The online racing simulator
iRacing
(13603 posts, closed, started )
Try iRacing with somewhere around 107% in the Profiler and 10 in-game, it really sounds like you have it turned up way too high. I have tried it turned up, and pretty much all you can feel is bumps, but when it's setup properly it feels fantastic.
107 and 6 - so no, sadly... I once was even at 117 and 3, to even out SAT and bumps a bit, but that didn't feel better either...
Quote from bbman :iRacing is bumps, bumps, then some more bumps, and way back in the distance there might be some tyre forces. It's like a whisper in a storm - if you concentrate hard enough, you might hear it...

completely agree
they might not be canned and actually based on track undulations but they certainly are way way too accentuated
you basically would get roughly the same ffb if you just added some noise
It feels spot on to me, I can feel exactly what the car is doing. Of course it's bumpy, we are all used to sims with glass roads.
Kind of agree with Deko (I know you're surprized).

I don't have to guess what the tires are doing in iRacing any more than I do in LFS, both sims are communicative. I never experienced "ripping the wheel from my hand" at LRP, either you're exaggerating beyond words or ... I don't know what else actually. (maybe you just need to work out a little :razz

The skippy is pretty lively for feedback but for example in the Miata or Solstice (since you brought up your own road car which probably has little in common with a Skip Barber) you barely feel those same undulations and bumps at LRP. It really depends on what car you're driving.


Quote from Shotglass : but they certainly are way way too accentuated

And you base this on what? I presume you have something to back that up since baseless statements are historically very un-Shotglass like

I'll snicker a little if LFS' Rockingham feels a bit like iRacing in the FFB department.

I should point out that I'm really 100% willing to be proven wrong on this, but so far it's only hand-waving.
Quote from Ball Bearing Turbo :And you base this on what? I presume you have something to back that up since baseless statements are historically very un-Shotglass like

ive driven cars on crappy roads and on good roads and thus far ive yet to see or drive across a bump that registers in the wheel without being clearly visible as a pothole or huge step from one patch of asphalt to the next
iracing on the other hand gives me a wheel that vibrates randomly and pointlessly while driving straight across a seemingly perfectly smooth bit of straight

also ive been round hockenheim granted not in the drivers seat but if you feel a bump in the wheel youll feel it in your spine as well
for one thing hockeheim was pretty darm smooth (except for the curbs) and even old well used non gp tracks like lime rock look like theyre at least as smooth as your average autobahn judging by pictures

actually my car recently did feel a little like the ones in iracing at around 110kmh... easily fixed by getting the wheels balanced
if i ever came across a car with a wheel feel like the ones in iracing id probably take it to a workshop immediately before the wheel falls off or whatever else causes those issues
I'm joining the group that say you feel bumps through FFB miles before you feel the tires.
The problem, it seems to me, is that in real life you get the bumps through your seat, whereas in iR you're getting them all through the steering column.

-LFSForums 2011 Cap'n Obvious Award Winner
Quote from DeadWolfBones :The problem, it seems to me, is that in real life you get the bumps through your seat, whereas in iR you're getting them all through the steering column.

-LFSForums 2011 Cap'n Obvious Award Winner

It's only January, someone will definately beat that.

And yes you feel it through the steering column, but that shouldn't silence the feel of the tires IMO.
It really doesn't cancel out the feel from the tyres, at all.
Forgive my ignorance here. But can someone explain why profiler:107% and game:10 is more realistic than say profiler:100% and game:20 ? Does clipping start happening after you go over 10 or something? What's the reason?

I've been using profiler:100% and game:20 because the iRacing devs suggested to use that in some of the game setup tips documents.
Quote from DeadWolfBones :The problem, it seems to me, is that in real life you get the bumps through your seat, whereas in iR you're getting them all through the steering column.

-LFSForums 2011 Cap'n Obvious Award Winner

so what youe saying is you think its using canned bump effects kinda like the right/left vibration in sync with the revs that isi games do which unless something is seriously wrong with our power steering would never happen on any real car?

Quote from JasonJ :Forgive my ignorance here. But can someone explain why profiler:107% and game:10 is more realistic than say profiler:100% and game:20 ? Does clipping start happening after you go over 10 or something? What's the reason?

logitech wheels all come with a nonlinear curve that reduces forces round the centre when set to 100
when you set them slightly above 100 you can reduce that effect and make them behave more linearly
trouble is thanks to no documentation whatsoever no one actually knows what you have to set it to exactly to get the proper result
Thanks

Switching up now
Quote from JasonJ :Forgive my ignorance here. But can someone explain why profiler:107% and game:10 is more realistic than say profiler:100% and game:20 ? Does clipping start happening after you go over 10 or something? What's the reason?

