realistic (or rather what we are used to seeing), because it was used in a more human-like manner. It's like dropping an apple from a building of from a table, physics dictate the apple will fall, the height will dictate the fate of the apple. you won't get the same result despite the input being the same because the method was different between the two tests.


Anyway.... So case solved? GTA behaves in a manner we would be used to seeing if the steering input is controlled in a more orthodox speed?

 

[Exorcet]

Yes, I get that physics are physics. But if we have never seen what would happen in RL with 100% lock applied instantly how can we compare?

Using the laws of physics. You don't need to have seen anything, physics tells you what will happen.

There is no reason why going lock to lock quickly should make the car use a different set of physics. Inertia is still going to exist.

Input is relevant, in fact it's 100% necessary to compare the results as it does affect the outcome. As you can see above the method in which the input was applied with a controller directly affected the outcome making the inertia look more realistic (or rather what we are used to seeing), because it was used in a more human-like manner. It's like dropping an apple from a building of from a table, physics dictate the apple will fall, the height will dictate the fate of the apple. you won't get the same result despite the input being the same because the method was different between the two tests.

Input doesn't matter, and the gravity analogy tells you why. You could predict what would happen in both cases despite different starting conditions. So if GT was a real gravity simulator, and some people threw things off buildings, and others tables, they could quite easily communicate with each other when it comes to physics.

Anyway.... So case solved? GTA behaves in a manner we would be used to seeing if the steering input is controlled in a more orthodox speed?

Input shouldn't enter the argument.

 

[Samus]

Regarding control and inertia when driving with a controller:



It's not possible with a wheel I'm told but that's still not right obviously.

 

[Johnnypenso]

I believe the input device is also relevant to the discussion. Although inertia and gravity don't change relative to the input device, the speed of the countersteer effort by the front tires can make a difference in whether you reach the point of no return when it comes to a slide or loss of control. Obviously there are times when no matter the input device the inertia of the slide cannot be overcome, but below that threshold, it is relevant whether you are going lock to lock before the critical moment or not.

 

[Exorcet]

I believe the input device is also relevant to the discussion. Although inertia and gravity don't change relative to the input device, the speed of the countersteer effort by the front tires can make a difference in whether you reach the point of no return when it comes to a slide or loss of control. Obviously there are times when no matter the input device the inertia of the slide cannot be overcome, but below that threshold, it is relevant whether you are going lock to lock before the critical moment or not.

The input changes the output yes, but that's not interesting. The question being asked is about the physics, so it's testable whether you are using a wheel, controller, or directly putting a voltage into the PS3's USB slot.

Input -> Physics -> Output

That's how you have to look at it. You don't test the physics against an output or an input. You test the output against an input and that tells you what the physics are doing.

 

[Johnnypenso]

The input changes the output yes, but that's not interesting. The question being asked is about the physics, so it's testable whether you are using a wheel, controller, or directly putting a voltage into the PS3's USB slot.

Input -> Physics -> Output

That's how you have to look at it. You don't test the physics against an output or an input. You test the output against an input and that tells you what the physics are doing.

What do you use as a benchmark to determine whether the car in-game is behaving according to the laws of physics? Anecdotal observations are too subjective, especially considering the DS3 doesn't exist in real life and it's debatable how close a wheel is to a real life wheel in terms of response time, degrees of rotation etc. How do we measure this with any reliability?

 

[Exorcet]

That's a good question. Anecdotal tests are subjective yes, but as long as you keep that in mind, you can probably make some use out of them. Look for a red flag, for example the lack of torque steer on launch. It's pretty clear that something is wrong even with an objective measurement. Lift off oversteer is another example. GT5 FF's don't have it, and you don't need any numbers to confirm this.

As far as objective measurements go, the game only really gives us a measurement of time. Figuring out distance, force, velocity, etc can be complex, but there is always a timer on the screen. The best bet for objective tests then would be time based ones. This still poses the problem of what to compare to. You can always compute the answer yourself and then you can test anything you want, but that's a bit of work. Another is looking for standard tests (like slalom data, etc) that can be copied in game, but for this you may need to use a standard input that copies that of the original test.

