Bad tire physics could actually be a weight and inertia issue

[Dmarc_Atl]

I noticed something odd when I was doing doughnuts in the 370 too, the weight does sort of feel centered now that I think about.Hopefully they fix what ever it is because it felt ( im assuming since I never did a doughnut in real life ) more natural in gt5.

 

[sk8er913]

I noticed something odd when I was doing doughnuts in the 370 too, the weight does sort of feel centered now that I think about.Hopefully they fix what ever it is because it felt ( im assuming since I never did a doughnut in real life ) more natural in gt5.

No, its actually slightly worse in GT5 as even more things are left out... and I heard some of PD check the threads every once in a while, maybe they will read this and it will give them some ideas to improve the simulation of the game.

When I tried the demo I thought it was the GT5 tire model.

I think it actually is the improved tire physics engine, but poor algorithms in other areas don't let the tires reach their potential.

 

[I'm Motarded]

Are you guys playing with a controller?

The front wheels dragging the car around happens even with PC sims with a controller unless you reach the point of the yaw-angle where there's no return. Using a wheel means physically, there's literally no way you can spin it so fast as to do that, unless you can set an extremely low degrees of rotation (like, under 100). You can, and do in some cases, have the maximum steering angle within milliseconds with a controller, I don't know of a real world automotive application where that would be possible, as steering is still mechanical, and not digital.

Having full lock at the flick of a switch would play havoc with physics as we know it, makes sense that it messes with the physics in games too. This is the reason why the right thumbstick and D-pad steering has assists, but it will never be perfect because of the limitations those control methods have, and the way people use them.

 

[APXCaramel]

The problem I have is that when drifting, the car just snaps around, like the grip level is exactly the same, even though the setup is tuned so that it doesn't have too much front grip but way more rear grip. I am on the gas and the car doesn't keep sliding. I have been drifting on PC sims for a while now and they all feel about right, but GT just feels weird. A lot of GT physics feel weird. I think the only time that I've really felt like the physics were good was when I drove the HSV. Most every other car feels pretty weird, comparing to rFactor, iRacing and a little bit of Race07. They just don't feel like cars.

 

[Wolfe]

@I'm Motarded: I don't think that's the problem. In Live for Speed, my PC sim benchmark, even 1-to-1 axis control (or mouse steering) isn't enough to "tug" the front end around like Gran Turismo, Forza, and other console games allow you to do. It only takes a handful of side-to-side drifts down a straightaway to lose control -- caused by the "point of no return" you mentioned -- and the "shimmying" occurs at a slower rate in the first place, communicating the car's angular momentum.

Theoretically, you could modify the electric-assist in a modern car to snap from lock to lock very quickly at the push of a button (with a beefy motor); there's no reason such input should "play havoc with physics as we know it." It doesn't perturb the simulation in Live for Speed, anyhow -- the resultant behavior is extreme, but it still looks natural within the context of having crazy fast steering. If Gran Turismo or Forza were truly top-notch simulators, that's what I would expect them to do with silly joystick inputs.

The input filters on GT/Forza don't even allow you to steer as quickly as you can in LFS, so there's no excuse.

 

[Zer0]

@I'm Motarded: I don't think that's the problem. In Live for Speed, my PC sim benchmark, even 1-to-1 axis control (or mouse steering) isn't enough to "tug" the front end around like Gran Turismo, Forza, and other console games allow you to do. It only takes a handful of side-to-side drifts down a straightaway to lose control -- caused by the "point of no return" you mentioned -- and the "shimmying" occurs at a slower rate in the first place, communicating the car's angular momentum.

Theoretically, you could modify the electric-assist in a modern car to snap from lock to lock very quickly at the push of a button (with a beefy motor); there's no reason such input should "play havoc with physics as we know it." It doesn't perturb the simulation in Live for Speed, anyhow -- the resultant behavior is extreme, but it still looks natural within the context of having crazy fast steering. If Gran Turismo or Forza were truly top-notch simulators, that's what I would expect them to do with silly joystick inputs.

