Good work, but still major problems with tyre physics

[RIP2004]

Hi there.
First of all thanks for the patch and great work. Definitly a step towards the right direction (realism).
Love how the F1 car is done, eventhough I don't like OW as much as "real" cars *g*

I love street cars. The small ones that are driven on demo servers. And there I can compare to the real world, because I am driving mine sometimes to the limit. But I didn't drive a F1 car yet.

The problem with tyres which still remains is the longitudinal behaviour as written in your progress report.
The way tyres behave under acceleration and braking. Its simply still wrong.

The FWD car on street tyres (GTI) accelerates best when pushing the petal to the metal. With hard spinning wheels it still manages a 7.13 and ace in training. You don't need any feeling whereas IRL you have to be careful with the throttle in a FWD car, if you wanna go off fast.
Same goes for braking. Locked wheels may be damaged while braking, but they still provide the shortest braking way, which is again wrong.

And because everything is calculated the same way, the ingame physics way, cornering behaviour can't be true this way, because longitudinal behaviour is also important for cornering. The correct grip behaviour of spinning or locked wheels is important for everything.

The RWD cars are better now, but still FZ50 without TC or RA are much more dangerous and difficult to drive than comparable cars in RL (like Porsche e.g.)

I also noticed, that it is very hard to get your car drifting by suddenly take away throttle. You can go fast in a corner and than full throttle, no throttle, full throttle. The car won't begin to drift because of the transfer of weight. If it really starts to drift, then because its a RWD car and got to much power full throttle on the rear wheels and make them spin. Not because weight was transfered away from the back. So its very hard to do that with a fwd car.

(EDIT: this may be the result of setup and is perhabs no fault of wrong physics)

In my oppinion rFactor still is more believable in the result of simulating street cars. No matter how they do it and even if they just readout data and don't calculate anything.

I really hope that tyre physics isn't supposed to be finished for a while. I hope there will be further improvements in the next patch regarding spinning and locked longitudinal behaviour.
This would also make starts and braking much more interessting. At the moment starts with street cars are boring because of this problem and because there is no free start at all.
And braking wrong with locked wheels isn't punished as hard as it should be (in RL you would slide off course ...)

Have fun with the new patch all
RIP

[mkinnov8]

OMG! :x

[deggis]

OMG! :x
What? RIP2004 has very good points.

[Tweaker]

Yeah has a few good points... especially the FWD full acceleration stuff.

And also the fact that lift oversteer is not very easy to perform in most of the RWD cars. Understeer counteracts this.

[Shotglass]

ive thrashed the new physics a bit in the other thread but some of your points are just not valid

Same goes for braking. Locked wheels may be damaged while braking, but they still provide the shortest braking way, which is again wrong.

i just tried at the car park with a gti and my results were that the distance is more or less the same with locking causing slightly longer braking distances which is correct

The RWD cars are better now, but still FZ50 without TC or RA are much more dangerous and difficult to drive than comparable cars in RL (like Porsche e.g.)

have you ever driven one of the old porsches without esp and with rather simplistic dampers ? from what ive heard theyre widowmakers

I also noticed, that it is very hard to get your car drifting by suddenly take away throttle. You can go fast in a corner and than full throttle, no throttle, full throttle. The car won't begin to drift because of the transfer of weight. If it really starts to drift, then because its a RWD car and got to much power full throttle on the rear wheels and make them spin. Not because weight was transfered away from the back. So its very hard to do that with a fwd car.

must be your setups ... whendrifting ive mostly used lift off in q and i still do in s ... and the fwd also get some nice lift off oversteer

edit:

about full acell on fwds ... havent checked with perf view or anything yet but judging by the sound the tyres spin a lot less than back in q when taking off full throttle and if im not mistaken a tiny amount of slip will give you the largest longitudinal forces

[Ball Bearing Turbo]

about full acell on fwds ... havent checked with perf view or anything yet but judging by the sound the tyres spin a lot less than back in q when taking off full throttle and if im not mistaken a tiny amount of slip gives you the largest longitudinal forces

Agreed.

In a car like the GTi I doubt dropping the clutch would make much difference to accel times. Only higher powered cars can lose enough traction for long enough to seriously affect acceleration. It's not just the fact that the wheels are spinning, more like just how great the differential is between road speed and wheel speed.

[RIP2004]

You don't need perf view, to see how hard they are spinning.

Just watch the revs. They are about 8000 from 0 to 40 or 50 kph. This is a HUGE difference between speed and engine revs and therefore the wheels are spinning badly.
And it makes a difference, even in very low powered cars. Ever accelerated a FWD car in rain? See what happens if you have to much throttle ... you will move hardly ... :shrug:

about oversteer problem :
As I said it could be the setup. But in rFactor for example it works with every car with the default setup. And I am not the only one which has a problem to use this kind of oversteer. See Tweaker. He described it perfectly. There is often understeer then ...


I am not really sure about the braking but in my first tests there was no real difference between locked wheels and a good braking method. But I will test it again ...

Yeah perhabs the RWD cars should behave like very old cars. But the problems in longitudinal behaviour make the behaviour in general wrong. The great thing about LFS is that all the physics act together. And if anything is wrong it will affect everything ... so I am not sure if it would be easier without that problem. But it would be different.

[Shotglass]

And it makes a difference, even in very low powered cars. Ever accelerated a FWD car in rain? See what happens if you have to much throttle ... you will move hardly ... :shrug:

theres huge difference between how a tyre acts on a wet road and on a dry one ... on a dry one the optimum slip is somewhere around 0.3 afaik on a wet road its much much lower

As I said it could be the setup. But in rFactor for example it works with every car with the default setup. And I am not the only one which has a problem to use this kind of oversteer. See Tweaker. He described it perfectly. There is often understeer then ...

most of the race_s setups use high coast lock => little to no lift off oversteer

[RIP2004]

Maybe with rain. But its also on the dry like this. My car has enough power on the dry (150 bhp on FWD) and it is anything but not a good start with hard spinning wheels. A very short spin is good, but not this bad spinning like with the GTI in LFS.

