GT5 Tire Model

[machschnel]

Just wondering if anyone can explain the GT5 tire model and it's limitations. Reading through a few threads (mainly F4 vs GT5 physics discussion) it seems GT5 is lacking a bit in it's tire modelling and I'd like to know more about it.

I've been playing Forza 4 for a few months now and whilst I love it, I refuse to buy another wheel (already own an G27). While the game pad experience is certainly excellent, I feel as though I could get more enjoyment out of GT5 and my G27 than Forza and a controller. My only concern is GT5's apparent lack of a tire model, so any information on the subject would be very useful.

Thanks in advance.

 

[mikecustom]

First, tire wear is completely wrong, in slow ff cars, it's fine but once you get into faster rwd cars, the tires wear too fast no matter the compound.

Second, tire compound in gt5 only act as grip multipliers and there's no deformation, this means that they give the same results for all cars, a slow hatchback can get the same g forces trough a corner as a mr supercar if both are using the same tires, you would think the solution would be either downgrade tires on slow cars or upgrade tires on faster cars, but if you upgrade, then you'll get too much grip, if you downgrade, you get realistic levels of grip but you'll get a car that's slower than it's real life counterpart when it comes to accelerating and braking,
it changes from car to car but with gt5's tire model, you get either too slow or too much grip.

 

[Olegonic]

Almost no production car is equipped with Sport tires. Exception - may be Enzo.
I use simple "Rule" :
- all my cars below 250 HP are equipped with Comfort Hard tires ( HONDA Civic, Toyota FT86 '12, MINI)
- below 400 HP are equipped with Comfort Medium tires ( HONDA NSX, Nissan 350, 370Z, BMW M3 '07, Ford Mustang, Dodge challenger, Chevrolet Camaro SS10 )
- super cars are equipped with Comfort soft tires.(Almost all Ferrari, Alfa Romeo Competizione C8, BMW M5, Vipers, Vettes, FORD GT, Mers SLS AMG, MacLaren, Nissan GTR)
Comfort tires are "road" tires and they almost match to real life lap times.
Besides, when cars hot lapping in real life they definitely tweak suspension and use if not sport tires then softer compound than is between Comfort Soft and Sport Hard tires. So, your time can be lower or different from real life. Example Nissan GTR record on Nurb. It is 7'21'xxx now. You can not do it on Comfort Soft tires. Close tire to that time could be Sport Hards. But, again, cars for hot lapping are different from production cars whatever they say about them. A few things tweaked.
Sport tires are semi-slicks.
My RM cars are equipped with Sport Hard tires ONLY.
My TC cars are equipped with Sport Mediums ONLY.
I don't use Sport Soft.
Race cars ( Formula 1, Audi R10, Mazda 787B, '65-'67 race cars) are equipped with Racing Hard tires.
My tool is tuned G27, No assists, ABS OFF.

 

[Olegonic]

EDIT:
Just did few tests on Monza. Car Ferrari 458 Italia NO tuning. Arcade Mode - means no engine brake in. All assists OFF. Brake ballance 2-1. Grip - real. Tires Comfort Soft.
Lap time - 2'00'xxx
Real lap time on this track by tuned Ferrari 458 GT3 is 1'39'xxx
Looks reasonable, huh?

 

[rumblina]

Might be better to compare same against same. Look for the official time of a car at the Nurb car and compare it that way, that would be a better comparison.

 

[Rotorist]

EDIT:
Just did few tests on Monza. Car Ferrari 458 Italia NO tuning. Arcade Mode - means no engine brake in. All assists OFF. Brake ballance 2-1. Grip - real. Tires Comfort Soft.
Lap time - 2'00'xxx
Real lap time on this track by tuned Ferrari 458 GT3 is 1'39'xxx
Looks reasonable, huh?

To be honest, no. This vehicle cannot be beaten by its racing brother in such a "straight" track by so much time difference. And if you search for it, you will find that every supercar nowadays wears semi-slick tires and special designed ones. So, the correct tire for 458 Italia is SH or even SM. But since SM have much grip in GT5, SH would be the right one.