The input filters on GT/Forza don't even allow you to steer as quickly as you can in LFS, so there's no excuse.

Sure? I don't see any input buffer in the GTA, check the front tyres:




"It only takes a handful of side-to-side drifts down a straightaway to lose control -- caused by the "point of no return"

^That is what happens in the GTA if you try with an steering wheel instead of a controller.

 

[sk8er913]

Sure? I don't see any input buffer in the GTA, check the front tyres:




"It only takes a handful of side-to-side drifts down a straightaway to lose control -- caused by the "point of no return"

^That is what happens in the GTA if you try with an steering wheel instead of a controller.

They way the car is spinning looks unnatural in that video, at 6 seconds the car should have been unable to snap back around the other direction, I think, but if it could snap back around it is surely going in the wrong direction after that, it completely lost it's forward momentum. And I do use a DS3.

 

[Stephanos82]

Here is Forza 3's version of spinning... see how much faster the car spins and the front tries are almost useless? Unlike GT grip. I think if GT adds rear inertia though this will be fixed... which will lead to more realistic throttle stomping, spinning, sliding... which may lower the amount of throttle stompers.

The way the car slides in Forza seems perfectly fine to me!

My mx5

My mx5 in Achna Raceway track

 

[Zer0]

They way the car is spinning looks unnatural in that video, at 6 seconds the car should have been unable to snap back around the other direction, I think, but if it could snap back around it is surely going in the wrong direction after that, it completely lost it's forward momentum. And I do use a DS3.

 

[Scaff]

Sure? I don't see any input buffer in the GTA, check the front tyres:

Both GT and Forza have quite clear input buffers with a controller, and its possible to use it in both to utterly fool the physics engine.

Just get a car doing donuts with full lock and then let go of the control stick, the steering of the car (and tyres) will centre but the car will still be doing donuts.

The buffer effectively 'ignores' the rapid return to centre of the steering from an 'engine' point of view but returns it from a visual point of view.

BTW - Not quite sure what you are attempting to say with your video above?

 

[sk8er913]

The way the car slides in Forza seems perfectly fine to me!

I was saying Forza's model is more accurate than GT5s and GTA 2013s.

 

[Wolfe]

Sure? I don't see any input buffer in the GTA, check the front tyres...

Sorry, I forgot that GT5 (and now GTA) actually shows 1-to-1 axis steering visually. But if the game was really giving gamepad users unfiltered steering, the game would be utterly unplayable without precise, nuanced joystick control. The majority of players would full-lock understeer straight into the gravel pit of the first corner they tried. No console game I'm aware of has ever defaulted to true 1-to-1 axis steering, for this reason.

That is what happens in the GTA if you try with an steering wheel instead of a controller.

Not being able to keep up with the steering wheel doesn't prove or disprove anything physics-wise. Regardless of how you play the game, super-fast steering should not nullify angular momentum. It's a sign of something fundamentally wrong that affects the entire game, not only silly drifts. The silly drifts just happen to be a straightforward test that's easy for anyone to try.

 

[dbarrade]

@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?

And this whole comment about GT not having rear inertia is ridiculous, this is what a game without inertia at the rear looks like:



and here is your far more realistic Forza:



Note the angle of the steering wheel in the cockpit view and the way that the car moves in Forza.

GTA 2013 improves on GT5, if you can't feel that, then you need a better Steering Wheel set up.

First thing I notice is that they have to counter steer much more than is required in Gt Academy.

With the wheel, the amount of countersteer required feels right to me, when I watch videos back, you find that you are keeping the front wheels in that spot where it is close to the direction of travel in order to maintain grip on them, too much either side and you lose grip and are sliding all four wheels, which sounds like what you are experiencing.

And just in case there is some doubt as to whether I've spun a car before, this is me spinning my car just to compare



And if you are wondering why those two Exiges pass me so quickly on the straight, I only have about 217HP at the wheels, they are close to 400BHP and 300BHP respectively.

