Dr. Kazunori Yamauchi Gives Lecture on Gran Turismo's Driving Physics & Production

You are taking my statement "not incorrect" out of context. I'm not implying that everything is perfect but it is not broken. But when i give a perfect 10/10 score its not just judged on physics alone. I also gave Forza 6 And Project Cars 2 perfect scores also. The way I feel about a game overall score doesn't have everything to do with this topic specifically.






PzR Slim, I never agreed to give you a video

You are very much implying that Scaff’s (and any who agree) interpretation of what is happening in GTS is wrong, whilst refusing to accept that something that happens in real life that doesn’t happen in GTS and is only because our perception from our inputs is wrong and perhaps we are using the wrong settings...

Why?
 
You are very much implying that Scaff’s (and any who agree) interpretation of what is happening in GTS is wrong, whilst refusing to accept that something that happens in real life that doesn’t happen in GTS and is only because our perception from our inputs is wrong and perhaps we are using the wrong settings...

Why?

Very good question. When in reality, you are going to sense much more than what you sense in a simulation of reality. A simulator is not reality; it will never be reality. It simulates reality and that is why we call it the "Real Simulator"
 
Very good question. When in reality, you are going to sense much more than what you sense in a simulation of reality. A simulator is not reality; it will never be reality. It simulates reality and that is why we call it the "Real Simulator"

But that is not the answer to what the Simulator does wrong...

(And now you begin this endless circle again)
 
How can it be the 'Real Simulator' when it has cars and tracks that do not exist in reality?

GT Sport does, just not all of it. You don't need real life thing to simulate reality. For example, if iRacing only had fictional cars and fictional tracks, but a realistic physics engine, would it no longer be a simulator?
 
GT Sport does, just not all of it. You don't need real life thing to simulate reality. For example, if iRacing only had fictional cars and fictional tracks, but a realistic physics engine, would it no longer be a simulator?

That's like saying a first person shooter even with alien space guns can be a simulator, A simulator of what? not reality, The whole point of a simulation is to simulate reality
 
That's like saying a first person shooter even with alien space guns can be a simulator, A simulator of what? not reality, The whole point of a simulation is to simulate reality

So if I'm driving an SMS-R car on Bannochbrae in Project CARS 2, it's not a simulator anymore?
 
Very good question. When in reality, you are going to sense much more than what you sense in a simulation of reality. A simulator is not reality; it will never be reality. It simulates reality and that is why we call it the "Real Simulator"
But it simulates it so poorly when compared to it's competition (which in turn don't do it perfectly) that it's a joke to call it the real simulator. Any luck with that video to prove your point?
 
GT Sport has some issues. That cannot be disputed. But at least I don't have to buy a economy size can of Raid just to play it like a certain other racing game that is very popular atm. ;)
 
I don't think the word simulator means what you think it means.

So realistic =/= simulation?

We have Project CARS 2 and Assetto Corsa, two games that pretty much everyone agrees are driving simulators. Yet, you can drive fictional cars on fictional tracks in both of those games. Where do you draw the line then?

People seem to twist the definition of "simulator" so much on this website that it's pretty much lost all meaning in my eyes.
 
Last edited:
One game does not need real world licensed cars or tracks to be a simulator with excellent physics engine.LFS is a good example of this,unfortunately with GT this is not the case.
 
GTS isn't as "sim" as other "sims", but a lot of us love it anyway. Can we stop this now? lol

True, people can choose to look at the game however they like, but if it gets in the way of their enjoyment then that's their problem.
 
GT has never been, nor will it ever be a sim. The "Real Driving Simulator" slogan I don't think is taken seriously by PD or anyone at this point. To me it's just kept there because it's always been there.
 
If the private road I have access to wasn't 20 miles away and it wasn't hammering down with rain I would go and film a 300 ftlb FWD car doing a hard launch right now just to put a nail in this nonsense.

However I do have tomorrow off and the weather report looks OK, so it will have to wait until then.