 

[jabofu]

Almost no production car is equipped with Sport tires. Exception - may be Enzo.
I use simple "Rule" :
- all my cars below 250 HP are equipped with Comfort Hard tires ( HONDA Civic, Toyota FT86 '12, MINI)
- below 400 HP are equipped with Comfort Medium tires ( HONDA NSX, Nissan 350, 370Z, BMW M3 '07, Ford Mustang, Dodge challenger, Chevrolet Camaro SS10 )
- super cars are equipped with Comfort soft tires.(Almost all Ferrari, Alfa Romeo Competizione C8, BMW M5, Vipers, Vettes, FORD GT, Mers SLS AMG, MacLaren, Nissan GTR)
Comfort tires are "road" tires and they almost match to real life lap times.
Besides, when cars hot lapping in real life they definitely tweak suspension and use if not sport tires then softer compound than is between Comfort Soft and Sport Hard tires. So, your time can be lower or different from real life. Example Nissan GTR record on Nurb. It is 7'21'xxx now. You can not do it on Comfort Soft tires. Close tire to that time could be Sport Hards. But, again, cars for hot lapping are different from production cars whatever they say about them. A few things tweaked.
Sport tires are semi-slicks.
My RM cars are equipped with Sport Hard tires ONLY.
My TC cars are equipped with Sport Mediums ONLY.
I don't use Sport Soft.
Race cars ( Formula 1, Audi R10, Mazda 787B, '65-'67 race cars) are equipped with Racing Hard tires.
My tool is tuned G27, No assists, ABS OFF.

I have the same approach when choosing the right tyre compount. But in reality many Sports cars come with very good street tyres that are nearly as grippy as sports tyres. It really depends on the car and the make. They wont deliver you a Ferrari 458 with a compount you will find on a Honda Civic for example. In GT5, when the tyres are heated up, they will grip quite good.

I think GT5 has a tyre model that must be quite advanced to some degree. On the other hand, the fact that you cant change tyre pressure, tyre width and height is not very convincing. Either way the tyre model is crap or PD is just very secretive about their technologies. It is definetly not on par with the tyre physics of Forza 4 or rfactor 2.

 

[Schwartz]

Just wondering if anyone can explain the GT5 tire model and it's limitations.

GT5's tire model has come under alot of scrutiny from the cynics and fans alike. While the grip physics have remained relatively unchanged through the updates, Polyphony has tweaked the tire wear model a number of times in the past. But before version 2.0 (I think?) soft compounds often times had a longer life span than hard ones, due to better grip resulting in less sliding. In the 2.0 update Polyphony increased tire wear rate in soft compounds, causing them to last an unrealistically short amount of time.

The grip physics of all racing tires are thought by most to be way too unrealistically grippy compared to their real life counterparts.

 

[Exorcet]

To be honest, no. This vehicle cannot be beaten by its racing brother in such a "straight" track by so much time difference. And if you search for it, you will find that every supercar nowadays wears semi-slick tires and special designed ones. So, the correct tire for 458 Italia is SH or even SM. But since SM have much grip in GT5, SH would be the right one.

Monza isn't straight. There are a lot of high speed corners where downforce will translate into massive speed gains on corner exit.

GT5's tire model is simply a grip multiplier that basically ignores the car when calculating cornering grip. This allows a Samba bus to corner like a sports car as long as the tires are the same. I think that braking is the same where brake distance is more or less determined by your tires. The only thing besides tire compound that makes a difference is downforce, but the aerodynamics model is just as wrong as the tires are.

The tires also don't transition from grip to sliding very well. It's like the tires only have two states, full traction and low traction. In reality limit behavior should vary with each tire. Real life racing tires like to lose grip quickly when pushed over the limit, but GT5 doesn't seem to capture that very well.

 

[research]

Great post, Exorcet. Very perceptive and well stated.

👍



Hopefully in GT6 we will see some of these crucial things addressed. Wouldn't hurt to have tire temp zones, tire widths/profiles, and tire pressures. Imagine the tuning and set-up possibilities.

(ok, yes, accurate suspension tuning would also be nice to have lol)

 

[Ridox2JZGTE]

I think that braking is the same where brake distance is more or less determined by your tires.

Brake distance will vary even with the same tires, it depends on the brake balance strength, higher number will generate more braking force and generally shorter stopping distance. A very low brake balance numbers like 2/0 will not use the most of tire grip during braking. I have done a thorough tests and posted the result here :

https://www.gtplanet.net/forum/showthread.php?p=7370477#post7370477

 

[Olegonic]

Please remember David Coulthard SLS AMG event where we can only use Comfort Soft tires.
More, in Asia Gran Turismo official competition event on Toyota FT86'12 the only tires could be used were Comfort hard. PD explained that they are close to real Pirelli or something tires in real life by performance.
On conclusion, I would like to say that tires you use depends on what you expect from GT5 game. If you want to race fast without effort from your side then Racing Soft is what you need. I want complete, as possible, submersion in driving. When I use less grip tires I am focused on car control much more than on gripper tires. I see the speed when I enter 90 degree turn, if it is unreal, then I equip my car with tires of less grip.
Polifony Digital just simplified tire model in GT5 to be accessible and adoptable for different users including DS3 ones.
In new cars dealership cars equipped with tires which are not equal to real life just to let most of the people to take that car on track and drive.
When you want the game to be more challenging there is nothing left but make driving harder and precise.

