Physics thread

  • Thread starter LVracerGT
  • 1,459 comments
  • 129,184 views
Well, I was young and had no money but a broken joystick. So I guess the idea came up. My Dad has a friend who has driven go karts with Michael Schuhmacher in the same league in his young ages as he started his career. He had a steering wheel for his PC at home and I was in love with it. But he never had let me played with it when I remember correctly. Wheels were expensive in the 90´s. In that time the wheels had no FFB. Or at least his wheel had no. It was just a rim and two springs. One on the right and one of the left. They were centering the wheel. That was all. But it was gooood :D

He was playing Formula One Grand Prix from Geoff Crammond and it was a blast. That started my sim racing career ^^

"Formula One Grand Prix from Geoff Crammond"

I think I remotely remember this one. Right up there along with Grand Prix Legends as one of the best F1 sims at the time.

👍

Now that we're on topic, my love for sims started with a game called Revs on the Micro Computer that my dad got for me when I must have been 5 or 6 at the time! Ah, good times.
 
"Formula One Grand Prix from Geoff Crammond"

I think I remotely remember this one. Right up there along with Grand Prix Legends as one of the best F1 sims at the time.

👍

Now that we're on topic, my love for sims started with a game called Revs on the Micro Computer that my dad got for me when I must have been 5 or 6 at the time! Ah, good times.

Yap good good times. Even though we got wayyy off topic, its nice to talk about retro games :D Sometimes I watch them on YouTube to remember that times.

One of the first 3D racing games I played was stunts. Played it at a friends house. It was the ****. And you were even able to make your own tracks. And it had better sound than Gran Turismo 1-6 :D

 
About the pCars physics: Overall, the handling has improved a lot for all cars. But, in general, the open wheelers handle quite better than the others. Some cars need careful throttle control even at full traction aid. The colision needs to improve a lot yet. The cars glue too much on each other or bounce too much. However, it's already quite better than Test Drive Ferrari Racing Legends.

Cheers.
 
Well, tell us how the physics are compared to FM5?
The car handling is quite similar to FM5, which is nice to me. But many cars are harder to control at pCars, mainly the most powered, even at full aids. The karts are quite funny to drive, but the superkarts are much sensitive to abrupt commands. Since I use xbox 360 controller for windows, it is a problem to me, so I have to be much careful with superkarts. Formula A also spins if you give it much throttle and you have to watch for the curbs in some tracks, like Mugello.
 
Yap good good times. Even though we got wayyy off topic, its nice to talk about retro games :D Sometimes I watch them on YouTube to remember that times.

One of the first 3D racing games I played was stunts. Played it at a friends house. It was the ****. And you were even able to make your own tracks. And it had better sound than Gran Turismo 1-6 :D


What's really sad/amazing is that a game in 1990 had modeled interiors while GT6, a late-2013 game has silhouettes for many cars.
 
Guys, I tried one of latest builds & I was not able to get my wheel config properly. I dunno why but I felt a LOT of grip riding up Megane (Carrer mode). I also notice I'm slow as hell racing against AI @spa GT3 (free mode).
Last months, I've been racing rFactor DTM, Enduracers & Blancpain series mods a lot in online server communities with great results in endurance races.So pCARS feedback got me lost.
I know the only way to get faster is understanding car behavior and reaction, physics itself, get used to the sim-mod, but pCARS feedback is driving me nuts.
 
If you need help with your wheel config & stuff, WMD is the place to ask, not here. There were lots of changes to that during the last 2 weeks.
 
After I started with iRacing seriously 1 week ago, I now understand whats the difference between direct FFB and canned effects like in AC. First I was hating the iRacing FFB because I had almost no curb and road effects etc. But now, after a while, I love it. It is so direct and you can feel the road and the car perfectly.

And I remember that Pcars had similair FFB like iRacing but still was missing some details. So I hope SMS is going the iRacing FFB way and do not implement these canned effects?
 
After I started with iRacing seriously 1 week ago, I now understand whats the difference between direct FFB and canned effects like in AC. First I was hating the iRacing FFB because I had almost no curb and road effects etc. But now, after a while, I love it. It is so direct and you can feel the road and the car perfectly.