The problem remains. A full throttle, hard spinning start is the perfect method to get away ... but only in LFS. Even GT4 on the playstation has better physics regarding this issue.

[Ball Bearing Turbo]

I must test this for myself.

One thought I had is that it would be very hard without an analogue clutch (which I thankfully possess) to truly launch a car the way one would in real life. If you don't, the only alternative is to floor the thing with no clutch slip which would not lend itself to rapid acceleration unless your gearing is setup perfectly for low end acceleration, which would lead to wheel spin once the revs get up a bit. So in the GTi I could still see dropping the clutch at 8000RPM (gear still determines wheel speed, and with low power, the tires probably are not spinning as wildly as you think even with the engine at 8000RPM) being faster than bogging the engine. A proper start with higher revs and some clutch slip and some minor wheel spin should be the best result, I agree.

[RIP2004]

I didn't really look at the clutch. But I think I didn't see any blue graphic left. So no clutch. And 8000 revs or so are about 40 or 50 kph, where you have to shift in second to continue acceleration.
Therefore wheels are spinning at lets say 40 kph all the time. So in the beginning there is a difference of about 40 kph, which gets smaller and smaller.
Sure, a proper start is difficult without a real clutch. I just want to see that full throttle isn't the best thing. And I don't want boring starts anymore.

Ever tried a start with these cars in rFactor? There are big differences between people starting there and it is great fun. I want this also in LFS (remove this arcade style signal and give us a free start also plz)

I have to test the thing with the diffs. Could be the reason for the problem with oversteer ...

[Ball Bearing Turbo]

(remove this arcade style signal and give us a free start also plz)

Alright, THAT I totally agree with!:thumb:

[RIP2004]

:thumb:

[deggis]

I agree with the starting thing. Can't be that hard thing to code. I kinda hoped it would've been changed for this patch... but still didn't. :shrug:

[Vykos69]

And it makes a difference, even in very low powered cars. Ever accelerated a FWD car in rain? See what happens if you have to much throttle ... you will move hardly ... :shrug:

Hmmm, where is the mindtrick in that thought? :scratchch

[Ball Bearing Turbo]

Rofl

[RIP2004]

There is no mindtrick.

Two seperate things.

1. Even with low powered cars you will achieve a massive wheelspin in the dry if you rev to the limiter and just take the foot of the clutch. It will last not as long as in a high powered car, but it should be enough for the first few kph and meters.

2. Everyone with every kind of car (low or high power and FWD) can experience the effects in the wet. Because usually noone tries that, but in the wet it happens not only voluntarily ...
It is a bit different and stronger in rain, but you'll see effect. Same in dry but not as easy to see usually ...

[Greek]

yes, more realism.

[Vykos69]

There is no mindtrick.

Two seperate things.

1. Even with low powered cars you will achieve a massive wheelspin in the dry if you rev to the limiter and just take the foot of the clutch. It will last not as long as in a high powered car, but it should be enough for the first few kph and meters.

2. Everyone with every kind of car (low or high power and FWD) can experience the effects in the wet. Because usually noone tries that, but in the wet it happens not only voluntarily ...
It is a bit different and stronger in rain, but you'll see effect. Same in dry but not as easy to see usually ...
well, in dry you loose grip WAY harder and regain grip WAY faster. I am not saying, actual tire-phys are right, but you base your accusations on pure speculations, you havent even tested yourself. SImple thing: Rain != Try => Not a real comparison is possible.

[vladimir]

The problem remains. A full throttle, hard spinning start is the perfect method to get away ... but only in LFS. Even GT4 on the playstation has better physics regarding this issue.gt4's tyres suck totally...you have virtually no grip when accelerating, which leads to acceleration times being way off from the real counterparts, but then way too much lateral grip where its almost impossible to get most cars to drift.
LFS is way more realistic there.

maybe you should try your theories with the FXO, the gti has not ernough power to find a proper conclusion here.

oh and you have hardly ernough feel in pc-wheels' brakes to really modulate the brake...but i have tested this extensively with the measure tool on autocross and locking the wheels will not procide you with the shortest braking distance! but its so hard to really brake properly on a pc simulation that you are not able to really see a difference.

[RIP2004]

Tell my summer tyres, that I did it in the wet. I have to change them this month. Ever smelled ruber? Would be a hardcore car to make the tyres smoke and smell in the wet :really:

[vladimir]

Tell my summer tyres, that I did it in the wet. I have to change them this month. Ever smelled ruber? Would be a hardcore car to make the tyres smoke and smell in the wet :really:jesus christ! would you please stop comparing the performance of your car in the WET with LFS???:pillepall

[RIP2004]

jesus christ : learn to read and understand. Then you would notice, that I tested it in the dry ... for real. And not only on the playstation or on pc. The only problem is, that it is bit more expensive for real. But they had to be changed anyway this summer ...

[Ball Bearing Turbo]

LOL...

Anyhow, I think a better test would be to setup the FX0 with some closer to road car gearing, then perform the test with the FX0 - a redline drop, versus a controlled partial clutch launch versus a dead stop "keyboard style" start without wheel spin. It should show the second type the fastest, the first test in the middle, and the last launch the slowest. And I seriously think it will. GMeter would be best to use. I know it has accel times, and can't remember if it has 60foot times etc...

Perhaps I'll try this later tonight if I feel the urge.

[RIP2004]

The car you choose won't matter.