 

[Ridox2JZGTE]

I'm curious about max wheel angle implementation in GT6, in GT5 all cars comes with 40 degrees and its locked, unless you can edit it via hex. Changing this alone makes a lot of difference to the way physics output,if anyone interested I can share one or 2 of my cars that have max wheel angle changed to closer one of that in real life - one tuned for track racing (20 to 35 deg or 8 to 23 deg for NASCAR ) and one for drift ( 70/72 deg ) - try these on steering wheel and stick.

Big max wheel angle, the quick lock to lock input of stick controller and the tire physics could contribute to the issues we see with the crazy countersteer and corrections.

 

[dbarrade]

Input -> Physics -> Output

You are forgetting this is a game, the laws of the real world don't have to apply unless the programmer wants them to.

Let's look at what happens when you go off track, you push a button and restart the race, all of a sudden your car is back at the beginning of the track:

Input -> Breaking all the laws of physics -> Output

This is an extreme example, but I'm just showing you that these assumptions don't hold in a game.

It is very hard to hold the controller at a certain point that correlates to an angle on the steering wheel, it's also very easy to move the controller faster than you would be able to turn a steering wheel and move a rubber wheel on tarmac. Therefore in order to make the game more playable they have to alter the way in which the Input is applied to their model:

Input -> Interpolation, smoothing or other adjustments -> GAME Physics -> Output

In this case, game physics can be vastly altered by settings such as Skid Recovery Force, clearly some non real world force that is being applied to the Physics model in order to Recover a Skid.

Until we see the final GT6 and possibly find out what all these settings are, it is very hard to make a call that what you are interpreting as bad tyre physics or weight or inertia issues is going to be an issue.

Personally playing with a wheel and every aid I have access to turned off, the physics, response and feedback felt through the wheel are a significant improvement over GT5. I'm very happy with the progress, and I feel a little sad that you are possibly getting a little put off by factors that may not affect the final product.

 

[LeMansAid]

Well, we can change it to input -> game physics -> output, but that's really just semantics. No matter what the input involves, the sequence is always going to finish with physics to output, and physics can be judged by output.

I don't think that anyone is forgetting that it is a game.

 

[Exorcet]

the laws of the real world don't have to apply unless the programmer wants them to.

That is the whole reason why anyone posts here at all.

To determine what the physics say in game, use input > physics > output. The physics are whatever the rules of the system you're testing are, be it reality of a game. It can apply to Need for Speed even because even though it's an arcade game, it has self consistent physics.

Let's look at what happens when you go off track, you push a button and restart the race, all of a sudden your car is back at the beginning of the track:

That doesn't really matter.

It is very hard to hold the controller at a certain point that correlates to an angle on the steering wheel, it's also very easy to move the controller faster than you would be able to turn a steering wheel and move a rubber wheel on tarmac.

That is OK, we don't need to exclusively test using wheel inputs because the physics does not care what inputs are generated. Also, the controller is not necessarily faster and there are players good enough to drive with wheel like precision on a controller, like those who golded the X1 challenge without wheels.

Therefore in order to make the game more playable they have to alter the way in which the Input is applied to their model:

Input -> Interpolation, smoothing or other adjustments -> GAME Physics -> Output

The first two things are the same thing, input. The controller assist is input. TCS is input, ASM is input, ABS is input. SRF is physics. Whatever the car does is output.

Until we see the final GT6 and possibly find out what all these settings are, it is very hard to make a call that what you are interpreting as bad tyre physics or weight or inertia issues is going to be an issue.

I have little doubt that there is an issue with the tires, and I don't see this changing unless PD changes the physics. This might happen for final GT6, but for the GT6 demo, the physics are off. Tire grip is still digital and apparently hard coded.

Personally playing with a wheel and every aid I have access to turned off, the physics, response and feedback felt through the wheel are a significant improvement over GT5. I'm very happy with the progress, and I feel a little sad that you are possibly getting a little put off by factors that may not affect the final product.

The demo is supposed to represent the final. If the final product is different, it will be reevaluated, but there is no guarantee that GT6 final will change.

 

[dbarrade]

To determine what the physics say in game, use input > physics > output. The physics are whatever the rules of the system you're testing are, be it reality of a game. It can apply to Need for Speed even because even though it's an arcade game, it has self consistent physics.

I think you missed my point here, which was essentially that this whole:

input > physics > output

Is your belief of how GT works and may not be the case at all.