Now that is dedication. :)
 
In my opinion , PD have a physic engine very powerfull and well done but they don t have the right people to set the cars properly in most of the cases.
I have heard a single person spend 3 monthes to build and digitalize a single car..... They don t have the same sense of details for the settings when you see all those same values spreaded all over the setting sheet for any car , even if each car have differents behaviors , just like different shape.
What make me say that they don t have the right people to set the cars properly , it s because when you drive a gem like the Porsche 911 GT3 RS , you can really feel the work of an artist behind the wheel. The car give so much pleasure to drive, i am sure that a porsche driver must find a lot of identical sensations. I can t say that because i have never driven a porsche , but this is the idea i have from what it is about.
I have never driven such a perfect balanced road car in any racing game i have played. Congrats to you Mister Kaz, i want to see much more please.
 
In my opinion , PD have a physic engine very powerfull and well done but they don t have the right people to set the cars properly in most of the cases.
I have heard a single person spend 3 monthes to build and digitalize a single car..... They don t have the same sense of details for the settings when you see all those same values spreaded all over the setting sheet for any car , even if each car have differents behaviors , just like different shape.
What make me say that they don t have the right people to set the cars properly , it s because when you drive a gem like the Porsche 911 GT3 RS , you can really feel the work of an artist behind the wheel. The car give so much pleasure to drive, i am sure that a porsche driver must find a lot of identical sensations. I can t say that because i have never driven a porsche , but this is the idea i have from what it is about.
I have never driven such a perfect balanced road car in any racing game i have played. Congrats to you Mister Kaz, i want to see much more please.
I'm glad you're enjoying it and that you feel its a great car to drive in GTS, and the following should not detract from that at all.

The problem is however that the 911 GT3 RS has a few issues with it that do not match with the car or cars from its family of drivetrains. Its not a situation that is new to the GT series, and while it is better than in past titles, its still not quite right. Its also not a fault that I believe is unique to RR drivetrains, but rather one that happens to get highlighted more by RR (and interestingly FWD) drivetrains, which is how load transfer is modeled and the speed at which it is transferred (which would be damper related).

The key ways in which this is demonstrated in RR cars in GTS is lift-off oversteer, however the first and most fundamental thing to understand is that a RR car will naturally want to understeer (anyone wishing to dispute this point, please take it up with them - I'm quoting the Porsche Driving Book).

What should happen.

A quite reference to physics makes this actually quite clear. A car is going to understeer when the front slip angle is greater than the rear slip angle (i.e. we are getting the front tyres to do more than the rear tyres), now a RR layout exaggerates this due to the rear weight bias when moving, so if you apply the throttle during cornering (or are carrying to much speed into a corner) the low load on the front is going to provide less grip and understeer ensues.

Two ways out of that exist, more throttle and if you have enough grunt to get the back tyres to exceed the front tyres then the balance will shift to oversteer. Its not however the best way to correct it, as the understeer is going to continue to build until such a point and the end result can be a rather abrupt transition.

The other way is to reduce the throttle input (and unwind the steering if required - I will come back to this in a minute), now you need to do this very, very gently and progressively, as the idea is to shift only enough load forward to get the balance of grip between the two end back in sync and to be able to then 'play' with that balance as you corner. However should you just lift off the throttle you will get a massive shift of load to the front of the car, the front will gain a massive amount of grip very suddenly and the rear will get unloaded to a very high level and very suddenly. The result is a very rapid shift from understeer to oversteer, and in a RR car you now have a rotating car, with the majority of the grip at the rear. A rear that has a lot of weight in it (not weight and load are not the same). That rear is now going to come around and do so very, very rapidly.

Now I said I would come back to unwinding the steering, the reason why this may be needed is that if you are in a heavy understeer situation you will far more lock applied than the radius the is travelling would suggest (that's the very definition of understeer, you are traveling a wider radius than than you should be for the steering input and tyre angle). When you return grip to the front end (by whatever means) and the understeer is reduced or removed, the car is now going to want to travel along the radius the steering input and tyre angle are at, which will almost certainly be to high for the radius of corner you are attempting to get around. If you haven't unwound the steering to correct for this then the nose will 'dart' into the corner, which combined with a now more mobile rear, is what is commonly referred to as 'very bad news'.

What GTS appears to be doing

Now in GTS (and this is again better than in past titles) that initial understeer is much better modeled than its been before (mainly because it was absent in the past - with RR cars going straight to power oversteer, even ones with bugger all power), its also possible to power through this, so in this regard to model is pretty good, not great, but it acts roughly in line with what should happen.

