The biggest reason is that LFS has only 1 person doing the code, as you may know. It results that sometimes we get a patch that doesn't really include anything seriously special. I'm sure there are different opinions about the latest patch in this regard as well, for me it was small fixes but for a person using Windows Vista it made the game to work, for example. Another person may like the new sounds, someone else () may use older version...
LFS may be dead to you and I have nothing against it, for myself, I just can't enough of it
That huge post of Todd.. About the feel of 'break away' being more related to the stop of force buildup and not a curve that drops.. Can I apply this to LFS for a moment?
There have been acceleration and braking tests in LFS and it was very hard to do better than a 'complete wheelspin' start or a 'locked tyres' stop. This might indicate, ever so perhaps, that the rise of the curves in LFS is on the steep side? I know the curves don't drop much in LFS, which is probably the main reason that its the best sim 'once it goes wrong' but the buildup seems very quick. Its easy to go from no wheelspin to loads of wheelspin in an instant, and in a turn, the difference between going through it smoothly and leaving skid marks is tiny.
I think it's been taken into account in the Papy engine, nK Pro maybe rF and RBR. Even Forza takes chassis flex into account. LFS need not do anything revolutionary RoR style just some kind of simple slight movement in the suspension pickups could probably be far better than the current completely rigid body simulation. AFAIK RoR is the only simulator to use a non rigid body physics engine and to get the basic chassis flex effect it really is overkill with current PC performances, that said RoR has awesome potential.
@ajp71:
Do you know the story of the guy who began a post in a discussion with "I think..." and didn't give references to his sources and then was executed by throwing flaming bananas at him?
But didn't you race on the 13th of January at Redline Racing 3 in a 6 lap race? And four races in July? And over 20 races in June etc.
More than one third of your total laps have been on the oval in just one car. Some tracks you've hardly visited at all and some cars you've hardly driven at all. I'm sure you can get a lot more enjoyment out of LFS than you have so far.
Hope to see you back online in LFS again soon. There's plenty of good times waiting for you.
Chassis flex is defenately not overkill, it just needs to be done well. Making a system that calculates the chassis flex exactly based on the 3d-model will of course use a lot more resources than a race sim way, where you have a sort of simpler model of the flex equations, or a table filled with values. A simpler system which doesn't need to be able to handle infinite amount of different 3d-models (like in RoR) but just the one, the car in question. What RoR does it that it calculates the flex from the beam elements of the vehicles structure, for LFS this kind of approach would be waste of resources and, ever more importantly, wrong.
The biggest effect of chassis flex is the tire angles and weight distribution. When the car is going through a corner on the knife-edge, the weight is distributed slightly differently between the tires. A car never lifts it's inside rear tire in air because body flex, it is the suspension that is compressed in the opposite corner of the car. The chassis flex has the big effect on how the weight is distributed between the 3 tires still touching the ground and on the angles (camber and toe-in/out mostly). To simulate this you don't need advanced equations and matrixes, just some reliable information how the chassis in question reacts under hard braking, braking and cornering, accelerating and cornering etc..
I remember reading that the biggest problem with chassis flex are the high frequencies...
the problem with this is you cant do much engine modeling with todays cpus ... for reference look at any of todds posts on engine modeling
in short the only way is to precalculate much of it put it in lots of tables and work from there
which coincidentally could be the reason why racing legends has gone into hiding
they simply want to simulate too much so theyve decided to dig a hole and hide down there until the next big thing in cpu history drops ... unfortunately while they were down there they probably developed enough new code to depend on the cpu gen after the one they initially planned to use
there really doesnt see to be any gear box sim
possibly all that happens in the gearbox model is that torque and rpm get multiplied with the gear ratio and some factor to simulate losses
i think youre generally reading too much into this
the brake away feel is mostly about the lateral curve where pushing harder (ie requiring more centripetal force) suddently doesnt work anymore as the curve drops into the plateau (btw i think lfs is too smooth in this which yields the entire floaty spongy feel)
the absence of any difference in spinning or locking up versus threashold braking and hitting the perfect start indicates that lfses longitudinal curves dont drop much at all past peak
this isnt entirely correct but at the least a good bit closer to the real thing than rfactors curves
about those skidmarks ... i think lfs does overdo them a little at this point to get a quicker rubber buildup on the racing line ... irl you dont get skidmarks that easily
Racing Legends is still going strong... I think a little fairy might have told me something about a virtual wind tunnel based on the car's 3D model creating files to be used by the physics engine in real time (don't quote me on it though, my memory of the description is rather vague). This approach is awesome, IMO, for anything that's just too stressful to do in real time. Provided that the files are still generated through simulation, you don't necessarily lose anything in terms of the physics. Engine simulation, chassis flex and aerodynamics can be modeled really well in this way I think.
dont take that part of my post too serious it was more of a comic relief
and i do think that every development team should incorporate cfd sims in their sim to base the aerodynamics on
problem is i dont think theres any software available atm thats able to simulate both tyres and fluid dynamics to base the model for the tyres in soaking wet conditions on
Ive been ill and had to move back across to Glasgow so ive been away, don't have time right now to reply or read any other posts right now but you bring up a good point. I do love the LFS engine, its ingenious, but some how the game just doesn't float my boat. I first got into real sims years back with F12002 (i also played LFS back then when it was Test Demo (is that right?)) just because of the mods and thats the way its been since then. I love mods and being able to mod the game(s). I just feel LFS would be a wasted buy for me, i would never play it.
My comments on break away were in reference to lateral/cornering forces rather than longitudinal. The picture in longitudinal is similar, but it doesn't quite pertain to what you're referring to directly.
High slip ratios do indeed usually cause a drop in longitudinal force after the peak for reasons I won't get into here (I need to keep some things to myself for my model ). If the amount of acceleration you're getting when spinning the wheels is too high, then indeed the longitudinal curves could be tweaked to provide more of a drop off. I doubt very much that this is due to a flaw in the LFS tire model. It's most likely just a matter of tweaking parameters to get some slightly different curves.
Same goes for braking of course. In reality it's not quite that simple as temperature and other things effect it too making the picture quite dynamic, but a very good approximation can still be made with fixed curves. This is after all how it's done in the racing industry. If it's good enough for them it's probably good enough for us What people want is to have increased stopping distances when you lock a tire up. That can probably be accomplished with some tweaking.
In one dynamic test I saw, at the instant of lock-up the tire was still providing quite a lot of force, almost as much as you get at the peak. It then trailed off to a lower value over a small period of time, so the curves in reality are not really fixed. Again, a completely locked tire is a different situation than one that's rolling a little bit but still at a high slip ratio. The patch of rubber in the contact zone is not getting a chance to cool off through a revolution of the tire, but is instead sitting there being absolutely fried. Back it off to slip ratio -0.9 (-1.0 is fully locked) and the picture is quite different.
Anyway, all they need to do to get a really good approximation is to get the longitudinal curves to drop off a bit more (not the lateral ones though!). That'd be close enough for government work
Regarding the tyre lockup, I guess we'd see a more severe effect if we had tyre surface temperature modelled, which would pop to "OMG HOT" in an instant as the tyre locks up. Either that or the (after peak) infinitely flat slip curves get tweaked a little
) may use older version...
). If the amount of acceleration you're getting when spinning the wheels is too high, then indeed the longitudinal curves could be tweaked to provide more of a drop off. I doubt very much that this is due to a flaw in the LFS tire model. It's most likely just a matter of tweaking parameters to get some slightly different curves.