I gave a specific example of where it isn't the case to show that a Game world isn't limited to the same laws of physics that the real world is.

For all you know there may be like you said "Hard Coded" Tyre grip fixes that Polyphony have to apply in order to make the overall simulation seem more realistic.

Obviously they cannot calculate everything that happens in real world physics, so their game is going to have to take short cuts.

These fixes may be applied differently to a player using the controller than they are to a player using a wheel, most likely due to the situations we are all aware of, that it is harder to hold a specific angle of the wheel and it is easier to overcome the friction and travel required to vastly alter that angle of wheel.

There is no guarantee that they have to compile all of the inputs, feed them into something that calculates the physics and then gives an output.

I gave an example of where this is not the most efficient way to do things in the game, as the player would have to sit and wait for all the cars to drive back to the start line again, as they would in the real world which wouldn't make the game much fun.

Another example is how the AI cars behave once they are collided with compared to how your car behaves when it is collided with. These are also vastly different. PD has had to take short cuts because it is more important for your car to behave realistically in that situation than it is for an AI car, so they don't perform as accurate a simulation for the AI cars which in turn saves them processor time allowing them to have more cars on track, or perhaps dedicate more time to the players car physics.

The fact that controller players can actually compete at all and are able to gold the X1 challenge (which is the only thing I haven't completed in the game ) is a testament to Polyphony's excellent work to balance the two.

I do not believe that the balancing applies only at an input level.

Tell me, can you also use the controller to steer whilst a wheel is the main controller? Can you use it to accelerate or brake? Why do you think that is the case if the balancing is simply down to adjusting the controller input?

Tell me, are there any cars in the real world that use a joystick similar to the way people play GT? Why do you think that is the case if Polyphony are able to fix the inputs, why can't car manufacturers? Obviously a lot of people prefer to drive that way otherwise they would have bought a steering wheel right?

I believe you can prove that GT does not hold to this process:

input > physics > output

by doing the following:

Take a car out on a large curve, and follow the curve.

Using the controller, input minor fluctuations either side of the wheel direction that is required to follow the curve, now increase those fluctuations slowly until the car no longer follows the curve.

Watch the replay and notice that the wheel can be altering direction quite erratically whilst the car still seems to follow an overall approximation of the curve. Note that these fluctuations in the steering wheel do not appear to have any significant bearing on the stance of the car.

Do exactly the same thing, but with a wheel.

Note that as soon as you start to fluctuate, the stance of the car will also fluctuate, you will find that it is impossible in the replay to duplicate those fluctuations in the direction of the steering wheels and still have the car follow the curve as you could with the controller.

This will prove that the balancing is happening in more areas than the input alone, and in my opinion also very likely to impact the physics that can be applied to a controller player and thereby resulting in players like yourself who are not just trying to follow that curve, but truly wanting to experience the full physics simulation being dissatisfied.

I believe you would be in the minority of GT Game Players, so Polyphony is likely to cater for the majority.

Most of us who want to experience the full physics have gone out and bought a wheel, I would highly recommend it, you will not be disappointed.

 

[Exorcet]

I think you missed my point here, which was essentially that this whole:

input > physics > output

Is your belief of how GT works and may not be the case at all.

It's not really a belief because it can't work any other way.

I gave a specific example of where it isn't the case to show that a Game world isn't limited to the same laws of physics that the real world is.

But everyone knows this already. It doesn't change anything (and just for the record, the restart button doesn't break the laws of the physics because it's not part of the physics).

For all you know there may be like you said "Hard Coded" Tyre grip fixes that Polyphony have to apply in order to make the overall simulation seem more realistic.

Yes (though I don't think it makes things more realistic, the opposite actually) and this would be under the physics.

Input > Physics > Output refers to any physics. Like I said, you could do this with Need for Speed.

These fixes may be applied differently to a player using the controller than they are to a player using a wheel, most likely due to the situations we are all aware of, that it is harder to hold a specific angle of the wheel and it is easier to overcome the friction and travel required to vastly alter that angle of wheel.

Such as the controller steering assist that is not used with wheels, I know. This is input. All you have to do is take it into account (as an input) and you're fine.

There is no guarantee that they have to compile all of the inputs, feed them into something that calculates the physics and then gives an output.

There is because it's a video game. Whenever you drive a car, the input to output path is present.