Just watch the online races and the starts. Everyone with the same setup is exactly as fast as you are. All start with full throttle.
Doesn't matter if it is FZ50, FXO or GTI ...

But try it. I wanna see data ... :nod:

[michaelnyden]

just got back from trying the new patch out...and I know I am going to get flamed for this...but it seems worse in some respects...it seems that they tweeked the FF instructions to the steering wheels and now I get a lot less info from my FF logitech momo than I used to...now it just feels flat in it's effects compared to the last patch/version, in addition, it is way to easy to drive the cars at the limit now with very, very progressive breakaway, almost to a rediculous point, cars are not twitchy at the limit you aren't walking on "egg shells" like in real life at the limit...it's too bad I almost like the last patch better and how it feels, I do like the new fonts used though and the new F1 car...but other than that, the game feels less like a sim and more like an arcade game now...

*let the flaming begin..lol*

[joeynuggetz]

I still get that annoying "making a left hander around a corner and the car decides to lose total grip and spin to the left instead" Where I expect some cars to push, all wheels lose traction and I spin. Shouldn't I at least start to push first?

[bal00]

have you ever driven one of the old porsches without esp and with rather simplistic dampers ? from what ive heard theyre widowmakers

That's because older 911's used semi trailing arms in the rear. Since the bushings aren't perfectly rigid, you get additional toe out if you brake or lift off in a vehicle with semi trailing arms. So if the rear gets loose and the driver lifts off, the outside rear wheel will actually steer into the drift and make the car go even more sideways. Newer 911's and the FZ shouldn't have these problems.

[michaelnyden]

also, it feels as if I have some driving aids on....but I don't see any options to turn them off...I noticed the throttle help and steering help have been removed from the options menu, only brake help is there as an option? does that mean they are on be default in this new patch? if so, how can I turn them off...the cars are way to easy to drive now...and grip like crazy...

[CombatWombat]

I always hated the LFS RWD physics.

Now they feel damn good.

Corner speeds seem about the same, but the transient behavior is much improved. They dont "grip like crazy" nor does it "feel like an arcade game". Remember that increased difficulty != realism. They only feel easier because the old behavior was so random. So. uh. drive faster or something?

[BWX232]

I also noticed, that it is very hard to get your car drifting by suddenly take away throttle. You can go fast in a corner and than full throttle, no throttle, full throttle. The car won't begin to drift because of the transfer of weight.

Yup- after a lot of testing today, that is right.. At first I thought everything was great - (it is better).. but after some more testing I found the same problems as before.

If the back end gets loose in a RWD car, in LFS, the way to stop from spinning is instantly letting off the gas. Also, in a corner in a normally set up RWD car in LFS, when you let off the throttle, there is no change in YAW (pic below). Some FWD cars can be set up to have this behavior, but the RWD cars don't act right still.. You should be able to transfer weight to the front and loose traction in the rear while going around a corner in most all RWD card.. sadly, this doesn't happen yet in LFS on a consistent basis. I think LFS is much better now though, and closer to being perfect.. and much more fun to drive.

It is funny to me to hear people here bash rFactor and call it arcade and whatnot, when some aspects of rF are actually more accurate than LFS.. rF's implementation of the physics engine is not as good, and the FF you get is no where near is good.. but in some ways it is still more accurate. Strange.


Also RIP2004, maybe you aren't talkign about incorrect tire physics in LFS, but incorrect vehicle dynamics when it comes to weight transfer and inertia. Maybe if those were calcualted correctly, the tires would behave pefectly fine.. Hard to say.

http://img73.imageshack.us/img73/1236/ayaw4rr.gif
A yaw motion is a side to side movement of the nose of the aircraft as shown in the animation.
http://www.grc.nasa.gov/WWW/K-12/airplane/yaw.html





edit-

LIFT THROTTLE OVERSTEER

A handling characteristic that causes the rear tires to lose some of their cornering grip when the throttle is released during hard cornering.
http://autorepair.about.com/library/glossary/bldef-199.htm




edit-
Lift-off oversteer (also known as trailing-throttle oversteer, lift-throttle oversteer, or drop-throttle oversteer) is a form of oversteer in an automobile that occurs when the vehicle's weight shifts from the rear to the front too quickly due to throttle release while cornering. Various suspension enhancements can be done to limit a vehicle's tendency to oversteer in this situation, including using a Weissach axle or a multi-link suspension.
http://en.wikipedia.org/wiki/Trailing-throttle_oversteer

[Michael Denham]

I have not played the new patch enough yet to have a really solid opinion, but I think one of the big things here is car setups. You can change the feel of a car immensely just by making half a degree change in camber at each end. Some of the default setups still seem pretty weird to me, not sure what anyone else thinks... Also, we don't really know about the road tires. Are road normal like regular tires or are they fairly sporty? Are Supers fairly sporty, or are they almost track day tires? I bet the XRT on street legal track day tires would feel a bit underpowered on the track with all that grip... There are so many variables at work.

Definitely feels better than it did before, a lot more controllable. Not 100% yet there perhaps, but it feels pretty good to me! I think it's being made harder for Scawen by the fact he's trying to do things as realistically possible. Now instead of maybe 3 or 4 variables in tire construction, he's working with say 7 or 8. I'm completely making this stuff up, but I hope you get my general point. Just that it may be harder to get things feeling right with a more complex model than with a simpler, more easily 'tuneable' model (ie GPL). I think to get it feeling right with such a complicated model, you need a huge amount of knowledge that the devs are still acquiring. Or maybe it's already pretty fantastic, and I just need to spend a bit more time working on setups and getting used to it all :)

One point about lift off oversteer...again I think it depends a lot on the settings. But I've found with default settings if I turn in hard and lift off I will usually get the back out in pretty much any car. If I'm in a steady state in a corner and lift, it will tighten the line. Maybe not get the back out, but again, I think it depends quite a bit on the setups.