I've been using profiler:100% and game:20 because the iRacing devs suggested to use that in some of the game setup tips documents.

not sure if this helps but here are a couple of interesting posts by the person that wrote the reelfeel ffb plugin.

Quote :Force Feedback Part 1 - The Cold Hard Facts

There seems to be a lot of mystery regarding force feedback in racing sims, iRacing in particular. More worringly, there is a also lot of mis-information floating around, which feeds the mysteriousness and serves no use but to cause confusion and mis-understanding.

This post/article (whatever it ends up being) is my effort to put an end to this mysteriousness and confusion. I don't know if it will work, I may end up just causing more confusion. My apologies in advance, if that happens.

Let's start at what we could call the beginning - the physics engine. Physics engines understand numbers, nothing else. Just like money is the root of all evil*, numbers are the root of all physics engines.

Most of these numbers represent forces of some kind - forces generated at the tyres, the mass & inertia of the vehicle, aero forces acting upon the vehicle etc.

One of these forces is at the heart of any good force feedback system - the force occurring at the steering wheel.

Given that the physics engine is constantly calculating the relevant force it should should be no problem to send it straight out to the force feedback device, right?

Well, for me, this is where things start to get interesting.

Most racing sims use DirectX DirectInput to generate their force feedback output. In particular, they use a method known as DirectInput ConstantForce. As the name suggests, ConstantForce provides a method for the developer to send a constant force to the force feedback output device.

Sounds perfect, but there's one stumbling block. And it's a big one.

The DirectInput ConstantForce method doesn't use any particular unit of force.

The ConstantForce method only understands numbers from -10000 through zero to +10000, in effect having it's own unique units for measuring force. -10000 means 'turn as hard as you can in one direction', +10000 means 'turn as hard as you can in the other direction', zero means no force in any direction.

The reason for this is quite simple - the actual force that will be output by the device is wholly dependant on how much torque it's motor(s) can produce (or are configured to produce).

Chances are that the physics engine will be calculating the torque that should be occurring at the steering wheel in Newton metres (Nm), or a unit that can be directly converted to Nm.

It is down to the developer to convert this Nm torque value into the range -10000 to +10000, so that it can be output via the ConstantForce method.

When I said above that this is where things get interesting, this is precisely where I mean - the conversion between a distinct Nm torque value and DirectInput's +/- 10000 allowed range.

For the purpose of clarity, let's assume we have a force feedback wheel that can produce 10Nm of torque.

Let's also assume that under normal driving conditions our simulated car produces a maximum of 10Nm of torque at the wheel.

In this case it's nice and easy, there's not really any conversion to be done, we can simply multiply the calculated force by 1000 and it will fall within the range that DirectInput ConstantForce will accept.

This means at 10Nm steering wheel force we'll send a ConstantForce of 10000, at 5Nm we'll send 5000 etc. Given that our wheel can produce 10Nm we'll end up with what I would call the holy grail of force feedback, where the physics engine's calculated torque value is being mapped to our force feedback wheel with total accuracy and fidelity. We will feel exact same forces on the wheel as the physics engine is calculating.

So, where does it all go wrong? Why all the confusion?

Firstly, not many of us have wheels capable of producing a realistic amount of torque. The Logitech G25, for example, tops out between 2Nm and 3Nm. My road car with skinny eco tyres and power steering produces more torque at the steering wheel than this!

Secondly, different cars produce different amounts of torque at the steering wheel. Just changing the car set up can have an effect on steering forces. Caster in particular has a huge effect as it introduces jacking forces. On a car with lots of caster as the steering wheel is turned the car is effectively being jacked up with the steering, manifesting as a return to centre force on the steering wheel. The further the wheel is turned the more the mass of the car attempts to return the wheels to the straight ahead position, the stronger this force gets.

To cater for both these situations the developer has to build in some way for the user to control the conversion from Nm to ConstantForce values.

Usually, as in iRacing, this takes the form of a slider. The RealFeel plugin for rFactor is slightly different in that it let's you set the amount of torque that corresponds to a ConstantForce value of 10000, but the theory is the same.

Which ever way it is accomplished, what the user is actually altering is the 1000 multiplier value that we used in the example above to scale the forces into the range accepted by the DirectInput ConstantForce method.

If this multiplier is set too high the result will be clipped output forces and a loss of fidelity. The sim might be sending values way above 10000 but DirectInput will interpret these as a maximum of 10000.