 

[Ridox2JZGTE]

^ I agree with that, I always have comfort hard/medium tires available for use on every street car up to 600pp

EDIT:
Just did few tests on Monza. Car Ferrari 458 Italia NO tuning. Arcade Mode - means no engine brake in. All assists OFF. Brake ballance 2-1. Grip - real. Tires Comfort Soft.
Lap time - 2'00'xxx
Real lap time on this track by tuned Ferrari 458 GT3 is 1'39'xxx
Looks reasonable, huh?

I did the same a while ago, Ferrari 458, stock, no tuning, Monza, rain 100% weather, all assist off, comfort medium, BB 7/5, got a lap time of 2:07.xxx

 

[ChevyFan1912]

Racing hards are good for durability. They will easily last 15 laps around daytona at 220mph with only right rear tire wear on a 800hp Z06

 

[sumbrownkid]

Racing hards are good for durability. They will easily last 15 laps around daytona at 220mph with only right rear tire wear on a 800hp Z06

If you toe them out slightly, you can probably eek out another lap or two.

 

[Samus]

Please remember David Coulthard SLS AMG event where we can only use Comfort Soft tires.
More, in Asia Gran Turismo official competition event on Toyota FT86'12 the only tires could be used were Comfort hard. PD explained that they are close to real Pirelli or something tires in real life by performance.
On conclusion, I would like to say that tires you use depends on what you expect from GT5 game. If you want to race fast without effort from your side then Racing Soft is what you need. I want complete, as possible, submersion in driving. When I use less grip tires I am focused on car control much more than on gripper tires. I see the speed when I enter 90 degree turn, if it is unreal, then I equip my car with tires of less grip.
Polifony Digital just simplified tire model in GT5 to be accessible and adoptable for different users including DS3 ones.
In new cars dealership cars equipped with tires which are not equal to real life just to let most of the people to take that car on track and drive.
When you want the game to be more challenging there is nothing left but make driving harder and precise.

Hilarious, you think the basic tyre model is on purpose, they actually had something far better? Yeah right.

 

[SavageEvil]

Hilarious, you think the basic tyre model is on purpose, they actually had something far better? Yeah right.

Considering many car companies use PD simulations for virtual test drives of theirs cars I would wager to guess they have a much deeper tire modeling system that you with your ignorant comment believes. PD themselves knows exactly which tires match closest to realistic lap times for each car, which is why when you had to do the contests the car was locked to a certain tire type. Disgruntled players like to yelp nonsense before they actually stop and think about what it is they are saying. Would be great if PD had recommended tire type for each and every car in the game would be a great help to those who want to drive with the least amount of in game help as possible.

 

[Griffith500]

I've no doubt they do have more "sophisticated" tyre models floating about at PD towers, but getting them to behave themselves probably isn't easy. It took iRacing long enough to roll out its improved model once it was technically working, and LFS is still waiting for its new model to even get to the implementation stage, as far as I know.

The problem with tyre physics in simulators is that we've long since reached the limit of usefulness of well-documented models like Pacejka (industry standard, not just for games, but the "public" version is a "bit" unstable at low speeds, which is why most sims and games that use versions of it have a low-speed "hack". Note that GT5 clearly has a low-speed hack) or the harder-to-ground in-reality "brush model". Most of these models are unsatisfactory at or near the limit of adhesion, partly because of a lack of real-world data for this region and because the models just weren't aimed at that behaviour region in the first place (Pacejka) or don't effectively model the underlying physics (brush model) etc.

These shortcomings have led to games developing their own proprietary models from first principles, for example, by modeling the entire carcass in fine detail and then finding a way to approximate that behaviour in real-time. This also means they're all different, and once again a lack of real-world data hinders progress. Most freely-available tyre data is only presented in terms of the Pacejka-like models - remember Forza 4's claim of no longer "interpolating" tyre data, which probably alludes to them using the underlying physical measurements instead of the magic formula coefficients used in Pacejka-like models, because of their new partnership with Pirelli.


Anyway, it's important to separate things like model fidelity and sophistication from "accuracy"; people often say GT5's "tyre model" is "unrealistic" because wear rates are too high. In reality, all that shows is that the wear rate is too high, and doesn't necessarily point to any underlying issue with the tyre model itself. A game like GT is doubly screwed because the vast array of tyres meant to be represented cannot ever be properly simulated, so we will always have to put up with gross approximations. Which is fine so long as we're always given lots of options, which effectively means more compound types, tread types and tyre construction and size options. Which will require a new model to get right without using broad, messy brush strokes.