[Blowtus]

I agree, car setup seems to be a factor influencing a lot of this discussion.

I also agree with the point that a fast spinning wheel doesn't lose traction enough. Sort of like the slip / traction curve levels out at some early point, when it should still be going sharply down.

The rwds exhibit far more normal cornering behaviour when set with a fairly neutral setup to me now - throttle on understeer, have to get a bit vigorous to shake the rear end loose. Previously I was unable to make the xrt drive at all like my rx7, my sets now feel pretty close. Tyres search for traction when slipping, rather than just floating around. All very good. Still doesn't feel like enough of a bite when correcting a slide, but part of this 'may' be to do with the lack of real forces being applied to my body :)

[BWX232]

One point about lift off oversteer...again I think it depends a lot on the settings. But I've found with default settings if I turn in hard and lift off I will usually get the back out in pretty much any car. If I'm in a steady state in a corner and lift, it will tighten the line. Maybe not get the back out, but again, I think it depends quite a bit on the setups.

Yeah- you are correct about that.. I mean.. nothing is set in stone, not even handling on real cars.


Also, to defend the lack of "lift off oversteer" in LFS in some cars and some setups.. and interesting note:


For most of us, lift throttle oversteer isn't a problem. First of all, it won't bite you unless you're driving at or near the limit. At reasonable street speeds, all cars have sufficent traction reserve to make the corner . Second, not all cars suffer from a lift throttle oversteer. Most rear wheel drive cars do not, particularly with an open or limited slip differential. Lifting during a corner merely helps the car tuck in tighter, though it does cost you time. Cars that suffer from lift throttle oversteer tend to have a very loose handling, and are often mid or rear engined. Famous examples of cars with significant lift throttle problems are the Porsche 911 and the second generation Toyota MR-2.
The first part is funny too--
http://everything2.com/index.pl?node_id=1258083
But maybe that guy doesn't know what he is talking about? LOL..



Anyway- All in all this patch is great.. I feel like I am driving a car, and I have better control of it, especially in the GTR cars.. You can push them much harder, just like you would think you would be able to, and you can feel the car take a nice solid set in a turn. MUCH more balanced. I'll keep testing before I make any final conclusions, all I know is that before, the "physics problems" made me not want to play LFS very much and frustrated me.. now I have an absolute blast, and can really feel the car underneath me- it doesn't feel "arcade-ish" at all either.. it feels very authentic most of the time.. :thumb:

[Michael Denham]

:thumbsup: I think another problem is that we are all 'testing' the new patch for 'problems' or 'issues' instead of just getting out there and driving and having fun! I know I am, and it really bugs me that I do it. I just want to drive and enjoy the cars. :)

[michaelnyden]

now I feel less connected or there in the car if that makes any sense..the cornering speed seem much higher now since they are much easier to drive and they do have more grip I honestly believe...I don't think the old behavior was random it was just that they gave the street tires very little grip under acceleration...but the physics at the limit were more realistic as anyone who has been to a road coarse or autoX will tell you how fragile a car is at the limit and oftentimes twitchy...it makes you almost hold your breath sometimes! in this patch, I feel nothing of the sort...no feeling of walking on "eggshells"

[Blowtus]

higher cornering speed? I was comfortably lapping blgp in the 1:24's in xrt before, stuggled to get under 1:26 during brief testing last night... felt like I was down a few kph each corner.

[GT Touring]

BIg improvement really, the double apex last corner in KY GP, anyone try it?
there is where you notice the tyres doing what they should- that said... KY oval is still flat and 0 degree wings for the OW cars...the drafting is much better- but it is harder to cacth it, as the effect seems less so.

[Yiots]

Ever tried a start with these cars in rFactor?
Let's not compare RF to LFS please... RF is NOT a simulator in any kind of way.

[Gabkicks]

that doesnt sound biased at all:tilt:

[deggis]

Let's not compare RF to LFS please... RF is NOT a simulator in any kind of way.
Too obvious troll comment to be taken seriously.

[KTy]

Some interesting points by RIP2004...

I just want to add my 2cents saying that I am also quite frustrated for not being able to start faster with an analog clutch, without making spining the wheels, compared to flooring the pedal to the metal and making spin the wheels... I dunno if I'm doing wrong or anyhting, but I never couldn't start faster "the real way"...:shrug:

Anyway, overall an excellent patch :nod:

[Sternendaal]

Let's not compare RF to LFS please... RF is NOT a simulator in any kind of way.

Have to agree with you.

[BWX232]

You guys should quit bashing rF, it is a sim... it may not have quite as "fine" of a feel as LFS, but it does have it's high points. You need to spend a lot of time tweaking that the same as LFS.. They both have their place in the sim racing world.

There are parts of LFS that put rF to shame, but there are also parts of rF that put LFS to shame.. it is all subjective.

[Krane]

The FWD car on street tyres (GTI) accelerates best when pushing the petal to the metal. With hard spinning wheels it still manages a 7.13 and ace in training. You don't need any feeling whereas IRL you have to be careful with the throttle in a FWD car, if you wanna go off fast.Actually careful acceleration produces best results (according to my testing) but the difference I think is too small, 0.02s between full throttle acceleration and a proper one on the 100m acceleration test. See attachments.

[Tweaker]

Also try the LX6 acceleration in like a 50 meter dash autocross layout. Full throttle is much slower. But yes, like you said Krane, it is very minimal, and it doesn't make a difference really if you try and start optimally, or full throttle.

Kind of silly to teach people through the training to do full throttle starts if you ask me :zombie:

[Niels Heusinkveld]

I just find it hard to believe that people call the tyres such a big improvement! It does seem that the slicks have improved 'more'... But then thats a risky statement as except one they all have considerable downforce. A bit more long. grip makes it harder to spin the tyres, and once at speed you have the magic downforce hand pushing them on the tarmac..