The issues that haven't been fixed are how the load transfers from the front to the rear when the throttle is lifted, its almost as if the cars are overdamped to a massive and unrealistic manner at this point, as the load transfer happens so slowly that not only is true lift off oversteer not occurring, but it takes so long (in relative terms) that you can't even trim the line as effectively as you should be able to.

As a result its almost impossible to unsettle cars that should be subject to high levels of lift off oversteer (RR and FWD with stiff rears/soft fronts) in GTS.

Now the reason why I suspect that damping may be the issue here is that lift off oversteer is absent from both RR and FWD cars to anything like the degree it should be, and the reason why both of these drivetrains share the same trait is down to quite different factors, as such the only key area of commonality is load transfer and the speed it occurs at. The key factor in regard to the speed at which load transfer occurs is down to damping (actually the reality behind this is a little more complicated - but the general concept is fine in this regard).

Now in fairness to GTS this has always been a very tricky area to model well, and in all honesty the vast majority of console titles have never got it quite right. Only one that I can think of nails it to a really acceptable level (and interestingly that's the platform Porsche use for its sim rigs at the driver centers), and another only managing to come close to that this year (with the previous iteration not getting it right either).


TL: DR
Yep it can be a fun car to drive in GTS, but its not quite as accurate as many think, because load transfer and damping issues.
 
I'm glad you're enjoying it and that you feel its a great car to drive in GTS, and the following should not detract from that at all.

The problem is however that the 911 GT3 RS has a few issues with it that do not match with the car or cars from its family of drivetrains. Its not a situation that is new to the GT series, and while it is better than in past titles, its still not quite right. Its also not a fault that I believe is unique to RR drivetrains, but rather one that happens to get highlighted more by RR (and interestingly FWD) drivetrains, which is how load transfer is modeled and the speed at which it is transferred (which would be damper related).

The key ways in which this is demonstrated in RR cars in GTS is lift-off oversteer, however the first and most fundamental thing to understand is that a RR car will naturally want to understeer (anyone wishing to dispute this point, please take it up with them - I'm quoting the Porsche Driving Book).

What should happen.

A quite reference to physics makes this actually quite clear. A car is going to understeer when the front slip angle is greater than the rear slip angle (i.e. we are getting the front tyres to do more than the rear tyres), now a RR layout exaggerates this due to the rear weight bias when moving, so if you apply the throttle during cornering (or are carrying to much speed into a corner) the low load on the front is going to provide less grip and understeer ensues.

Two ways out of that exist, more throttle and if you have enough grunt to get the back tyres to exceed the front tyres then the balance will shift to oversteer. Its not however the best way to correct it, as the understeer is going to continue to build until such a point and the end result can be a rather abrupt transition.

The other way is to reduce the throttle input (and unwind the steering if required - I will come back to this in a minute), now you need to do this very, very gently and progressively, as the idea is to shift only enough load forward to get the balance of grip between the two end back in sync and to be able to then 'play' with that balance as you corner. However should you just lift off the throttle you will get a massive shift of load to the front of the car, the front will gain a massive amount of grip very suddenly and the rear will get unloaded to a very high level and very suddenly. The result is a very rapid shift from understeer to oversteer, and in a RR car you now have a rotating car, with the majority of the grip at the rear. A rear that has a lot of weight in it (not weight and load are not the same). That rear is now going to come around and do so very, very rapidly.

Now I said I would come back to unwinding the steering, the reason why this may be needed is that if you are in a heavy understeer situation you will far more lock applied than the radius the is travelling would suggest (that's the very definition of understeer, you are traveling a wider radius than than you should be for the steering input and tyre angle). When you return grip to the front end (by whatever means) and the understeer is reduced or removed, the car is now going to want to travel along the radius the steering input and tyre angle are at, which will almost certainly be to high for the radius of corner you are attempting to get around. If you haven't unwound the steering to correct for this then the nose will 'dart' into the corner, which combined with a now more mobile rear, is what is commonly referred to as 'very bad news'.