 

[Raitziger]

Forza 4's claim of no longer "interpolating" tyre data, which probably alludes to them using the underlying physical measurements instead of the magic formula coefficients used in Pacejka-like models, because of their new partnership with Pirelli.

I think that just means that they now use the Pacejka equation, and not the look-map with interpolation. That is of course more accurate as look-up maps (3d tables) have limited resolution. But in practice I do not believe there would be any difference. And if there is, the map resolution was not correctly selected.

 

[Scaff]

Hilarious, you think the basic tyre model is on purpose, they actually had something far better? Yeah right.

Considering many car companies use PD simulations for virtual test drives of theirs cars I would wager to guess they have a much deeper tire modeling system that you with your ignorant comment believes. PD themselves knows exactly which tires match closest to realistic lap times for each car, which is why when you had to do the contests the car was locked to a certain tire type. Disgruntled players like to yelp nonsense before they actually stop and think about what it is they are saying. Would be great if PD had recommended tire type for each and every car in the game would be a great help to those who want to drive with the least amount of in game help as possible.

The pair of you need to learn to post without the borderline personal digs and attacks. Fail to understand the difference between arguing the point and the person and you will not be posting here for much longer.

Now in regard to manufacturers using GT for 'test drives', its a marketing exercise and has no bearing on the accuracy of the tyre model at all. That PD recommend certain tyres for each car doesn't then equate to accuracy in the tyre model.

Enough clear issues exist with GT5's tyre model to indicate that, including:

• Grip multiplier used between compunds
• Tyre width and profile have little to no effect
• Grip transition is digital, lacking almost any degree of progression
• Deformation is not modeled, either withing the tyre model or visually
• Tyre pressure appears to be a fixed value and certainly can't be adjusted

I've no doubt they do have more "sophisticated" tyre models floating about at PD towers, but getting them to behave themselves probably isn't easy. It took iRacing long enough to roll out its improved model once it was technically working, and LFS is still waiting for its new model to even get to the implementation stage, as far as I know

If they do have them then I would love to know why at least some of the work has not been included in the tyre model for GT5 (either at launch or since). Personally I would rather have a partially accurate tyre model that has a good degree of transition than what we have today.

The problem with tyre physics in simulators is that we've long since reached the limit of usefulness of well-documented models like Pacejka (industry standard, not just for games, but the "public" version is a "bit" unstable at low speeds, which is why most sims and games that use versions of it have a low-speed "hack". Note that GT5 clearly has a low-speed hack) or the harder-to-ground in-reality "brush model". Most of these models are unsatisfactory at or near the limit of adhesion, partly because of a lack of real-world data for this region and because the models just weren't aimed at that behaviour region in the first place (Pacejka) or don't effectively model the underlying physics (brush model) etc.

Quite agree that tyre models are hideously complex things, and getting them even partially right is tricky, but I don't see that as a reason to move things on at such a mind numbingly slow pace. What PD have as a tyre model right now is a long way from we should have and they should be capable of.

I also don't buy the 'perfection' path either, PD have managed to get as close to accurate as has been possible within the limits of the titles hardware in the past, yet not with GT5 and tyres. Which means that either the PS3 is not capable of crunching the numbers (which I don't buy) or a tyre model that covers as many bases as they can (within the limits of the system) is not a priority.

These shortcomings have led to games developing their own proprietary models from first principles, for example, by modeling the entire carcass in fine detail and then finding a way to approximate that behaviour in real-time. This also means they're all different, and once again a lack of real-world data hinders progress. Most freely-available tyre data is only presented in terms of the Pacejka-like models - remember Forza 4's claim of no longer "interpolating" tyre data, which probably alludes to them using the underlying physical measurements instead of the magic formula coefficients used in Pacejka-like models, because of their new partnership with Pirelli.

You are quite right that T10 and Pirelli didn't just take the Pacejka models (which are very basic in reality and not heavily used outside of broad modeling) but compiled the data straight from tyre tests, and when that data don't exist Pirelli carried out the required tests (for drifting as an example), as the following explains:

Tires:
In Forza, we’ve always focused on our tire simulation. As a result, we’ve struck up multiple partnerships over the years to get tire data. In the past, we’ve worked with Goodyear, Toyo, and Michelin. We also used reference data from the Society of Automotive Engineers (SAE). In past Forza Motorsport games, we even featured tire carcass flex, tire surface heat, as well as tire gas heat and expansion. However, the model was based on a combination of data from all of our previous partnerships. And we had to do fix-up on the data to make all of the different data formatting play nicely together.