Here is a collection of replays where, most of the time, I don't countersteer.. There is just SOOO little required, both on and off throttle to make the car oversteer. And even when you don't truly 'spin out' the car still wants to turn so eagerly.

I don't run 'oversteery' sets, on a whole my front wheelrates are a bit stiffer than the weight thats on that side of the car and the front rollbar is considerably stiffer than the rear. Diff locking is non insane.

Out of the box, LFS still is wrong. Fixing 'normal' front tyres helps.. but we're masking a real problem. I don't even know if the tyres are the real cause, perhaps yaw forces are 4 times too high or inertia calculations are wrong.. I have no clue. But its too obviously not realistic imo.

[Blowtus]

I don't know if anyone ever really anticipated we'd be able to run perfectly normal setups and have it behave as expected... but imho, being able to run a slightly less than normal setup, and have it actually behave a lot closer to 'normal' (as opposed to before, where 'normal' behaviour just didn't happen) is a huge improvement.

[Tweaker]

Niels, but wouldn't you say that P/Q physics were a lot worse? Just go try that version if you can, right now, because overall it was worse in the old version.

As of now, SOME things have been improved, but there is something about the lateral grip that I cannot comprehend either. Longitudinal grip is great, maybe could use a bit of fine tuning, but overall good. It is just that the longitudinal grip is so powerful that the lateral grip is hardly affected at times. Then once it IS affected, the strong longitudinal grip is lost completely until you "flick/countersteer" the wheel forwards to recover. I also think recovery is ok, but the loss of grip is uninformative with tests like you showed in your replays.

Low speed grip is just the problem in my opinion. Your replays are demonstrating the low speed grip, and that is the main issue here. High speed lateral grip seems ok to me, you undergo some heavy loads, and it is forceful on the car. I've taken some cars at 100 mph in corners very fast, and at the limit, the strength and loads on the tires in huge... feels just like we have now at high speeds for the road cars. But like I said, low-speed grip is still a case here... only thing that bugs me to be honest (because I autocross, and a slalom is a joke atm)

[Niels Heusinkveld]

Yes there is an improvement over the previous version. However, over the years LFS has gone from 'some really huge problems' to 'some pretty huge problems'...

I don't know and I'm annoyed Scawen isn't joining discussions.. But all seems to indicate that the actual tyre MODEL hasn't really changed! Only the parameters that feed it have gotten a bit better. In many ways exactly the same things happen as before, only now they happen a little bit less 'easy' but that is only under certain circumstances. Straight line stuff improved, but as soon as there are lateral forces (and no downforce to mask the problem) its the same old issue.

I don't think you can say its 'high speed' or 'low speed'. This has often been mentioned in sims, and its true that most have different low speed physics.. but that is near 0 speed. Above 10km/h most sims will use their pacejkas or whatever model they run. Problems might be more apparant at certain speeds. At low speeds the car can change direction a bit faster, perhaps highlighting some issues with that. At high speed the car carries a lot of momentum which might cause different symptoms to show.

The thing is that it seems that tweaking a 'not optimal' tyre physics model isn't the sollution. Sadly I don't know how you *can* model tyres properly so I'm not of much help.

Scawen's 'avoiding of physics forum threads' doesn't help. How has he debugged LFS physics? How much situations has he put in the sim to calculate later by hand to see if weight transfer etc is 'close'? We all think its the tyres, and that is quite likely, but perhaps thats not the only thing..

Since I can have absolutely 0 input on making LFS 'better' (up to a point) I do get annoyed at its very slow physics progress. LFS is an awesome achievement in many ways but there are just a few too big things too wrong.

[Tweaker]

Yes there is an improvement over the previous version. However, over the years LFS has gone from 'some really huge problems' to 'some pretty huge problems'...
You've got to be kidding..... S1 was like.... stupid compared to S2. I wouldn't even compare what S1 was presenting us, because that was almost an entirely different feel. S2 is what we should be focusing on, since it is the focus for the physics. Would be nice if we could just get the S2 Final a solid tire code that feels great, rather than saying it hasn't even improved over S1, that is just absurd. Doesn't let the game evolve at all if you think S1 was somehow better.

[Niels Heusinkveld]

Language barrier I suppose :) I mean it DID get better from s1 to s2 to yesterdays patch. But it was 'far off' and still is 'far off'.. just less far off..

[RIP2004]

I totally agree with Niels. Its exactly what he sais. Problems are still there in some way, but its harder to see them. Now you have to push harder to get there ...

[EeekiE]

My LSD'd Polo suffered from lift-off understeer :pillepall So long as you kept the power on it would pull hard round the corner, but if you suddenly let off it would drift straight on LOL. Then again Mk3 Polo suspension sucks.

[Tyrion]

Isn't part of the problem that people are actively seeking out and performing tests they know will show the weaknesses of the tire physics? It seems hardly an objective approach.You can find flaws in anything with that mindset.

When that is said, your criticisms may be totally correct, I haven't the first clue about tire physics.But I am not questioning that of course, merely suggesting that preconcieved notions of the physics -or perhaps the game itself- may cause some people to go out of their way to find flaws, and be overly critical (this goes both ways of course).

Would love to see some more threads that focus on both what is good and bad.It would also reduce the risk of people running to the trenches as the discussion gets underway.

Hope someone gets my meaning.

Now please, carry on :)

[col]

...

It is funny to me to hear people here bash rFactor and call it arcade and whatnot, when some aspects of rF are actually more accurate than LFS.. rF's implementation of the physics engine is not as good, and the FF you get is no where near is good.. but in some ways it is still more accurate. Strange....