What GTS appears to be doing

Now in GTS (and this is again better than in past titles) that initial understeer is much better modeled than its been before (mainly because it was absent in the past - with RR cars going straight to power oversteer, even ones with bugger all power), its also possible to power through this, so in this regard to model is pretty good, not great, but it acts roughly in line with what should happen.

The issues that haven't been fixed are how the load transfers from the front to the rear when the throttle is lifted, its almost as if the cars are overdamped to a massive and unrealistic manner at this point, as the load transfer happens so slowly that not only is true lift off oversteer not occurring, but it takes so long (in relative terms) that you can't even trim the line as effectively as you should be able to.

As a result its almost impossible to unsettle cars that should be subject to high levels of lift off oversteer (RR and FWD with stiff rears/soft fronts) in GTS.

Now the reason why I suspect that damping may be the issue here is that lift off oversteer is absent from both RR and FWD cars to anything like the degree it should be, and the reason why both of these drivetrains share the same trait is down to quite different factors, as such the only key area of commonality is load transfer and the speed it occurs at. The key factor in regard to the speed at which load transfer occurs is down to damping (actually the reality behind this is a little more complicated - but the general concept is fine in this regard).

Now in fairness to GTS this has always been a very tricky area to model well, and in all honesty the vast majority of console titles have never got it quite right. Only one that I can think of nails it to a really acceptable level (and interestingly that's the platform Porsche use for its sim rigs at the driver centers), and another only managing to come close to that this year (with the previous iteration not getting it right either).


TL: DR
Yep it can be a fun car to drive in GTS, but its not quite as accurate as many think, because load transfer and damping issues.
I really don t have enough real life driving knowledges and experience to compare how close this car is from the reality.
What i wanted to say is that a wonderfull job was done on it and that the PD crew member who build it must have spent time ,passion and competence to do this. I would love to see the same level of perfection in others cars.
 
I'm glad you're enjoying it and that you feel its a great car to drive in GTS, and the following should not detract from that at all.

The problem is however that the 911 GT3 RS has a few issues with it that do not match with the car or cars from its family of drivetrains. Its not a situation that is new to the GT series, and while it is better than in past titles, its still not quite right. Its also not a fault that I believe is unique to RR drivetrains, but rather one that happens to get highlighted more by RR (and interestingly FWD) drivetrains, which is how load transfer is modeled and the speed at which it is transferred (which would be damper related).

The key ways in which this is demonstrated in RR cars in GTS is lift-off oversteer, however the first and most fundamental thing to understand is that a RR car will naturally want to understeer (anyone wishing to dispute this point, please take it up with them - I'm quoting the Porsche Driving Book).

What should happen.

A quite reference to physics makes this actually quite clear. A car is going to understeer when the front slip angle is greater than the rear slip angle (i.e. we are getting the front tyres to do more than the rear tyres), now a RR layout exaggerates this due to the rear weight bias when moving, so if you apply the throttle during cornering (or are carrying to much speed into a corner) the low load on the front is going to provide less grip and understeer ensues.

Two ways out of that exist, more throttle and if you have enough grunt to get the back tyres to exceed the front tyres then the balance will shift to oversteer. Its not however the best way to correct it, as the understeer is going to continue to build until such a point and the end result can be a rather abrupt transition.

The other way is to reduce the throttle input (and unwind the steering if required - I will come back to this in a minute), now you need to do this very, very gently and progressively, as the idea is to shift only enough load forward to get the balance of grip between the two end back in sync and to be able to then 'play' with that balance as you corner. However should you just lift off the throttle you will get a massive shift of load to the front of the car, the front will gain a massive amount of grip very suddenly and the rear will get unloaded to a very high level and very suddenly. The result is a very rapid shift from understeer to oversteer, and in a RR car you now have a rotating car, with the majority of the grip at the rear. A rear that has a lot of weight in it (not weight and load are not the same). That rear is now going to come around and do so very, very rapidly.

Now I said I would come back to unwinding the steering, the reason why this may be needed is that if you are in a heavy understeer situation you will far more lock applied than the radius the is travelling would suggest (that's the very definition of understeer, you are traveling a wider radius than than you should be for the steering input and tyre angle). When you return grip to the front end (by whatever means) and the understeer is reduced or removed, the car is now going to want to travel along the radius the steering input and tyre angle are at, which will almost certainly be to high for the radius of corner you are attempting to get around. If you haven't unwound the steering to correct for this then the nose will 'dart' into the corner, which combined with a now more mobile rear, is what is commonly referred to as 'very bad news'.

What GTS appears to be doing

Now in GTS (and this is again better than in past titles) that initial understeer is much better modeled than its been before (mainly because it was absent in the past - with RR cars going straight to power oversteer, even ones with bugger all power), its also possible to power through this, so in this regard to model is pretty good, not great, but it acts roughly in line with what should happen.

The issues that haven't been fixed are how the load transfers from the front to the rear when the throttle is lifted, its almost as if the cars are overdamped to a massive and unrealistic manner at this point, as the load transfer happens so slowly that not only is true lift off oversteer not occurring, but it takes so long (in relative terms) that you can't even trim the line as effectively as you should be able to.

As a result its almost impossible to unsettle cars that should be subject to high levels of lift off oversteer (RR and FWD with stiff rears/soft fronts) in GTS.

Now the reason why I suspect that damping may be the issue here is that lift off oversteer is absent from both RR and FWD cars to anything like the degree it should be, and the reason why both of these drivetrains share the same trait is down to quite different factors, as such the only key area of commonality is load transfer and the speed it occurs at. The key factor in regard to the speed at which load transfer occurs is down to damping (actually the reality behind this is a little more complicated - but the general concept is fine in this regard).
Now in fairness to GTS this has always been a very tricky area to model well, and in all honesty the vast majority of console titles have never got it quite right. Only one that I can think of nails it to a really acceptable level (and interestingly that's the platform Porsche use for its sim rigs at the driver centers), and another only managing to come close to that this year (with the previous iteration not getting it right either).


TL: DR
Yep it can be a fun car to drive in GTS, but its not quite as accurate as many think, because load transfer and damping issues.

911’s handle exactly as you describe – they all understeer at the limit (unless you know what you’re doing), and then the rear comes round when you come off the throttle when you have them loaded laterally. The more modern the version, the better they are, but the overall balance is inherent in a rear engine car.

I’ve only done 1 lap of the 'Ring in the GT3RS in GTS and it felt horrible – you just can't load the front. Whilst I haven’t driven an GT3RS IRL, I have driven a 991GT3 extensively (my brother-in-law had one for a year before changing to a DB11 last month)… and it does not feel like the RS does in GTS.

But I don't think the reason they don't tuck the front in when lifting the throttle is damping related – damping adjustments have a relatively minor effect on handling in GT (based on experience of tuning in GT4/5/6).

Almost certainly this is down to the stock rear toe and LSD decel settings - I haven't checked the RS, but I expect rear toe will be c.+0.50, and LSD decel will be c.30, as this seems to be the setting that PD inflict on pretty much every car.

Toe and LSD are by far the most influential settings in GT (ie; they have the biggest effect on how cars handle of all available adjustments)... and at the stock levels, they both create masses of understeer... toe all the time, LSD decel when braking or coasting off throttle.

GT stock settings also have very low front camber, which also contributes to reducing front end grip. And they run –ve front toe (c.-0.25), which is there to try and make the cars feel a bit livelier on entry, but has the side effect of reducing front mid corner/exit grip.

So when you feel the understeer created by the toe and low front camber, and you lift to shift some weight on to the front tyres, the toe and LSD decel just make the car want to continue to plow.

To make the car handle properly, I’d start by setting front and rear toe to 0.00, LSD decel to 5 (minimum), and front toe to -2.0. If it then oversteers on turn in whilst trail braking, I’d then add LSD decel (in increments of 1) until I found a stability level I felt comfortable with.

Once I’d stabilised entry, I would reduce LSD accel to adjust mid corner and exit traction – this is a much more effective way to fine tune mid corner traction than rear toe.

Interestingly, 911 race cars do run high LSD decel – it’s there by design to minimise risk of losing the rear whilst trail braking heavily. But the levels in GT, combined with the high +ve rear toe, are just too much (and they are there in all cars, so equally hamper the handling of those too).

Edit:

I see the online event was on SS tyres. My go in the RS at the 'Ring was on SH.
 