I gave an example of where this is not the most efficient way to do things in the game, as the player would have to sit and wait for all the cars to drive back to the start line again, as they would in the real world which wouldn't make the game much fun.

Restart is not part of the physics. I guess I get what you were trying to say, but it's not really relevant.

Another example is how the AI cars behave once they are collided with compared to how your car behaves when it is collided with. These are also vastly different.

Physics. The AI cars are represented differently in the physics engine.

PD has had to take short cuts because it is more important for your car to behave realistically in that situation than it is for an AI car, so they don't perform as accurate a simulation for the AI cars which in turn saves them processor time allowing them to have more cars on track, or perhaps dedicate more time to the players car physics.

Yes. But why is that important here?

We're trying to determine if the physics for a player car are correct, or reasonably close.

I do not believe that the balancing applies only at an input level.

It does not have to. But so far, no evidence exists for it being anywhere but at input.

Tell me, can you also use the controller to steer whilst a wheel is the main controller? Can you use it to accelerate or brake? Why do you think that is the case if the balancing is simply down to adjusting the controller input?

If you can't, it's about avoiding input conflicts probably.

Tell me, are there any cars in the real world that use a joystick similar to the way people play GT?

What would this prove or disprove?

Also yes:

Why do you think that is the case if Polyphony are able to fix the inputs, why can't car manufacturers? Obviously a lot of people prefer to drive that way otherwise they would have bought a steering wheel right?

Every car can be driven via controller. They aren't because a controller is inferior and no one would know what to do with them outside of people who use them regularly.

I believe you can prove that GT does not hold to this process:

input > physics > output

by doing the following:

Take a car out on a large curve, and follow the curve.

Using the controller, input minor fluctuations either side of the wheel direction that is required to follow the curve, now increase those fluctuations slowly until the car no longer follows the curve.

Watch the replay and notice that the wheel can be altering direction quite erratically whilst the car still seems to follow an overall approximation of the curve. Note that these fluctuations in the steering wheel do not appear to have any significant bearing on the stance of the car.

Do exactly the same thing, but with a wheel.

Note that as soon as you start to fluctuate, the stance of the car will also fluctuate, you will find that it is impossible in the replay to duplicate those fluctuations in the direction of the steering wheels and still have the car follow the curve as you could with the controller.

This will prove that the balancing is happening in more areas than the input alone, and in my opinion also very likely to impact the physics that can be applied to a controller player and thereby resulting in players like yourself who are not just trying to follow that curve, but truly wanting to experience the full physics simulation being dissatisfied.

I believe you would be in the minority of GT Game Players, so Polyphony is likely to cater for the majority.

Most of us who want to experience the full physics have gone out and bought a wheel, I would highly recommend it, you will not be disappointed.

The above can be achieved with input and you've also pointed out that visual output does not need to match with physics output. I will say I've never noticed what you're talking about though. Small changes in stick position will change your path through the corner, but exceeding the maximum steering angle won't be allowed with a controller. Maybe you can clarify with video?

I've been buying wheels since GT2. A wheel doesn't help GT5 much, nor does it provide much more insight into the physics. I have not tried with GT6, but I don't see much to make me think it will be different.

 

[Wolfe]

I believe you can prove that GT does not hold to this process...

...Watch the replay and notice that the wheel can be altering direction quite erratically whilst the car still seems to follow an overall approximation of the curve. Note that these fluctuations in the steering wheel do not appear to have any significant bearing on the stance of the car.

Do exactly the same thing, but with a wheel.

Note that as soon as you start to fluctuate, the stance of the car will also fluctuate, you will find that it is impossible in the replay to duplicate those fluctuations in the direction of the steering wheels and still have the car follow the curve as you could with the controller.

This will prove that the balancing is happening in more areas than the input alone, and in my opinion also very likely to impact the physics that can be applied to a controller player and thereby resulting in players like yourself who are not just trying to follow that curve, but truly wanting to experience the full physics simulation being dissatisfied.

Oh, that? For whatever reason, Gran Turismo 5 (and GTA) depict a visual steering angle that matches the literal, 1-to-1 joystick input, while the physics are calculated from filtered input, like practically every other 3D racing game on the market. What you see in the replay is atypical for a sim, but the end result is very typical. It's the same as Forza Motorsport and all the rest, and it follows the input > physics > output sequence that Exorcet described.