A clock that loses 1 second per day only shows the correct time every 236 years - a clock that has stopped completely shows the correct time every day.

col.

[Niels Heusinkveld]

Yes true.. Look at the different opinions on current sims. Simracing is very subjective.. some points:

1) Skill level varies a lot: someone doing a 2 minute lap isn't really driving the same as someone who does a 1:55..

2) What is skill? You can learn the limits of the sim but does that mean you're applying real driving skill? Are you stomping on 3 pedals at the same time with overly stiff / oversteery setups that noone would ever use in real life? Even two people who do a 1:55 (lets say thats a near WR time somewhere) will do this in quite different ways. It seems like you drive your 'style' and form an opinion on how the sim reacts to your style...

3) Style is overrated, 95% of simracers suck! (me included!) A LOT of simracers with their opinions have the relative skill level of ending up somewhere mid/backfield in a national celebrity Citroen Saxo race.

4) Computer limitations.. Some points here.. a 'flat' 3D image, some sound.. thats it.. Sims using a different FOV or use more 'head shaking' will make people conclude they 'feel different' even before you get to the physics.

5) Controller limitations. If 95% of simracers are poor (again me included), 98% probably use horrible controls. Standard off the highstreet store shelve pedals are very poor, and my recent trials with a Logitech DFP made me quite sad! Horrible! It made me understand part of the subjective differences. FF settings vary from person to person, totally changing the virtual drivers capability of judging the simulation. Its obvious that FF is flawed with the affordable wheels. Everytime a sim comes out people discuss the FF.. It shouldn't be a point of discussion. A good wheel with a good sim simply 'does' FF, yet the incredibly compromised Logitech designs are causing more grief than giving you feedback. So, sadly, mass produced wheels and pedals are more of a gimmick than good.

High quality pedals are often poor when it comes to brake sensitivity (I ranted about that in AutoSimSport jan/feb issue page 47 if you're interested). Thankfully the high end wheels don't have FF so in that respect they don't add another subjective parameter to the equation.

Then what is happening! We use office chairs and often race in fairly uncomfortable positions, mostly pressing and turning Logitech toys. A squashball makes the brake 'feel nice' (whilst screwing up its sensitivity).. the FF 'is great' while having little relation on the actual steering torque calculated by the sim....

Auch! What can I conclude? We call it simracing but in most cases this is certainly not true looking at the hardware we use. It seems that people look for the immersion and 'thrill'.. This is fine, but hampers your ability to objectively judge something. I can fall for it too! GTR with the canned FF effects and shaking view and sampled sounds has the 'wow' factor. It should however have nothing to do with judging its realism as far as physics go.

7) sims? I think Stefano (Netkar) said that sims are getting more and more alike and that this was a 'good thing' as it showed that basically 'we're getting there'. I disagree. First of all sims are worlds apart, and that is using proper pedals and for the sake of there not being good FF, a non FF smooth wheel. Sims are not all alike, and even if they would be, it might well be that they all would be making the same mistakes..

I always wanted to believe that sims came a 'fair way' since GPL in 1998 but I honestly don't think so. Yes we now have multilink suspensions, brake fade, semi complex aero... but thats no use if the tyre model and/or the data that feeds them is far off. Or if weight transfer is wrong..

Sims aren't good enough yet. Now I can't code more than a multiple choice QuickBasic questionaire so its a stellar achievement that these small programming teams come up with physics and graphics engines at all, but from a pure realism point of view, sims are not truly sims yet.

8)games? I think in 2006 racing sims are not professional and there are few people actually interested in ultimate realism. For some rF does great, for some LFS, for some NKpro, for some N2003. The general maturity of the players shows a lot too.. "I want this car in the game" "I want special drift tyres" "The reversed weight bias Porsche mod is GREAT" etc..

Am I a moaning old man? Well.. yes.. I demand a lot and can't be satisfied when all sims on my harddrive are so different. What if THE real sim showed up tomorrow? I think peoples opinions would still vary a lot because of their poor controls, poor interpretation skills and poor demand for realism.

There is light on the horizon with Todd Wasson's new ideas about tyre modelling, Eero's new sim, and good ol Kaemmer apparantly working on something. I can't wait to try these as its about damn time the bar gets raised! I just hope there are more than 3 people out there who actually care about it.

[JeffR]

The FWD car on street tyres (GTI) accelerates best when pushing the petal to the metal.Depends on the tire compound. Some tires lose very little grip when slipping. A bit off topic here, but because of the slipping grip limitations of clutches in most RWD cars, magazine testers just find the sweet spot rpm, drop the clutch and let the tires spin for acceleration measurements. The slipping tires provide more grip than slipping clutches in most cars. I think that stock clutches are designed like this on purpose to keep from breaking transmissions.

The RWD cars are better now, but still FZ50 without TC or RA are much more dangerous and difficult to drive than comparable cars in RL (like Porsche e.g.)Porche sets up their modern cars to have a significant amount of understeer. With all that mass at the back end, you don't want oversteer. Try drving a race prepped 911 and it will be a lot more sensitve.

I also noticed, that it is very hard to get your car drifting by suddenly take away throttle. Not because weight was transfered away from the back.Weight transfer is normally the smallest factor with lift throttle oversteer. In a rear wheel drive car, lifting the throttle applies a braking torque to the rear wheels only, so they have less grip available for cornering. The differential setup can exaggerate lift throttle oversteer reaction. Except for low speed, low gear corners where there's a lot of engine braking torque, the weight transfer from lifting on the throttle isn't much, as the braking torque at the rear wheels is small in higher gears.

(Repeating myself, but how many here are finding it hard to believe I'm defending LFS's tire physics now?).

[Ball Bearing Turbo]

Yes true.. Look at the different opinions on current sims. Simracing is very subjective.. some points:

blah blah blah


That is an extremely negative view, clearly you're a glass half empty person. First of all, do you actually have any real world racing experience?