Last edited:
A simple question to all the Gamephysicsexperts in here: On which Naturephysical Law realbased since GT6 the ingame Physics Engine?

Answer: The gravity of Earth, refers to the acceleration that is imparted to objects due to the distribution of mass within the Earth. Near the Earth's surface, gravitational acceleration is approximately 9.8 m/s2, which means that ignoring the effects of air resistance, the speed of an object falling freely will increase by about 9.8 metres per second every second. This quantity is sometimes referred to informally as "little g" and in contrast, the gravitational constant G is referred to as "big G". The gravitational constant, also known as the universal gravitational constant, or as Newton's constant, is an empirical physical constant involved in the calculation of gravitational effects in Sir Isaac Newton's law of universal gravitation and in Albert Einstein's general theory of relativity.
https://wikipedia.org/wiki/Gravitational_constant

An gravitational effect in this physical calculatin is for example: Downforce is a downwards thrust created by the aerodynamic characteristics of a car. The purpose of downforce is to allow a car to travel faster through a corner by increasing the vertical force on the tires, thus creating more grip.
This effect is referred to as "aerodynamic grip" and is distinguished from "mechanical grip", which is a function of the car's mass, tires, and suspension. The creation of downforce by passive devices can be achieved only at the cost of increased aerodynamic drag (or friction), and the optimum setup is almost always a compromise between the two.
https://wikipedia.org/wiki/Downforce

No gravitationaler Effect in this physical calculation is for example: Cold inflation pressure is the inflation pressure of tires before the car is driven and the tires warmed up. Drivers are encouraged to make sure their tires are adequately inflated, as suboptimal tire pressure can greatly reduce fuel economy, increase emissions, increased wear on the edges of the tire surface, and can lead to premature failure of the tire. Excessive pressure, on the other hand, may lead to impact-breaks, decrease braking performance, and cause uneven wear (i.e. greater wear on the center part of the tire surface). Tire pressure is commonly measured in psi in the imperial and US customary systems, bar, which is deprecated but accepted for use with SI or the kilopascal (kPa), which is an SI unit. Ambient temperature affects the cold tire pressure. Absolute cold tire pressure (gauge pressure plus atmospheric pressure) varies directly with the absolute temperature, measured in kelvins.

Fact: The Moonmission in GT6 shows clearly pure Gravity, as working function in the ingame Physics Engine. Gran Turismo as "Real Driving Simulator" used real Physics Law calculated physicmodels for real working dynamic Drivingphysics and has no static Databased physical-simulated pseudo-realdriving Gameengine.

Carsetting since GT6:

1) tune the Carsetup balanced out for Handling, NOT for grip on setting Tyres!

2) when you only stiffen anything on one axle, SpringRate, AntiRollbar or Damper, you are removing mechanical grip from that axle.

3) Anybody who says Camber removes/reduces cornering grip, doesn't understand points 1 & 2.

4) Increasing SpringRate Push the wheels into the tarmac.

5) Increasing ARB tends to make the inside wheel light.

6) roll resistance is the total of SR and ARB front & Rear.

7) When tuning the Suspension AND Differential most often you need to compensate, for adjusting one setting by adjusting a few more.

8) Toeangles are used to dial in fine handling between the Axles, unless from the compromise tuned SR, ARB, Damps, Diff and Ride Height setting.

For Example: a low power FR car has wheel spin but your car is balanced in the corner, what to do??

Lower the ARB and raise the SpringRate. This will keep the roll resistance coupling the same BUT will make the inside tire more planted and capable of putting down the power. THIS WILL REQUIRE adjusting the differential to get the most out of. High ARB with Low SR will call for a very tight Diff, a low ARB with high SR will respond better to a more open Diff set up.

GTS Basicsetups: The same FF FR AWD MR RR default settings for each specific Racingclass, are allready compromise inused-Driveraids Handlingtunes and makes only Sense as Fairplay-Groundbasesetup for GT Sport Onlinecompetitions.

Did you think the Tyremodel or Gamephysics are don't real enough? Go to the Missions 6/8 30-Lap Race, set the Breakbias to +1 because it's a Porsche with minimal weightshiftingeffect and realize, that results in a Front-Rear-balanced Tirewearing and this is simply like the Reality.

bye
 
Last edited:
Back