 

[Ridox2JZGTE]

Try to correct a slide driving a high power car like my Wangan Midnight replicas on comfort soft, losing rear at just over 100kmh exiting a corner or over 300kmh on high speed curve at SSR7 will be virtually unsalvageable even with stick controller.

 

[sk8er913]

@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?

And this whole comment about GT not having rear inertia is ridiculous, this is what a game without inertia at the rear looks like:

and here is your far more realistic Forza:

Note the angle of the steering wheel in the cockpit view and the way that the car moves in Forza.

GTA 2013 improves on GT5, if you can't feel that, then you need a better Steering Wheel set up.



With the wheel, the amount of countersteer required feels right to me, when I watch videos back, you find that you are keeping the front wheels in that spot where it is close to the direction of travel in order to maintain grip on them, too much either side and you lose grip and are sliding all four wheels, which sounds like what you are experiencing.

And just in case there is some doubt as to whether I've spun a car before, this is me spinning my car just to compare


And if you are wondering why those two Exiges pass me so quickly on the straight, I only have about 217HP at the wheels, they are close to 400BHP and 300BHP respectively.

I have not played Forza 4, I was referring to Forza 3, which I posted a video of, and I don't use any driving aids except for ABS: 1 90% of the time, also I never posted a video of me driving.

 

[Ridox2JZGTE]

If you know how, you can change each car max steering angle in GT5, default max angle on every car in GT5 is 40 degrees - too extreme IMO. I usually use between high 20 degrees to mid 30 degrees. On most cars, 30 is enough. Drifting or countersteering with appropriate steering angle of the real car used on in GT5 will be more realistic and harder to recover.

On my latest S15 Silvia Time Attack build, I use 30 degrees, this is more than enough and for drift tune - real drift car S15 uses up to between 62 to 70 degrees, and this is possible in GT5 now

With such a huge angle, you can do manji drift at higher speed in real life, drift zig zag like mad

 

[LeMansAid]

@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?

Does it really matter? If the topic is intertia, input method and aids are irrelevant. The video in question does look odd to me, but there also is a depth factor. It is difficult to gauge how far forward the car moves at that 6 second point. It's a bit like when a documentary films hungry lions and the crazy zebras that are seemingly standing right next to them. I scream "ruuuun!!", but the view is deceptive. Then again, just as the lions end up getting the zebra anyway, GT cars may well be a little anorexic.

 

[ReverenceSatans]

and here is your far more realistic Forza:

Oh look, a replay from FM4.

Anyone who has actually played it (i.e not you) would know data compression for replays completely destroys steering and car movement accuracy i.e they look almost nothing like what was actually happening while driving. They would also know the steering angle animation (i.e what makes the wheels look like they are turning) is limited and doesn't actually show correct or maximum angles.

 

[Raitziger]

Steering animation sucks in Forza and it sucks in GT. It should be 1:1.

 

[Rotorist]

To just drop an idea, what you think about the tire physics problem to be connected to the factor that dramatically changes the tire-to-surface contact between wheels in a straight angle and when starting to steer or be turned already when gas or brake or weight applied or not?

This variable is critical for the whole car balance during the turns, exiting the turns, accelerating from stance, in-corner braking and all this changes the pressures in the same time depending to how much the tire struggles.

Very complicated and only a cooperation with a tire manufacturer like Yokohama could give the data to put this together and reach a sim level. We all hope that until GT6 launches some things will be even better. Signs of improvement already exist in GTA2013.

 

[dbarrade]

Does it really matter? If the topic is intertia, input method and aids are irrelevant.

Of course input method and aids are relevant, I believe the game compensates vastly for DualShock Input as it feels like a different game playing with that compared to the wheel.

Aids like Skid Recovery obviously break the laws of physics in order to make the game easier, this probably includes affecting tyre physics and therefore weight distribution and most likely inertia.