What exactly are you expecting anyways? We have limited machines with limited hardware, and like it's some kind of divine revelation you draw the conclusion that the sim experience is limited on a PC... Excellent research :razz:

Whatever problems remain, I would bet that Scawen is far more aware of it than any of us. He's stated earlier that he's created test "software" to test the LFS tires under different conditions. Personally, I think he probably knows what he's doing. And if I was in his position, there is no reason to get involved in this discussion - he knows how LFS works and where it's weaknesses are already, and if his knowledge happens to contradict someone's opinion on this forum then all it generates is bad feelings. We already have had people telling them why their busniness model is poor. Should Scawen care? Obviously not.

In fact, based on your redundant essay above, we might as well not have opinions about LFS, since it's subjective. Guess what, opinions are by nature subjective, and so is experience. Since you cannot generate reality out of a PC, it will always be this way no matter what.

Here's the point:

Even if someone plopped on our hard drive today, a sim that was 100.0% perfect in every way conceivable, there would still be the RIP2004s and others that would be on the applicable forum, "explaining" why it was not right. All your essay proves is the fundamental human problem of perception within the limited tools we have to recreate a massive experience within tiny confines. In truth, of course it's not really possible. Excellent work discovering that!

Hopefully you don't think I am being rude, I'm just trying to get a point across. No matter what happens with any sim, people will always complain no matter what, period.

Edit: I am not trying to say LFS is perfect, but it's far more perfect than it was two days ago.

[Linsen]

one more note concerning laptimes pre- and post s-patch:

could it be that the roadcars/streettires in general are slower now, whereas the racingcars/slicks in general are faster? I haven't tested it, but my impression from reading several posts about laptimes is just that. Anybody know more about this?

edit: oops, this post should actually have gone to the "first impressions" thread (damn tabbed browsing ;)). I guess I can't move it, can I?

[RIP2004]

Depends on the tire compound. Some tires lose very little grip when slipping. A bit off topic here, but because of the slipping grip limitations of clutches in most RWD cars, magazine testers just find the sweet spot rpm, drop the clutch and let the tires spin for acceleration measurements. The slipping tires provide more grip than slipping clutches in most cars. I think that stock clutches are designed like this on purpose to keep from breaking transmissions.


That's correct. They find the sweet spot. If it was LFS, they would just floor the throttle and drop the clutch and it would be perfect. You need a certain amount of wheelspin but not that much.
The only car to start perfectly with this method is a AWD car ...

Note also that in race series there is a huge difference between starters. If it was LFS all would start equally with full throttle ...

It is a very old weakness, which is still there ... and most important it is tyre physics, which was one of the main topics of this update ... hope he will improve it again with next update

[Niels Heusinkveld]

LFS *can* rule though, thats obvious. With normal front tyres, trying to do scandinavian flicks in the Fz50 is such a joy / challenge. It really doesn't want to just throw the back around, you have to do some serious provoking. This replay shows some non expert but still cool action imho. ;)

Another very interesting thing might be TC. Now that you can have it on the Fz50, its kind of 'strange'. With an understeery setup and normal front tyres, you can provoke understeer: i.e. the car more or less goes straight. However, if you do that with TC enabled and your gas pedal to the floor, throttle will drop to 0% by the TC! So you're more or less going straight with loads of understeer, yet TC thinks there is 4% slip and cuts throttle? If TC on the Fz50 is cutting the throttle when rear wheelspin is >xx% then I'm not sure why front tyre understeer would lead to enough rear wheelspin to make the throttle stop completely... Who can explain that?

Of course if TC on the car acts when any of the wheels, driven or not driven, slip this is easy to explain..

Anyway even though I'm not that good yet this replay shows off the best side of LFS imo. :)

[Resound]

LFS *can* rule though, thats obvious. With normal front tyres, trying to do scandinavian flicks in the Fz50 is such a joy / challenge. It really doesn't want to just throw the back around, you have to do some serious provoking. This replay shows some non expert but still cool action imho. ;)

Another very interesting thing might be TC. Now that you can have it on the Fz50, its kind of 'strange'. With an understeery setup and normal front tyres, you can provoke understeer: i.e. the car more or less goes straight. However, if you do that with TC enabled and your gas pedal to the floor, throttle will drop to 0% by the TC! So you're more or less going straight with loads of understeer, yet TC thinks there is 4% slip and cuts throttle? If TC on the Fz50 is cutting the throttle when rear wheelspin is >xx% then I'm not sure why front tyre understeer would lead to enough rear wheelspin to make the throttle stop completely... Who can explain that?

Of course if TC on the car acts when any of the wheels, driven or not driven, slip this is easy to explain..

Anyway even though I'm not that good yet this replay shows off the best side of LFS imo. :)

Do the front wheels turn more slowly under heavy understeer? If they do and the difference between the front and back wheel speed is what triggers the TC, then that could explain it.

[Niels Heusinkveld]

Hmm yes that is probably it.

[jtw62074]

But all seems to indicate that the actual tyre MODEL hasn't really changed! Only the parameters that feed it have gotten a bit better.

I feel quite strongly that this is not the case, actually. After having written several tire models for my own simulator and then driving and really analyzing the reactions on not just the seat of the pants, but also the numerical level, it feels very much to me as though Scawen has made a large leap forward in the model itself, specifically in regards to lateral/longitudinal force combination. I.e., when you're sitting at some slip angle and then change throttle/brake, the lateral force changes too.

Before this patch I really thought it was not right at all and was the cause of all these tire problems. However, to me it really appears to have been drastically improved. I've run my own models with and without these combination errors (took a very long time to figure out a way to do it properly) and the difference is really very similar as to what we see in the change with this new patch.