And I remember that Pcars had similair FFB like iRacing but still was missing some details. So I hope SMS is going the iRacing FFB way and do not implement these canned effects?

Yep, direct FFB, no canned effects.
Can one or both of you explain what this means please? I like the feel of AC personally, best FFB I've ever had in a game IMO, but my experience is limited. What parts of AC are "canned". What exactly is canned? Wouldn't it imply that if a game has any tuning options at all in FFB, it can't be 100% direct FFB, that increasing certain things is in fact "canning"?
 
Can one or both of you explain what this means please? I like the feel of AC personally, best FFB I've ever had in a game IMO, but my experience is limited. What parts of AC are "canned". What exactly is canned? Wouldn't it imply that if a game has any tuning options at all in FFB, it can't be 100% direct FFB, that increasing certain things is in fact "canning"?

Because of the physics update frequency you won't get all the road details from direct FFB, for example at 200 km/h you make 55.5 m/s that at 600hz physics rate (pCars one) makes about one physics tick every 10cm. This will make two things happen: if you have a step between two sampling points, the game will just react as if it was a slope (faster you're going, lesser will be the impact on FFB due to the slope being longer). The second problem is that due to Nyquist sampling theorem (http://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem) you can only exactly recreate signals that are 1/2 of the sampling frequency, so everything above 300hz will be distorted/not reproducted depending on how they filter it. the canned effects go mostly for things like curbs, flatspots (not needed in pCars), wheel locking and things not 100% accurately represented in physics model but you still want to get in FFB.

Another reason to add canned effects are the G-forces that in reality you'd feel on the body (so-called "Seat Of Pants" or SoP effects) that some games mix in the FFB for a better feel of under/over steer. Those are 100% configurable in pCars and i think are present in AC.
 
Because of the physics update frequency you won't get all the road details from direct FFB, for example at 200 km/h you make 55.5 m/s that at 600hz physics rate (pCars one) makes about one physics tick every 10cm. This will make two things happen: if you have a step between two sampling points, the game will just react as if it was a slope (faster you're going, lesser will be the impact on FFB due to the slope being longer). The second problem is that due to Nyquist sampling theorem (http://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem) you can only exactly recreate signals that are 1/2 of the sampling frequency, so everything above 300hz will be distorted/not reproducted depending on how they filter it. the canned effects go mostly for things like curbs, flatspots (not needed in pCars), wheel locking and things not 100% accurately represented in physics model but you still want to get in FFB.

Another reason to add canned effects are the G-forces that in reality you'd feel on the body (so-called "Seat Of Pants" or SoP effects) that some games mix in the FFB for a better feel of under/over steer. Those are 100% configurable in pCars and i think are present in AC.
I wonder if there is any conflict between the 300/600hz sampling weight and the various wheels and peripherals? Are some wheels more capable of reproducing the FFB more accurately than others? So @FuriousDemon says no canned effects only direct FFB and you say that yes, you can dial them in if you want to. The two positions aren't necessarily mutually exclusive but some more clarity would be helpful.
 
I wonder if there is any conflict between the 300/600hz sampling weight and the various wheels and peripherals? Are some wheels more capable of reproducing the FFB more accurately than others? So @FuriousDemon says no canned effects only direct FFB and you say that yes, you can dial them in if you want to. The two positions aren't necessarily mutually exclusive but some more clarity would be helpful.

The strength of the effects and forces acting on the wheel is configurable, but these additional effects that are not going to be felt through the real vehicle's steering wheel (canned effects) are not there.
 
I wonder if there is any conflict between the 300/600hz sampling weight and the various wheels and peripherals? Are some wheels more capable of reproducing the FFB more accurately than others? So @FuriousDemon says no canned effects only direct FFB and you say that yes, you can dial them in if you want to. The two positions aren't necessarily mutually exclusive but some more clarity would be helpful.
To add to Furious Demon's answer, there are some optional synthesized effects available, such as "Seat of Pants", "Gut" and "Arm", and also the more traditional canned effects, such as center spring, rumble strip, gear change and so on. I don't know if all of that will be in the final game.