What does this mean in practice? Well, imagine we're in a steady state corner, well below the peak slip angle of the tyres and the physics engine is calculating a torque at the steering wheel of around 15Nm. With the 1000 multiplier we used above the sim would be sending a force value of 15000 to DirectInput. DirectInput will interpret this as 10000 and your force feedback wheel will be putting out as much force as it can. Now let's say the corner tightens slightly and the car starts to understeer. The force at the steering wheel will drop as the tyres pass their peak slip angle**, the physics engine calculates there is now only 10Nm of torque at the steering wheel. That 10Nm of torque results in the sim sending a ConstantForce of 10000.

In this situation the physics engine would calculate a 33% drop in force at the steering wheel but the force at the force feedback wheel would not change at all, it would remain at a ConstantForce of 10000, i.e. maximum.

If this multiplier is set too low the result will be the opposite, the force feedback wheel will never reach it's maximum torque output.

If you've read this far, thank you, I appreciate it. There will be more to come, specifically regarding Logitech wheel settings and their effect on the relationship between the forces calculated by the phyics engine and the forces output at the wheel.

Any questions, please, fire away.

(Tech)Ade.


* Not neccessarily my personal belief, just a useful metaphor
** Assuming suspension geometry that doesn't have too much mechanical trail which could drown out any useful pneumatic trail from the tyres




==================================================================================================================


Force Feedback Part 2 - Logitech Muddy The Waters

Ok, so hopefully you've read Force Feedback Part 1 - The Cold Hard Facts and now understand the quest for the holy grail of force feedback, where the physics engine's calculated torque value is being mapped to our force feedback wheel with total accuracy and fidelity. We will feel exact same forces on the wheel as the physics engine is calculating.

This thread is about the settings available in the Logitech Settings screen and how they impact the quest for this holy grail.

First up, Combined axis. This has no effect whatsoever on force feedback. It should only be checked if you're using a sim that doesn't support the throttle and brake being on seperate axes. Using combined axis will mean that you can't use the throttle and brake at the same time as they will cancel either out.

Next up, degrees of rotation. Again, no direct impact on force feedback. For iRacing this is best left at 900 degrees, the sim will then give you 1:1 linear steering for each car.
Next we have a checkbox to enable force feedback and three sliders, Overall Effects Strength, Spring Effect Strength and Damper Effect Strength.

As long as the Enable Force Feedback is checked the three sliders below come into effect. I don't think I need to explain what will happen if Enable Force Feedback isn't checked, but, just in case - you'll get no force feedback!

On to the Overall Effects Strength slider. This defines how the Logitech driver scales the ConstantForce effects coming from the sim.

At 100% the output force at the wheel is mapped 1:1 to the input force (apart from a slight non-linearity built into the firmware in order to prevent unwanted oscillations around the centre).

With this slider at 50% an input force of 10000 (the maximum) will only drive the FFB motors at 50%. Unless this slider is at 100% or over the wheel will never drive it's motors at 100% and you'll never feel the maximum torque your wheel is capable of.

The force at the wheel will be the same in both the following situations:

a) ConstantForce of 5000 from the sim, Overall Effects Strength of 100% = motors working at 50%
b) ConstantForce of 10000 from the sim, Overall Effects Strength of 50% = motors working at 50%

As I mentioned in the Part 1 thread, the G25 is very weak and slow in comparison to a real steering wheel.

This means that if the sim is sending a ConstantForce of 5000, at 100% Overall Effects strength, we're only looking at a force of about 1-1.5Nm.

At a constant force of 2500, around 0.5-0.75Nm, we'll be lucky if the motor generates enough torque to even turn the wheel (when there's nobody holding it!).

In practice this means that on a smooth surface when the wheels are pointing straight ahead, i.e. when the forces on the simulated steering rack are relatively low, we'll be lucky to feel much at all on the G25.

Turn the wheel a bit or hit a bump and forces will build up, especially in a heavy car with lots of caster, or a car with lots of downforce (i.e. tyre load), but this weak feeling around the centre is a big put-off for most people as it just doesn't feel 'realistic'.

This leads us nicely on to what happens when we move the Overall Effects Strength slider up over 100%.

The further towards the maximum of 150% we go with this slider the more the Logitech driver amplifies the lower input forces. Going over 100% makes no difference to the force on the wheel when the sim is driving it at 100% (i.e. sending a ConstantForce of 10000) - 100% is 100%, the motors can't be driven any harder.

Going up to 150% has a massive effect on the lower magnitude input forces.