For Forza 4, we took an entirely new approach to our tire simulation. This time, we threw all of the old data away and asked Pirelli to provide us with all-new data for everything. Pirelli did custom tests on a huge variety of tires to cover all of our cases—including tire width and height, compound, inflation pressure, heat, wear, sidewall height, load, angle, etc. We then changed our system to accept the real-world data directly and without any fix-up from us at all. This means that the tires in Forza 4 behave exactly as the Pirelli test tires did, even in complex situations where multiple parameters are changing rapidly.

Just looking at graphs of the Pirelli data versus our Forza 3 data, there are several obvious differences. Unfortunately, I’m not permitted to expound on the data differences. As part of getting this level of access and partnership from Pirelli, we have to keep their data completely under wraps. From a gameplay perspective, these changes are most noticeable as you move load from tire to tire. There is a lot more nuance in the model than ever before.

Source - http://forzamotorsport.net/en-us/underthehood2/

While it is of course a marketing piece for both companies it does give an insight into what data and tests were carried out, and this kind of thing from PD is long overdue. That T10 went on the record to say that they got parts of the (low speed) tyre modeling wrong for FM3 due to assumptions they made also points towards a degree of transparency that PD could well do with.

Now that's not to say that T10's approach is without compromises, it is and to be honest one that is quite clear to see. By using data from only one company (and that's down to the level of secrecy tyre firms have, the lack of standardised testing and the limits of the hardware being able to handle the volume of differing information) any car that doesn't come with Pirreli as OEM tyres is not going to be factory right. Now its not going to be wrong, but if the OEM tyres are Dunlop then its going to feel different.

T10 are also still missing a few areas (not a major surprise given how complex the subject in question is), mainly in regard to deformation. While lateral deformation is modeled reasonably well, vertical deformation is limited and longitudinal deformation is none existent. So T10 still have a way to go with regard to the tyre model. That doesn't however change the situation that PD are playing catch-up in this regard. It also shows that you don't need a 100% complete tyre model to make a big difference.

Anyway, it's important to separate things like model fidelity and sophistication from "accuracy"; people often say GT5's "tyre model" is "unrealistic" because wear rates are too high. In reality, all that shows is that the wear rate is too high, and doesn't necessarily point to any underlying issue with the tyre model itself. A game like GT is doubly screwed because the vast array of tyres meant to be represented cannot ever be properly simulated, so we will always have to put up with gross approximations. Which is fine so long as we're always given lots of options, which effectively means more compound types, tread types and tyre construction and size options. Which will require a new model to get right without using broad, messy brush strokes.

I quite agree that the difference between issues such as wear (i.e. the rate being too high) and unrealistic (grip progression) need to be determined. However that doesn't mean that simply having a wide range of compounds available is a valid reason to not have a tyre model in place that does the job better than the one we have right now.

These differing compounds may well only differ in terms of a grip multiplier, which negates the 'range = more difficult' argument, and even if that were not the case it doesn't get over the fact that T10 manage it with a range of profile and width options for each compound. They may have less compounds to model, but in terms of number of 'options' to model I would suggest that a lot more exist.

PD do a lot right and have done a lot right in the past, however right now GT5 is stuck with a tyre model that doesn't cover the bases that its reasonable to expect in this day and age.

 

[PAPPACLART]

We are currently running a stock Supercar event using sports Medium tyres, grip real no aids etc

Our members are all pretty quick with some being top 50 in the UK/US in the recent GT Academy finals.

As an example of Lap times here are some times. (note we are 2 events in, 2 races per event using 2 cars)

Suzuka GP:

Lexus LFA
2.12.859


Ferrari F430
2.14.062


FUJI GP/F

Alpha 8C
1.51.779


Corvette ZR1 C6
1.46.340


The fastest real recorded road car time around Suzuka is an F40 at 2:25, not sure how competitive that time is but either way I still think that the GT5 times with sports mediums are too fast, comfort soft would deliver more realistic times. Source Fastest laps


Best Fuji Speedway (post 2005 and not sure if F or GT track) is 1:50 from a GT2 Porsche. Source Fastest Laps

We basically run sports tyres on most race cars. To get a racing tyre (Race Hard) you got to have more than 500bhp at least and be a full on race car but still most of that power end up on sports softs.

Our Current MX5 TC event is using sports mediums, and we are running them at 276bhp.

Sports cars in real life may indeed wear stickier semi slick tyres, but in GT5 you have to run a least 2 tyres lower/harder then what you would assume to get a realistic cornering/G. Besides, running RM or RS along with being way too grippy just simply wear way too fast!!