As for the physics tick rate vs steering wheel's update rates, that could make a difference. For wheels with slower update rates there will be less detail than there is "in the numbers". I don't know which wheels work at which rates (LogiForce?), but I suppose most work at a couple of hundred Hz or more, so it shouldn't be a big problem.
 
Last edited:
The strength of the effects and forces acting on the wheel is configurable, but these additional effects that are not going to be felt through the real vehicle's steering wheel (canned effects) are not there.

I played the 83x build by a friend. The wheel was pretty dead on straights but the wheel in the game was shaking on bumps. So that was really strange. We played with an Fanatec GT2 and tried FFB @80 and tire force from 30 - 100. But the wheel was totally dead on straights and in corners there was only the centering effect (no matter what car). Almost like in iRacing with the MX5. But in iRacing the FFB depends on the car. An open wheeler has way more road effects than an racing car. I hope the newer builds have good FFB already?!
 
To add to Furious Demon's answer, there are some optional synthesized effects available, such as "Seat of Pants", "Gut" and "Arm", and also the more traditional canned effects, such as center spring, rumble strip, gear change and so on. I don't know if all of that will be in the final game.

As for the physics tick rate vs steering wheel's update rates, that could make a difference. For wheels with slower update rates there will be less detail than there is "in the numbers". I don't know which wheels work at which rates (LogiForce?), but I suppose most work at a couple of hundred Hz or more, so it shouldn't be a big problem.

Logitech = 500hz
Thrustmaster = 500hz
Fanatec = 500hz
SimSteering = 500hz or 1000hz (not exactly sure which)
AccuForce = 2x1000hz or "Dual-Stream 2000hz" as they call it



The problem isn't so much the sampling rate, but how the FFB in Project CARS works. The FFB system isn't the issue, but the fact that you get everything through the steering rack at real world levels means you also need a FFB motor that is able to produce real world torque levels.
After all, the stronger the output is at a set volume level the louder small details will become as well. So the stronger the motor with a fixed strength level, the more kerb detail will become appearant as well.

As far as canned effects goes. This is why I hate old fashioned games like Assetto Corsa that still use canned effects that 'add in extra information'. Because now that I have a strong wheel (CSWv2) all these canned/fake effects just cause me to loose control over my car, as the FFB motor's effect execution has direct influence upon the position sensor, which actually steers the car in game. So every effect influences car control and stability (weight transfer).

I don't mind them being in a game, as long as I can switch them on or off. Which I can't in most, if not all, games out there. :boggled:


:yuck::yuck::yuck::yuck: Canned Effects :yuck::yuck::yuck::yuck:
 
Logitech = 500hz
Thrustmaster = 500hz
Fanatec = 500hz
SimSteering = 500hz or 1000hz (not exactly sure which)
AccuForce = 2x1000hz or "Dual-Stream 2000hz" as they call it



The problem isn't so much the sampling rate, but how the FFB in Project CARS works. The FFB system isn't the issue, but the fact that you get everything through the steering rack at real world levels means you also need a FFB motor that is able to produce real world torque levels.
After all, the stronger the output is at a set volume level the louder small details will become as well. So the stronger the motor with a fixed strength level, the more kerb detail will become appearant as well.

As far as canned effects goes. This is why I hate old fashioned games like Assetto Corsa that still use canned effects that 'add in extra information'. Because now that I have a strong wheel (CSWv2) all these canned/fake effects just cause me to loose control over my car, as the FFB motor's effect execution has direct influence upon the position sensor, which actually steers the car in game. So every effect influences car control and stability (weight transfer).

I don't mind them being in a game, as long as I can switch them on or off. Which I can't in most, if not all, games out there. :boggled:


:yuck::yuck::yuck::yuck: Canned Effects :yuck::yuck::yuck::yuck:

Are the canned effects in AC still there when you set them to zero? Curbs, slip and road effect.
 
As far as canned effects goes. This is why I hate old fashioned games like Assetto Corsa that still use canned effects that 'add in extra information'. Because now that I have a strong wheel (CSWv2) all these canned/fake effects just cause me to loose control over my car, as the FFB motor's effect execution has direct influence upon the position sensor, which actually steers the car in game. So every effect influences car control and stability (weight transfer).

I don't mind them being in a game, as long as I can switch them on or off. Which I can't in most, if not all, games out there. :boggled:

I just read an article with one of the top guys working on Assetto Corsa who stated quite clearly that they do not use canned effects, but all effects are generated by the physics engine and that and only that is what is felt at the wheel.
 
Are the canned effects in AC still there when you set them to zero? Curbs, slip and road effect.

There is no slider for the weight transfer effect. So that is still present no matter what.

I just read an article with one of the top guys working on Assetto Corsa who stated quite clearly that they do not use canned effects, but all effects are generated by the physics engine and that and only that is what is felt at the wheel.


Amazing that they say that when they use a weight transfer (also called Seat of Pants by some) effect in their game.
So stating something at some point in time and what is currently executed by the game can still differ. Everything is open to change after all.
I have to say that I have not touched a wheel in weeks though, so my info about AC might be out of date already.

That said I do not wish to get into an AC fanboy debate. Its a waste of time. Everyone should simply play what entertains them the most dispite what everyone else says. Who cares what someone thinks about something, the point is that the game player should have fun. ;)

The point should be that independent from the game, a fake effect will always influence the steering input negatively. The stronger your FFB wheel gets the harder it is to control these fake effects, and thus they will cause a bigger excursion of the wheel as it gets harder to keep the wheel dead still. Bigger excursion, means bigger direction change of the front wheels and bigger effect upon the cars weight transfer.

It is totally undesirable for powerful wheels. With a CSWv2 you can more easily read what the car does from how you feel the peak slip angle move through the wheel. That combined with the visuals give you enough information to tell you where your front wheels point to and what the rear of your car is doing.

All in all, I just hope game developers will realize that there is a feedback loop when a force feedback wheel is used.
Every FFB effect has direct influence upon the Physics if this is not front wheel positional data calculated by the Physics. This means your own wheel input has influence, but also fake/canned effects inserted post Physics calculation.

Feedback loop: Physics > FFB Motor > Position Sensor in Wheel > Physics > Etc...
 
Last edited:
(ffb wheel update rates)
Thanks Logi, I knew I could count on you there (I bet you can recite FFB wheel specs is your sleep!) :)

As you say, motor strength, and also the specific of the mechanicals and the inertia of the rotating parts, are factors that affect the FFB detail levels, and more so than a little bit higher or lower update rates (500Hz is more than enough).

Although, I like the idea of a pure physics/spindle derived FFB as a base, some optional effects can help with immersion and understanding of the car's behavior, even if it's not strictly realistic. It has to be optional of course.
 
Last edited:
Thanks Logi, I knew I could count on you there (I bet you can recite FFB wheel specs is your sleep!) :)

As you say, motor strength, and also the specific of the mechanicals and the momentum of the rotating parts, are factors that affect the FFB detail levels, and more so than a little bit higher or lower update rates (500Hz is more than enough).

Although, I like the idea of a pure physics/spindle derived FFB as a base, some optional effects can help with immersion and understanding of the car's behavior, even if it's not strictly realistic. It has to be optional of course.

Well, it isn't that bad yet. Getting close though. ;)

Optional effects do not help with a powerful setup in my opinion, but I sure liked them when I still had a Logitech wheel. Which are too weak to produce useful real effects to tell you how the weight is transfering and all that good stuff.
So yeah, I am for it but it should be given an on/off switch. Than again we all were nagging about having 'Options' with everything in Project CARS. ;)

:bowdown: Options :bowdown: <-- holy for simracers :P
 
sorry for the OT, you can answer me on the right thread (https://www.gtplanet.net/forum/threads/ps4-t300-gte-or-t500rs.319404/page-2) but since i'm going to buy a new wheel for christmas and i'm still not sure which to take between T300 and T500 (i'd pay the T500 just about 30 more euros) i wanted to hear your opinion on them.. As for now i'm probably picking the T500 because of the pedals and stronger motors, reading what you say it seems that actual motor strenght is quite important, but still i see most people suggesting the T300 over the T500 because of the more detailed FFB.
Another thing that made me want to buy T500 was how naturally you can drift letting the wheel go, something impossible to do with my g25.
-see from 2:05 for g25, 3:53 T500

I wonder if T300 is capable of this too
 
It is totally undesirable for powerful wheels. With a CSWv2 you can more easily read what the car does from how you feel the peak slip angle move through the wheel. That combined with the visuals give you enough information to tell you where your front wheels point to and what the rear of your car is doing.