I haven't tested things too much yet, but after a few minutes playing around in the car park with some of the cars it's pretty obvious this part of the model has been changed, and in a very good way. Good job, Scawen :thumb:

A lot of the discussion in this thread about handling quirks and so forth I think really can for the most part be attributed to the setup. You can engineer a real car to have lift-off oversteer or not, and so forth. That's what the setup is for and why suspension systems are adjustable. It's to allow you to adjust these under/oversteer tendencies in different situations. It's perfectly possible to build a real car that drives horribly, spins or doesn't spin when you hit the throttle, loses the back end or not on constant radius corners, and so on. That's chassis engineering and a great number of people make their livings at sorting that out with a real physics model (nature!) :)



I don't think you can say its 'high speed' or 'low speed'. This has often been mentioned in sims, and its true that most have different low speed physics.. but that is near 0 speed. Above 10km/h most sims will use their pacejkas or whatever model they run. Problems might be more apparant at certain speeds. At low speeds the car can change direction a bit faster, perhaps highlighting some issues with that. At high speed the car carries a lot of momentum which might cause different symptoms to show.


Yes, you're right about the high/low speed models. I don't know how it's handled in LFS and personally don't really care as the only time I drive that slow is after I inadvertantly punt someone off the track and wind up in the dirt next to him :D


How has he debugged LFS physics? How much situations has he put in the sim to calculate later by hand to see if weight transfer etc is 'close'? We all think its the tyres, and that is quite likely, but perhaps thats not the only thing..


Stuff like weight transfer and so on comes straight out of general rigid body dynamics calculations and you really can't get that wrong. As a result of the forces from the springs and so forth the weight transfer will just "happen correctly." I.e., there's no guessing or anything to check there.

[BWX232]

Isn't part of the problem that people are actively seeking out and performing tests they know will show the weaknesses of the tire physics? It seems hardly an objective approach.You can find flaws in anything with that mindset.



No it is that people play the game and notice these things.

[BWX232]

A clock that loses 1 second per day only shows the correct time every 236 years - a clock that has stopped completely shows the correct time every day.

col.
Ok, that's great, but I would rather have accurate look-up tables than incorrect real time calculations. edit- not implying either sim has any of those qualities.

[Ball Bearing Turbo]

Todd-

Thank you for the post, that was what we needed here... Someone with the arithmatic background and ability to analyse the recent updates to LFS. I could not agree more about the fact that setups affect what many self-appointed experts claim to be infallible evidence more than they realise.

[Michael Denham]

A lot of the discussion in this thread about handling quirks and so forth I think really can for the most part be attributed to the setup. You can engineer a real car to have lift-off oversteer or not, and so forth. That's what the setup is for and why suspension systems are adjustable. It's to allow you to adjust these under/oversteer tendencies in different situations. It's perfectly possible to build a real car that drives horribly, spins or doesn't spin when you hit the throttle, loses the back end or not on constant radius corners, and so on. That's chassis engineering and a great number of people make their livings at sorting that out with a real physics model (nature!) :)

Perfectly said!

[Honey]

A lot of the discussion in this thread about handling quirks and so forth I think really can for the most part be attributed to the setup. You can engineer a real car to have lift-off oversteer or not, and so forth. That's what the setup is for and why suspension systems are adjustable. It's to allow you to adjust these under/oversteer tendencies in different situations. It's perfectly possible to build a real car that drives horribly, spins or doesn't spin when you hit the throttle, loses the back end or not on constant radius corners, and so on. That's chassis engineering and a great number of people make their livings at sorting that out with a real physics model (nature!) :)

Perfectly said!
i did a lot of testing changing setups in every way even with extreme values, before speaking...to me seems that the handling problems arose here are related to this http://www.lfsforum.net/showthread.php?p=111845#post111845

[axus]

I think we need to see chassis flex modelled before we can asses the tyres much further - I think it might help the issue. There's been some pretty fine tuning going on with tyres and while the problem still exists under certain conditions it is close to being abolished. It has not merely been a case of changing a curve or two - as you see the tyre deformation has changed a lot so they must have looked at the roots of the problem. While LFS is still not perfect the problem is now quite easy to ignore because it doesn't appear under most driving scenarios that occur in proper racing. If you try to provoke it you will still get to it though.

[Honey]

the problem is now quite easy to ignore because it doesn't appear under most driving scenarios that occur in proper racing. If you try to provoke it you will still get to it though.
actually is the contrary: it always show up too evidently in normal conditions, that's why i'm spending so much words, everyone is caught by the "BF1 fever" and nobody is testing the slower cars...the few glitches in tire phisics are ruining this patch wich is at 90% a huge improvement

[Boris Lozac]

Anyway even though I'm not that good yet this replay shows off the best side of LFS imo. :)

Exactly! It feels soo good to watch this kind of replays and admire this brutall physics engine.. :) :thumb:

[axus]

actually is the contrary: it always show up too evidently in normal conditions, that's why i'm spending so much words, everyone is caught by the "BF1 fever" and nobody is testing the slower cars...the few glitches in tire phisics are ruining this patch wich is at 90% a huge improvement

I find it happens on a far more rare ocassion now, before it had gotten to the stage that the problem was predictable. Now because you don't know the conditions which bring the same problem about, you don't know when to expect it so it seems like not much has changed but a lot has.

[Honey]

I find it happens on a far more rare ocassion now, before it had gotten to the stage that the problem was predictable. Now because you don't know the conditions which bring the same problem about, you don't know when to expect it so it seems like not much has changed but a lot has.
i'm not sure i understood well what you say, but as i said in the other thread: take xfg put 2.2 bar and 3.5 camber (in RL you would never do such much) and drive around BL or SO at 50 kmh with F9 info and tell me if it is absurd or not