 

[TomBrady]

Almost no production car is equipped with Sport tires. Exception - may be Enzo.
I use simple "Rule" :
- all my cars below 250 HP are equipped with Comfort Hard tires ( HONDA Civic, Toyota FT86 '12, MINI)
- below 400 HP are equipped with Comfort Medium tires ( HONDA NSX, Nissan 350, 370Z, BMW M3 '07, Ford Mustang, Dodge challenger, Chevrolet Camaro SS10 )
- super cars are equipped with Comfort soft tires.(Almost all Ferrari, Alfa Romeo Competizione C8, BMW M5, Vipers, Vettes, FORD GT, Mers SLS AMG, MacLaren, Nissan GTR)
Comfort tires are "road" tires and they almost match to real life lap times.
Besides, when cars hot lapping in real life they definitely tweak suspension and use if not sport tires then softer compound than is between Comfort Soft and Sport Hard tires. So, your time can be lower or different from real life. Example Nissan GTR record on Nurb. It is 7'21'xxx now. You can not do it on Comfort Soft tires. Close tire to that time could be Sport Hards. But, again, cars for hot lapping are different from production cars whatever they say about them. A few things tweaked.
Sport tires are semi-slicks.
My RM cars are equipped with Sport Hard tires ONLY.
My TC cars are equipped with Sport Mediums ONLY.
I don't use Sport Soft.
Race cars ( Formula 1, Audi R10, Mazda 787B, '65-'67 race cars) are equipped with Racing Hard tires.
My tool is tuned G27, No assists, ABS OFF.

This is a very accurate solution in almost all cases. Good job.

For those who want the most out of GT5, use this.

 

[Cote Dazur]

Very interesting and informative thread, a must read for all GT5 enthousiast. At the end of the day the more we are going away from how fast we can lap to how real the lap felt, the closer we are getting to simulation and away from arcade racing.
I agree with all saying Comfort tyres are the most beleivable to drive on most cars.

 

[Griffith500]

...
If they do have them then I would love to know why at least some of the work has not been included in the tyre model for GT5 (either at launch or since). Personally I would rather have a partially accurate tyre model that has a good degree of transition than what we have today.

It might be arguable as to whether the differences are that noticeable, but it is likely the updates to the physics with Spec II and around the GT-Academy weren't just plucked from thin air. Besides, this sort of coding is all about experimentation - if you don't "play", you get nothing. They'll have plenty on their lab benches. I wonder if the lack of a truly new tyre model has anything to do with the Standards - yet another reason for people to hate them!

Quite agree that tyre models are hideously complex things, and getting them even partially right is tricky, but I don't see that as a reason to move things on at such a mind numbingly slow pace. What PD have as a tyre model right now is a long way from we should have and they should be capable of.

I also don't buy the 'perfection' path either, PD have managed to get as close to accurate as has been possible within the limits of the titles hardware in the past, yet not with GT5 and tyres. Which means that either the PS3 is not capable of crunching the numbers (which I don't buy) or a tyre model that covers as many bases as they can (within the limits of the system) is not a priority.

I'm not advocating such conservatism, and if I seemed to be defending it, it wasn't my intention. I'm not sure PD have ever been truly accurate regarding tyres (unless we're talking about laptimes, but you could get accurate laptimes with Wipeout physics), so there's no reason to see GT5 as a step backwards even if it's not the step forwards some people wanted. Forgive me if I see "perfection path" as being deliberately provocative, but the fact stands that you cannot use fixed data to model dynamic processes, so at some point you have to move away from those models which only use such data. If you're suggesting they should just update GT5's model with better values, then I agree. But that doesn't mean they shouldn't try to improve the model at all - which I can't quite understand your reasons for saying so when in the rest of your post you're "complaining" that GT5 is not moving things forwards.

You are quite right that T10 and Pirelli didn't just take the Pacejka models (which are very basic in reality and not heavily used outside of broad modeling) but compiled the data straight from tyre tests, and when that data don't exist Pirelli carried out the required tests (for drifting as an example), as the following explains:

http://www.youtube.com/watch?v=u02Wq138ouw
Source - http://forzamotorsport.net/en-us/underthehood2/

While it is of course a marketing piece for both companies it does give an insight into what data and tests were carried out, and this kind of thing from PD is long overdue. That T10 went on the record to say that they got parts of the (low speed) tyre modeling wrong for FM3 due to assumptions they made also points towards a degree of transparency that PD could well do with.

Now that's not to say that T10's approach is without compromises, it is and to be honest one that is quite clear to see. By using data from only one company (and that's down to the level of secrecy tyre firms have, the lack of standardised testing and the limits of the hardware being able to handle the volume of differing information) any car that doesn't come with Pirreli as OEM tyres is not going to be factory right. Now its not going to be wrong, but if the OEM tyres are Dunlop then its going to feel different.