For example:

a) ConstantForce of 2500 from the sim, Overall Effects Strength at 150% = motors working at 100%
b) ConstantForce of 5000 from the sim, Overall Effects Strength at 150% = motors working at 100%
c) ConstantForce of 10000 from the sim, Overall Effects Strength at 150% = motors working at 100%

As you can see, with the Overall Effects Strength up at 150% it doesn't really matter what magnitude of ConstantForce the sim is sending, the output at the wheel will always be at it's maximum.

Think back to Part 1 and what happened in an understeer situation when the force at the wheel was staying at 100%. The same thing is happening here. As far as the force feedback wheel is concerned the forces haven't dropped off at all, whereas the physics engine could have calculated that they've dropped by 50% or more.

With this in mind, I find 101% or 102% Overall Effects Strength is enough to improve the straight ahead feel somewhat without losing too much fidelity.

Spring Effect Strength and Damper Effect Strength

I need to do more research on the effect of these sliders and how they interact with the DirectInput ConstantForce method. The theory says that these sliders should have no effect unless the developer issues a Spring or Damper command via DirectInput. In practice, I'm not so sure - I'm almost positive I've tried 100% damper previously and it does what it says on the tin, dampens the forces coming from the sim, however in a quick test the other day I could feel no difference at all between 0% and 150% Damper. As I said, more research most definitely needed here.

For now I'll just say that I recommend setting both at zero. They are going to do nothing but distort the ConstantForce output from the sim and take us further away from the holy grail.
Centering Spring

Centering spring is designed only to be used in games that do not provide their own centering forces, it has no place being enabled in iRacing or any sim that uses ConstantForce. It will counteract the forces from the physics engine, muddying the waters and hiding that holy grail out of sight forever.

Allow Game To Adjust Settings

Irrelevant as, to my knowledge, no sim yet makes use of this functionality. It is meant for developers to be able to control the wheel rotation settings from in-sim.
And there we go, the end of Part 2.

Again, any questions or confusion, ask away.

(Tech)Ade.

Added later:

I forgot to mention something about the Centering Spring... and why I have it checked but on zero.

Unchecked the centering spring is in effect during normal operation of Windows, even if the slider is at zero. Once the sim initialises force feedback via DirectInput if centering spring is unchecked then it will not have any effect, even though it does in Windows. With it checked and the slider to some positive value it will have an effect in-sim, and counteract the forces the physics engine is generating.

Excellent Deal : 1 Year sub + Williams + Zandy + Silverstone + basic cont for just $75 ! Looks like they're pushing for new members.

I wish that offer was available when i signed up.
Quote from anttt69 :Excellent Deal : 1 Year sub + Williams + Zandy + Silverstone + basic cont for just $75 ! Looks like they're pushing for new members.

I wish that offer was available when i signed up.



(Guess who did).

So close to 4 on road now, 3.8, and oval's already D and 3.43.

FFB advice is nice, but I don't really know to what I should set. Maybe the problems lies there, that I always have to re-plug my G25 when joining sessions in rF, LFS and iR...
Quote from Swiss-Spirit :[snip]

FFB advice is nice, but I don't really know to what I should set. Maybe the problems lies there, that I always have to re-plug my G25 when joining sessions in rF, LFS and iR...

I find the FFB settings vary from car to car, annoying cause there is no separate car customization.
@107% profiler and in game:
Road cars (SS,Leg,mx5,sols,jett,stang) : 18
Single seaters (SM) : 10
Single seaters (SRF,FW31) : 15
FordV8 (FG01) : 5
Oval cars : not owned

I don't know for sure (not driven any of these cars), but these feel kinda right* to me. Some suggestions would be appreciated. 103% feels better I think.

*which is still difficult to judge because the G25 wheel is almost half the size of a real wheel so leverage is completely different


BTW: Thanks StableX
Quote from JasonJ :I find the FFB settings vary from car to car, annoying cause there is no separate car customization.

Yes there is.. look again!
Ah, I thought that option was only for button/key customizations. Didn't realize it applied to FFB. Thanks.

I just nailed my first full grid of mustangs - No Aids, Cloud 9
No idea why exactly these are made, but some of you might like them.

iRacing iPhone and Android Apps:

iPhone

Android
F1 season looks really good, I'm going to concentrate on that and the Nascar class A series this season, should be a good one with tons of participation because of the Pro qualifying. I quite fancy running the Indy car aswell, especially on the ovals, with probably the Radical and C6R on the road aswell to give me a break from the F1 car. I quite fancy running a bit of Grand-Am aswell, but i'm going to skint myself with all the content I need to buy.

New build is being rolled out at lunchtime GMT tomorrow, should be interesting.
I have a 1 month code if anyone wants it (New members only)
This thread is closed

iRacing
(13603 posts, closed, started )
FGED GREDG RDFGDR GSFDG