I've played Forza 4 a couple of times with a controller, and with the Microsoft Wheel, I didn't realise that the replays were that inaccurate that they don't show how the car really moves... I guess that rules them out as a learning tool, also that would make me wonder if Forza 5 will do the same thing, destroying any chance of the Drivetar of mimicking a players driving.

 

[gtracedriver1]

Hmm ... who knows for sure ??? From my observation GT looks very accurate like real life but as someone said hand animation needs improvement

 

[NFSCARBON1]

Of course input method and aids are relevant, I believe the game compensates vastly for DualShock Input as it feels like a different game playing with that compared to the wheel.

.

It does. With the DS the faster you go the smaller the maximum steering angle becomes (so you don't put full lock on at say, 100mph and crash). Also, when the car moves in an unusual way (e.g a spin) you can almost apply full lock to correct it. However for the most part you can put the lock thats available on quicker than with a wheel.

 

[Greyout]

lol thats got to be like 20 degrees of caster

 

[Ridox2JZGTE]

lol thats got to be like 20 degrees of caster

We can have the wheel angle ( via hex ) but not the caster yet on GT5, the car does counter steer differently than with default locked 40 degrees, I need stick sensitivity high at more than 5 - I use 7 to be able to counter steer faster, makes for cool manji drift. This is with 70 degrees wheel angle on S15 Silvia drift tune 570HP SR20 TT

BTW, if I lower the max wheel angle to mid 20s or low 30 degrees like most real car have, it becomes very hard to counter steer and recover big oversteer even with high steering sensitivity ( on comfort tire at least ) - very hard to drift consistently using stick. The default 40 seems to help for easier correction.

 

[ZoidFile]

And this whole comment about GT not having rear inertia is ridiculous, this is what a game without inertia at the rear looks like:

VIDEO

Lol, that Ridge Racer video made my day. Especially making a 360° left spin into a right hand corner.^^

But its not so funny how much Ridge Racer reminds me of seasonal events.

 

[LeMansAid]

Of course input method and aids are relevant, I believe the game compensates vastly for DualShock Input as it feels like a different game playing with that compared to the wheel.

Aids like Skid Recovery obviously break the laws of physics in order to make the game easier, this probably includes affecting tyre physics and therefore weight distribution and most likely inertia.

There are either inertia issues or not. If SRF breaks the laws of physics, that is an indictment on the developers, but it does nothing to show that inertia issues are not present with SRF disabled.

Input method is irrelevant when talking about inertia. If a robot was programmed to steer a real car at a turning rate that no human would be capable of, the car would still behave realistically. RC has been wired up to real cars, the input method changes but the car knows no difference, it's still cause and effect. Again, there are either intertia issues or not. What the car is "told" matters naught, what is does in response is what determines accurate physics or not.

 

[I'm Motarded]

There are either inertia issues or not. If SRF breaks the laws of physics, that is an indictment on the developers, but it does nothing to show that inertia issues are not present with SRF disabled.

Input method is irrelevant when talking about inertia. If a robot was programmed to steer a real car at a turning rate that no human would be capable of, the car would still behave realistically. RC has been wired up to real cars, the input method changes but the car knows no difference, it's still cause and effect.

The car would look to be behaving differently if the steering ratio was 100x faster than human steering input. If the RC motors are designed to act as a human would, which most instances are, we see human behavior in return. And the opposite is also valid there. All we can do is observe, because almost literally that's the only tool of measure. The fact that the SRF can break the laws of physics, means you can't dismiss any reactions [of the car] to any other assists the developers have also put into the game.

Again, there are either intertia issues or not. What the car is "told" matters naught, what is does in response is what determines accurate physics or not.

Every action has a reaction. How can you measure a response without looking at the input? Without input nothing happens.

 

[LeMansAid]

Every action has a reaction. How can you measure a response without looking at the input? Without input nothing happens.

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.

 

[Ridox2JZGTE]

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