T10 are also still missing a few areas (not a major surprise given how complex the subject in question is), mainly in regard to deformation. While lateral deformation is modeled reasonably well, vertical deformation is limited and longitudinal deformation is none existent. So T10 still have a way to go with regard to the tyre model. That doesn't however change the situation that PD are playing catch-up in this regard. It also shows that you don't need a 100% complete tyre model to make a big difference.

Yes, more of this in the games industry, please. Let's hope Avon get a prod, they like to release their data. Also, the original Pacejka model is "basic", but there are newer versions which have added other "features". The models were all intended to be used for vehicle engineering purposes, as far as I know. I'm not sure that's the best angle for a game, and most hardcore sims' reliance on bespoke models may reinforce that idea.

I quite agree that the difference between issues such as wear (i.e. the rate being too high) and unrealistic (grip progression) need to be determined. However that doesn't mean that simply having a wide range of compounds available is a valid reason to not have a tyre model in place that does the job better than the one we have right now.

I want a model that can do all of that. But the problem is there are too many tyres in the world, you cannot recreate them all, especially not with empirical data - about which you've yourself gone into some detail. Thus, lots of lovely options are a must (in conjunction with a better, physically-based model to reliably reflect those options) - and better documentation of what tyres are supposed to be will help, too, so we're not arguing over what exactly Sports tyres are (failing that, a table of recommended tyres for each car would be good, like GT5).

These differing compounds may well only differ in terms of a grip multiplier, which negates the 'range = more difficult' argument, and even if that were not the case it doesn't get over the fact that T10 manage it with a range of profile and width options for each compound. They may have less compounds to model, but in terms of number of 'options' to model I would suggest that a lot more exist.

I meant the same range you described in FM4, not that GT5 has more compound choices. FM4 is a game like GT5, so is doubly screwed for the same reasons - which you've just described.

PD do a lot right and have done a lot right in the past, however right now GT5 is stuck with a tyre model that doesn't cover the bases that its reasonable to expect in this day and age.

Agreed, but I expect PD know this, so hopefully we can expect improvements in GT6.

 

[Denilson]

How about this?

I've uploaded a spreadsheet with the recommended tires to Google Docs:

GT5 Stock Tire Recommendations


*************

In an effort to figure out what is going on with GT5's tire models and which tires should go on which cars, I decided to do some skidpad testing. I used the 2010 Camaro SS and the Corvette ZR1, since I'm familiar with those cars and actual data is readily available.

We don't have an actual 200' skidpad to calculate lateral g force with in GT5, but what we do have is a g "meter" and a datalog. For the values I came up with, I created a delineated scale and taped it under the HUD g-force bar graph, and also used a scale against the datalog graph during replays as verification. The measurements were taken on the widest part of the TGTT, by turning a continuous steady-speed circle after warming the tires. Lateral g force was recorded up to the point where the car started to skid and could no longer hold the established circle. I also ran laps "on the edge" to verify the numbers, and repeated all the tests twice. (Note that I rounded the numbers to the nearest .05, due to my screen resolution).

My setup is a racing simulator chassis with a G25 wheel, and a Sony 50" HDTV. I ran each test with no aids and a manual tranny in "bumper" cam. (I hate that inaccurate view name ).

First up was the Camaro, with comfort hard (CH) tires. I performed the test on each tire type, trying to be as consistent as possible. I only tested comfort and sport tires; once I got to the racing compounds the grip started getting ridiculous, and was beyond what I wanted to test with this setup.

Here are the numbers (Notice that each softer tire compound increases lateral acceleration by approximately .05g):

CH - .85
CM - .90
CS - .95
SH - 1.00
SM - 1.10
SS - 1.15

The real-life Camaro SS scores a 0.87 on R&T's skidpad test. So it would appear that CM tires would be closest to stock for the Camaro, based on lateral acceleration. (I'm going to the next higher number, just because ).

Now for the 'vette numbers:

CH - .85
CM - .90
CS - .95
SH - 1.05
SM - 1.10
SS - 1.15

Virtually identical as far as the lateral acceleration numbers for each tire type. The real-life ZR1 scores a 1.10 on R&T's skidpad, so it would appear that SM tires would be the best stock equivalents for it.

Note: Just for reference, RH lateral g values were around 1.25, and RS were around 1.35 with the ZR1.

Here's where it starts getting weird. The real-life Camaro comes equipped with Pirelli P Zero tires, and the ZR1 comes with Michelin Pilot Sport PS2 ZP tires. According to TireRack.com data sheets, both of these tires have identical speed rating (186+mph), tread wear (220), and traction rating (AA).

The only thing I can come up with to explain the unexpected test numbers is that the size of the contact patch is not figured into GT5's tire equations. In other words, to duplicate accurate lateral acceleration numbers for the ZR1, you have to use a softer tire compound to make up for the larger contact patch on the real-life car. (The 'vette has considerably more rubber on the road than the Camaro, especially in the rear).