Most of us who want to experience the full physics have gone out and bought a wheel, I would highly recommend it, you will not be disappointed.

I've had gaming wheels for a decade and a half. Not a single console game has proved itself worth the setup or effort. In my experience, bringing "real world" inputs into the mix just magnifies small problems into bigger ones. It's distracting to perform an action that produces "X" in a real car, but comes up with "Z" in the game. I prefer to leave the wheel for PC sims like Live for Speed.

If you ask me, between poor FFB implementation, varying quality in wheels, and effects applied by FFB that have no actual bearing on in-game physics (only affecting the car through input), a controller is kind of a superior way to evaluate game physics than a wheel. Like Exorcet said, as long as you're aware of how the game treats your inputs, there are no external factors to skew the results.

For an example, consider uneven surfaces in Gran Turismo. Wheel users will tell you that a lumpy track surface can send you off course, while controller users tend to sail through unscathed; this isn't because controller users get some sort of physics handicap, it's because uneven surfaces don't affect your heading on a physics level in the first place. It's just a FFB effect. Unless you unplug the FFB motors, you could be left completely unaware as a wheel user. Interference like that doesn't really lend itself to an objective evaluation, IMO.

 

[dbarrade]

Well, at this stage, I have to agree to disagree, the Wheel feels much better to me, and as a programmer, I know that software including Gran Turismo is not limited to input > physics > output. They can ignore the input and the physics and alter the output in any way they choose, Gran Turismo has been doing this for a long time, there are plenty of examples, roll over, the way the car moves through the air, the way the car behaves during a collision.

 

[Exorcet]

I think you're still hung up on physics being real physics. It's not. The game physics are what take the input, no matter what those physics are. What about roll over, etc., ignore the physics or input? They make look unrealistic, but that doesn't mean they aren't results of the physics.

 

[Wolfe]

Anything that involves the motion or control of the player vehicle is the game's physics. This includes things like SRF and collisions, which act directly upon the physics. By "output", we basically mean the end state immediately before the image and sound are sent to the TV. It can't be modified because it's what you get after all of the modifications are done.

In terms of programming, yes, it's possible that plugging a wheel into the game could switch the physics engine to a different "Full Simulation" model. But there are few problems with that:

1. It would be patently unfair to one group or another in online multiplayer, or internet time trial series, like those here on GTP.
2. No other simulator I'm aware of has ever had to resort to that solution, because input filtering works just fine.
3. With all the people like us who pick the game apart, if it was in GT5, I think we would know about it by now.

 

[Griffith500]

I've not been following this properly, really, but surely the fact that FFB contains feedback (from the output to the input) means that the stuff around the process in the middle ("physics" in this case) kind of is important?

In a lot of the stuff I mess around with, the input and the process are coupled; as are the process and the output, and there is usually some form of feedback into the input. Looking at a real car, that schematic of a coupled input -> physics -> output -> input loop is more accurate (schematically) than a discrete input -> physics -> output chain (depending on the models contained within, obviously).

So the question is whether all input methods offer the same total "input possibility space" as each other, and whether that makes a difference.

Or am I missing the point?

 

[dbarrade]

So, back on topic, I believe it's far more likely that the OP impression of bad tyre physics is more likely his interpretation of the input filtering in combination with non realistic elements of the game physics that can be turned off.

And along the way you've come to the conclusion that

input > input filtering > game physics > output of car body orientation

input > output of front wheels orientation

Clearly showing that we have output that bypasses the game physics because the programmers chose to do so.

You are now also assuming that they are keeping track of the filtered wheel positions and their input on the physics from frame to frame as opposed to calculating the highly erratic friction and forces introduced through the controller variations altering the angle of the front wheels.

 

[acidreach]

I like cheese

 

[finnracer]

I like cheese

Wow, that point really benefited this thread. Please expand as I, the rest of gtpanet and possibly Polyphony and Kaz himself would be most interested to learn how whether you like a certain dairy product or not could help to resolve the concerns with the tyre model in the GT Academy demo...

 

[Exorcet]

I've not been following this properly, really, but surely the fact that FFB contains feedback (from the output to the input) means that the stuff around the process in the middle ("physics" in this case) kind of is important?