👍

Will get my V2 next week I guess. Can´t wait. My actual wheel is the Porsche GT2.
 
sorry for the OT, you can answer me on the right thread (https://www.gtplanet.net/forum/threads/ps4-t300-gte-or-t500rs.319404/page-2) but since i'm going to buy a new wheel for christmas and i'm still not sure which to take between T300 and T500 (i'd pay the T500 just about 30 more euros) i wanted to hear your opinion on them.. As for now i'm probably picking the T500 because of the pedals and stronger motors, reading what you say it seems that actual motor strenght is quite important, but still i see most people suggesting the T300 over the T500 because of the more detailed FFB.
Another thing that made me want to buy T500 was how naturally you can drift letting the wheel go, something impossible to do with my g25.
-see from 2:05 for g25, 3:53 T500

I wonder if T300 is capable of this too


Good FFB on the hardware side needs 3 main things. The triple S.

1. Strength
2. Speed
3. Smoothness

1. Strength is needed to be able to give the right resistance to the wheel as well as being able to reproduce all details, and produce them at (close to) realistic levels of strength.
2. Speed is necessary so that your steering wheel will be able to accurately follow direction changes (of the front wheels!) determined by the game's Physics engine. This is especially necessary for drifting, because if your wheel lags behind what the car/Physics in-game wants at any given time you will have a high chance on spinning out.
3. Smoothness is necessary to be able to feel micro-details. 'Feeling the pebble on the road' that kind of micro-detail. If your wheel is grainy/cogging/feeling like a handsaw than it will overshadow any micro-details that were send by the game, as the FFB motor is too rough to be able to reproduce them.

The T500 is a fast wheel, but it is a very rough wheel too. I really noticed this when I went back from the CSWv2 to the T500 for testing and comparing in GT6 for a moment. I really thought I had a saw in my hands and I was cutting some wood. If you don't know any better than it is not something you'll notice, but if you do... :(
This is especially important when racing on the limit cause the peak of the Mu Slip Curve is hard to notice, and with the T500 there is no way you can really notice it. You'll only be able to feel the rough estimate.

So the T500 only has 2 downsides, the cogging and the noisy fan. The latter can be fixed by swapping in a Noctua fan (a post of mine should be in the T500 thread somewhere on this).

The T300 is a device I never had my hands on. So I can't really comment. But although it is less powerful, it does seem to be smoother from what I heard. So you should feel more detail and it should be better for normal racing.
Also going by what Darin Ganghi said about the differences, is that the T500 has more dampening and thus feels heavier. But the T300 he said was faster and smoother, which is what you'd want for drifting. You want the wheel to be able to keep up with the front wheels. It just feels less strong according to him, but you don't really need weight for drifting. Or at least it is not as important as the wheel needs to 'flow' in unison with the Physics/virtual car.


So yeah, it really depends on what you like to do. Both have a down side, one is cogging and the other doesn't have a proper set of pedals.
Personally I think I would pick the T300 with a set of pedals. That way I would at least be ready if I ever would want to get a Playstation 4. Besides I would gamble on the T300 being a bit nicer for drifting.
 
There is no slider for the weight transfer effect. So that is still present no matter what.

Amazing that they say that when they use a weight transfer (also called Seat of Pants by some) effect in their game.
So stating something at some point in time and what is currently executed by the game can still differ. Everything is open to change after all.
I know what weight transfer is obviously, but what is the "weight transfer effect" in AC and how is it different in Project Cars? Is there a slider for it in PCars?
 
Back