So far it looks like each car would have to be tested independently to come up with the best GT5 tire type to simulate real life. I plan on doing some more as time permits, but it will be a slow process. First up will probably be one of the Ferrari's that come with the Pirelli P Zero's (599 I think?), so that we have a side-by-side comparison.

Thoughts?

*************

For those that are late to the party and want a quick summary:

My testing so far has revealed that the 9 tire types (CH, CM, CS, SH, SM, SS, RH, RM, RS) in GT5 form what appears to be a simple grip multiplier, with each tire type adding approximately .06g of lateral grip. The only thing that changes is where the scale starts for various cars. (i.e. for the ZR1, CH = .85g and for the '71 Cuda, CH = .80g). It also appears that the width of the tire is not being considered in the grip equations; for any specific tire type, the '02 Mini Cooper has the same amount of lateral grip as the '09 Corvette ZR1! And as softer tires are equipped, the amount of grip increases equally for both cars.

The implications of this are that in order to get close to IRL grip performance (based on lateral acceleration anyway), you have to equip different cars with different tires. As an example, just throwing sport mediums on all performance sports cars means nothing. One car may need CM tires to reproduce IRL performance numbers, while a very similar car may require SH tires.

Here are my "recommendations" for the cars I've tested so far (take it for what it's worth and do with it what will you will ).

Edit: See link at top

 

[Cote Dazur]

How about this?

Very interesting.

 

[leeson65]

"the '02 Mini Cooper has the same amount of lateral grip as the '09 Corvette ZR1"
That just about sums it all up, doesn't it. Tires between all cars are essentially the same in GT5.
So, GT5 cars don't have accurately modeled tires, suspension, aerodynamics, or transmission gear ratios, as has all been proven pretty thoroughly in past threads. What did they get right?

 

[Raggadish]

"the '02 Mini Cooper has the same amount of lateral grip as the '09 Corvette ZR1"

I may be in the open blue for saying this but might this statement not be true? Can't a lighter car with smaller tire width compared to a heavy car with wider tyres have the same lateral grip?

What did they get right?

I guess the real life times to the times one usually sets in the game is pretty accurate.

 

[Scaff]

I may be in the open blue for saying this but might this statement not be true? Can't a lighter car with smaller tire width compared to a heavy car with wider tyres have the same lateral grip?

The two cars have totally different curb weights, static weight distribution, tyre width and profile, Polar Moment of Inertia, center of gravity, ride height, wheelbase, track width and suspension design (therefore roll centers).

Its therefore pretty much impossible that they would record the same lat-g on the same compound on the same radius turn.

Even if it was possible it doesn't then explain the numerous other cars in GT5 that are wildly different in all the areas listed above and have the same lat-g on the same compound on the same radius corner.

I guess the real life times to the times one usually sets in the game is pretty accurate.

Unfortunately lap times are not a good indicator of anything.

 

[TomBrady]

I think it's funny that people are bashing the tire model. What's especially strange is people saying that they have the progression from grip to no grip all wrong. Quite the contrary. That's one of the few things they do very well. The progression is one of the best things about GT5, because it's more spot on than even some PC sims. The only sims that do the progression from grip to no grip better than GT5 is LFS and netKar Pro. That's it. Now I'm not saying GT5 is on par with those 2 overall, but it doesn't really have to be because the progression is done so well. To me that progression is very important in making a sim enjoyable or not. I love iracing but for me the one reason why it will never top netkar pro (realism wise) is the progression from full grip to none is just far too steep of a curve. Has anyone wondered why you don't see people making drift videos in iracing? Well there you go. Rfactor & GTR 2 as well. The drop off is far too rapid and very unnatural. Drifting in GT5 is natural. That's what's so great about the tire model. It doesn't need built in aids to help the physics feel natural *cough* forza *cough*. It seems apparent to me you guys have lots of knowledge of how sims model the tires, but I question how you really can compare it to real life because no amount of knowledge on paper can match on track experience. For the most part, GT5 does a superb job replicating the feel of the cars it models, even with it's shortcomings. I know that from experience on real tracks. They may not get the job done in the best way, but they do get it done and it's certainly a huge step forward compared to past GT's, and other console arcade sims. Fact is, the room for improvement you all have pointed out only goes to show how much better things can get if they decide to go in the right direction with GT6

Of course I would agree in some ways the tires are terrible, racing softs are just unusable in my opinion but overall, the physics are great for the console. Keep in mind that most of the people who play GT5 use controllers. Compromises obviously had to be made. Considering that, they did a great job