In a lot of the stuff I mess around with, the input and the process are coupled; as are the process and the output, and there is usually some form of feedback into the input. Looking at a real car, that schematic of a coupled input -> physics -> output -> input loop is more accurate (schematically) than a discrete input -> physics -> output chain (depending on the models contained within, obviously).

So the question is whether all input methods offer the same total "input possibility space" as each other, and whether that makes a difference.

Or am I missing the point?

The original point as I saw it was to explain that physics are constant between wheel and controller. Input effects output and vice versa, but the physics don't change.

 

[LeMansAid]

The original point as I saw it was to explain that physics are constant between wheel and controller. Input effects output and vice versa, but the physics don't change.

At the very least, that it shouldn't. It goes back to this:

@sk8er913 it just looks to me like that GTA video you posted and possibly you are playing with the controller and aids, do you have skid recovery on?

It seemed to be suggesting something like: "Well, if you are playing with a DS3 controller with aids on, what do you expect?" Well, I expect the car to respond as realistically as possible, and at least as realistically as the game otherwise manages. I expect to be able to hook up something as ridiculous as a theremine to act as input and still have the car behave realistically, since that would be the case in real life.

To be honest, I have never actually tried the SRF thing, I always assumed that it would be similar to all of the other aids that make it feel like you are interacting with a story rather than a track surface. Is it unlike traction control and ABS? Apart from it not being a real world feature, of course. I would have imagined that it would just be another way of the game modifying your input to help steer you clear of trouble.

 

[Exorcet]

SRF is a unique aid. All aids besides SRF modify input. SRF modifies the physics.

 

[KinLM]

SRF is a unique aid. All aids besides SRF modify input. SRF modifies the physics.

^ This is very well put and completely true. Because if I'm not mistaken, SRF increases grip to the tires when they break their threshold of traction, right?

If I'm right then SRF directly changes the physics to compensate for a loss of traction, where aids such as TCS and ABS alter the input (how much throttle is applied and how much braking is applied, respectively) to help control the loss of grip.

TCS and ABS act as if a computer is helping you drive your car, SRF is acting as if the laws of physics change (when your tires break traction in real life, you begin to lose loads of grip, where SRF basically eliminates this.)

As a result, it isn't seen as very realistic like the other two are, and for a true physics test it should be turned off considering the very definition of what it does is alter how things would be in the real world, which isn't good for a realism test.

Hope I wasn't completely terrible at explaining that. 👍

 

[Ridox2JZGTE]

If GT6 simulate proper max front wheel steering angle ( inner wheel ) like their real life counterpart for each car, and then allow changes to it when full custom suspension fitted, along with individual wheel settings for camber, spring, damper, toe and caster. It will be more than enough for me


I think Ackerman steer angle is calculated in GT5, having felt the effect of altering the max turning angle.
I noticed the inner and outer wheel have different max angle and default 40 degrees in GT5 is for inner wheel.

Understanding Steering

Having 40 degrees inner wheel angle, the input controller and aids, along with the tire physics can make for some unrealistic car behavior.

GT5 and GTA demo has all cars with 40 degrees max inner wheel angle - totally unrealistic, rarely in real world cars are sold with such huge angle Even the NASCAR car in GT5 has 40 degrees I shared on my account, a NASCAR car with 12 degrees and everything else stock for oval and road course. One of GTP member here who raced NASCAR car since release day gave great feedback, he said the car felt good and more life like than with stock 40 degrees.

My BlackBird replica handles a lot better with proper max wheel angle My latest replica build : Roger Clark Motorsport WRX Gobstopper also have reduced max wheel angle - a stock GC Impreza WRX has low 30 degrees inner wheel angle, the car I build is reduced for time attack to 27 degrees. I'm sure the new owner will love it

Ferrari 458 Italia with max steering angle, I think this is around low to mid 30 degrees.

 

[Griffith500]

The original point as I saw it was to explain that physics are constant between wheel and controller. Input effects output and vice versa, but the physics don't change.

Oh, yeah, that's probably an important distinction (I guess I take it for granted!). But the different input methods are marked in the way they affect the control and, to some extent, driving style. Quite interesting, really - if the physics are spot on, but the input sucks, what you have basically sucks (GIGO).

I have to say I wasn't aware that the visual representation of steering angle wasn't the actual "physical" one the game was using, though. That strikes me as plain odd (and calls into question some of my "observations"...).