I was wondering if (or "when", actually) ScaViEr are going to implement similar car body physics in the future versions of LFS that can be seen in Rigs of Rods, for example. Currently the car is just a solid rigid model.
While watching slow motion replays I'd like to see the whole car frame (including doors, exhaust pipes and whatnot) react to forces, wobbling, shaking, bending/reverting and so on. I hope you guys know what I'm trying to say here.
I don't know what you're saying really. The car body does pitch and roll and bounce and hop, the suspension compresses and extends underneath it. Maybe the exhaust could rattle about a bit (not that you'd tell unless it was a slow-mo vid, but I don't know what the doors, or any other body part, should do independantly of the body - thankfully on every car I've driven they just stay put........ unless, in the case of the doors, I've left them wide open. None of the cars in LFS would suffer from noticeable body flex.
In a slow motion video, chassis and body parts under heavy stress would show some bending, most noticeable in a cabrio I think. But most important would be possibility of adding proper karts
The problem is that chassis flex is a very high frequency process, so it would eat tons of CPU time. Maybe in a future if Scawen can make the simulation to use multiple cores it could be suitable.
Scawen would probably add it if it added meaningfully to the racing, but not if it was simply used to make replays look better. I love ROR, but the way I see it- those physics aren't really that realistic. They look as if they've been pumped up an awful lot to showcase the technology underneath. (?)
2 major problems with this
1) very high computational effort needed to get it right ... youll notice that ror looks bad runs bad and despite all this the trucks are still nowhere near as stiff as they should be (i still love it but truth is its more of a tech demo than a complete technology)
2) collision detection with ror physics is a nightmare since the cars only physically exists at the nodes
i dont think they have ... probably just made softer to keep the calculation frequency lower and save a few cpu cycles
In ROR, the chassis flex a TON, quite literally it looks like bodies are made out of rubber bands or something.
Thus far I don't think LFS needs chassis flex just yet, I may be wrong, but I don't think 10 or 20 minute races will have an adverse effect on the shape of a chassis.
24 hour races on the other hand.. possibly.
However, in the MRT, yes that would be noticed, otherwise I don't think there would be much effect.
I personally can't say I've ever felt chassis flex before so I don't know.
Yes! That's it. Thanks. I was having trouble finding the correct term
You guys are correct. The flex in RoR is way too exaggerated, that's why I used the word 'similar'.
Thanks for the comments so far
EDIT: You think we'll see some form of chassis flex when the damage model gets ScaViEr's full attention in S3? With detachable doors, hoods, wind shields, lamps and so on...
It's important for LFS to model this eventually if it wants to continue down the simulation route. Karts cannot be simulated properly without chassis flex and even formula cars have more flex than you would first thing, they use stiff ARBs to stop the body twisting, so low ARB values just won't work well.... a problem we don't have in LFS. Due to the typical car chassis being very stiff though, the natural frequency is very high so you need a very high physics rate. LFS at 100Hz would not cut it (only the tyres are run at 2000Hz in LFS). I'm not sure how high you would need to go but obviously there could well be a substantial performance hit involved. I can't see this happening before LFS goes multi-threaded.
But even then, it wouldn't be necessary for it to be accurately modelled visually (which is the suggestion here), and it certainly isn't necessary for it to be modelled visually on the big cars in any way at all, even if it's there in the physics. It's just not something you see unless you're looking very, very, very, very, very closely, or the car is big yank tank drop-head, or equivalent.
well I had 2 karts and never saw them flex...prolly because it was an offroad kart...but still, they had no stress cracks in the paint anywhere....just some BENT steering rods and tyre axels..
I assume that as Karts have no suspension, the chassis itself must act as suspension. As suck the material has a certain "flex" to it that allows it to be a basic suspension system while keeping complexity, costs and weight down.
Circular metal tubing, of similar dimensions, is weight for weight STRONGER and STIFFER than square tube. That is why roll cages are round...
Cars aren't made of out square tubing anyway - they are a monocoque construction these days, where the skin, rather than separate tubes, carry the stresses.
A kart needs the flex because it DOESN'T have 'suspension' as such (i.e. no springs, dampers, rockers etc), and thus relies on the controlled 'give' of the chassis to provide compliance. Without chassis flex a kart would be undriveable. It's absolutely required.
Road cars flex too - it might not look like they do watching the doors (for example), but it's there, and measurable in millimeters easily. Some, cheaper, cars flex so much the doors won't open when the car is jacked up on one corner. And the bonded glass will provide a lot of the stiffness too - break the windscreen and your chassis rigidity can drop by up to 60% (depending on the car).
My offroad kart had no suspension really, had NO rear suspension at all and the front had not even a 2" spring, and still I dont think there was any stress cracks, might have been but i never washed it really maby only once or twice within the 2 years i had it......
It had no suspension, and no springs, but it still had chassis flex - it won't have been concentrated flex (which would cause stress cracking in the paint/coating, and ultimately failure of the chassis through fatigue), but flex throughout the chassis. It will be measurable also in millimeters. You can quite easily rig up a home-measuring system if you want, with a torque wrench, some clamps, and something to measure an angle with, e.g. a plumb line, a rule and some basic trig'.
They would all suffer from body flex, there's a reason why an F1 car is 600 times stiffer than a typical road car (likely an unscientific statistic that I can't remember the source of but you get the idea). Chassis flex is still a limiting factor in single seaters, there are two very similar single seaters in F4, same engine, same tires, same wings, similar drivers, similar design and weight distribution and suspension setup. The big difference is one is a lot lighter than the other because the tubular steel chassis has less thiner chassis tubes. The heavier car is about 30kgs over the minimum weight limit but it is still a lot faster because it is so much stiffer. Not a scientific example but it illustrates the point nicely.
RoR shows the limits of current computing power, one or two vehicles in a very low detail system running at an uncomfortably low frequency can grind todays top notch machinery to a halt with very little in terms of non-chassis physics demands.
Chassis flex isn't the permanent physical distortion of the car body, that's only noticeable in extreme cases, like the Capris. Chassis flex is the dynamic stretching of the chassis, it happens in every car and makes a big impact on the handling characteristics.
Why do you think the MRT would be particularly badly effected? It should be a lot stiffer than any of the road cars given that it is a small, if anything over engineered racing car designed with all the latest equipment. The UF1 would be the worst effected car, in standard form and with a soft top conversion a Mini would either fold in or need a huge amount of strengthening material (weight) added.
TBH given RoR performance multi-threaded and with minimal demand on anything else I can't see a multi-vehicle non-rigid physics modeling happening in an acceptable level of detail for years to come.
parallel vector processors are quickly being introduced into the pc market so while i wouldnt hold my breath chances are by the time s3 comes out a parallel coprocessor which should easily chew through ror magnitudes faster than any current cpu might be standard in pcs
:hide: