There are many many bumps in LFS you hit with one wheel first and the other later (or never), like sewer, tramway rails, curbs, grass, most bumps in south city. IRL you can feel through steering wheel longitudinal forces caused by track irregularities when they affect only one wheel or both wheels differently, and it happens very often.
For example IRL steering wheel is pulling one side when you have a wheel on a curb isn't it? Here roads are very bumpy and I can feel everything with the steering wheel. On smoother roads it requires to go faster but there is always a speed for which you can feel road surface.
In LFS no irregularity will pull your steering wheel, and this is really missing badly because it even suggests tracks are flat which is not the case, and hides amount of work done to improve track surface detail.
We should feel all bumps, even large ones because most of time we do not hit them exactly at the same with both wheels, and this bump feedback should be stronger when speed increases, as more energy is lost on bumps.
Last edited by Juls, .
I am still thinking again about it, and doing various experiments
I can clearly feel in LFS how the ground normal is changing below each tyre. When I run over a curb with the left wheel, the track normal is no more vertical under this wheel (assuming the car is horizontal). As a result, the tyre receives a lateral force, and I can feel this lateral force through caster and scrub radius. This is perfectly OK.
Similar thing when track is banked. The wheels receive lateral forces, and the steering wheel pulls.
But something is missing. When the tyre runs over a track irregularity like a bump, the wheel is losing a lot of energy and is braking because it is in fact like a little collision. And this is perfectly modelled visually in LFS, as we can see how the tyre deforms on a bump. This deformation shows that the bump is strongly braking the wheel, and we should feel this braking force through scrub radius. It should even turn the car slightly.
Scrub radius is modelled in LFS, tyre deformation too. But I do not feel at all the longitudinal braking force corresponding to this tyre deformation on track irregularities.
Last edited by Juls, .
Do a test. Show LFS and rFactor to many many persons, and each time, ask them what sound is the most realistic. Then shout at them they are stupid because they do not understand LFS sound is more realistic than stupid tape records of real cars, because it uses shorter tape records of real cars, but played more often. And you will have no friends anymore
(maybe you are right but you will have no friend anymore...that's destiny)What makes you think we were not using MVC ? Of course we used MVC we are not totally crazy
Sorry for OT
Unfortunately for me I have been doing everything on the other side of the line. For a long time my job has been to develop big GUIs for real-time trading client-server solution. The kind of stuff where dozens furiously clicking traders were connected through several stock exchanges through our poor GUI and generating all possible conflicting situations as they could interfere with each other in all possible asynchronous ways...woooohooooo
People from the server side looked soooo peaceful with their unit tests.
The main problem with unit tests is how to use them with code giving graphical result.
It's pretty easy to make a unit test for parts of code doing precise things like numerical functions...but there is no easy way to do a unit test that will detect when your buttons are misaligned, or something is flickering, or something is hidden when you resize this or that, or combobox does not have the right behavior. Same problem with code using directx.
To write this 6 million characters source code, you can bet Scawen had to type at least 3 times more...because it is never a direct process. You write, correct mistakes, come back later and add something, remove, refactor...
Last edited by Juls, .
To keep versions of source code for free, I use Tortoise SVN (not CVS as I wrote). No need for a special tool working with visual C++.
It integrates directly in windows browser. Just commit the changes from time to time and all versions are saved in repository. It makes life a lot easier and you can quickly commit some sources before tweaking the code without having to do some archive and name it
It integrates directly in windows browser. Just commit the changes from time to time and all versions are saved in repository. It makes life a lot easier and you can quickly commit some sources before tweaking the code without having to do some archive and name it
Last edited by Juls, .
Found it in the browser history!
It is a curve in a ppt presentation from Michigan University.
http://www.mae.wmich.edu/facul ... ture%2003-Tires%20pJZ.ppt
Page 8. It shows tyre friction coefficient for slip percentage. I suppose slip percentage is tan(slip angle)...15% slip percentage 0.15m sideway for 1 meter straight->8.5 degrees of slip angle. It goes quite far...15 degrees slip or more, and there is not this infamous friction coefficient drop you can feel in most sims. The drop looks like 5% of the peak value, no more.
I know cornering force drops, but FFB should not drop a lot because the main force friction stays very important and we can feel this strong force through caster.
And intuitively it is correct. If you push a block of rubber on a concrete plane, it will stick very well and then start moving, with a huge drop of friction coefficient. A tyre is different...whatever it does, a part of the contact patch is sliding...so even in straight line or very little slip angle, it already has partially the same behavior than a slipping block of rubber. Hence friction coefficient peak is lower, and after there is less drop.
edit: my mistake, this curve does not concern at all lateral behavior, but longitudinal slip! But the same intuitive thinking applies
I should sleep more! Lateral force is later in the document but unfortunately curves do not go a lot further than the peak.
Finally, there is one interesting chart in this doc. Page 20 it shows lateral force vs slip angle, up to 16 degrees
And it shows several times this curve for different acceleration and braking slip ratios!
And there is almost no drop after the peak!!!!! Even less than longitudinal force! Whatever acceleration or braking conditions!
Haha, knew I did not dream! ISI default FFB setup is absolutely wrong when it removes all FFB as soon as you lose grip.
Last edited by Juls, .
I remember wishing you turned LFS into a company with employees. Many times. It's not to offend you or to dictate what you should do...it's logical thinking when one appreciate LFS.
It always happen like that and I suppose not only for me. I run other sims, some mods, some cars, some tracks and it feels weird.
Then I come back to LFS, and it feels right. It feels right and it is the only sim for me that feels right until now. So how could I possibly NOT wish to have more of it, to have every track every car from other sims in LFS? How could I possibly NOT beg for it on forum?
Yeah yeah, I know it's selfish
Last edited by Juls, .
Yes, concerning this lift off issue, the 3.0 was the best for me. Do not misunderstand me. This mod is great and you did an awsome work.
I just think there is a problem with physics models in sims. Developers sometimes seem to think realistic means unforgiving. Because of that, catastrophic events like spin are too frequent in sims, and driving sims fast often means "finding the way to drive fast without spinning", whereas in reality it is more a question about proper trajectory, timing...etc.
I do not know if I manage to explain well. In real racing, spinning and crashing is far less present, and even at very high speed there is still a large choice of trajectories, possibilities, because tyres are more forgiving.
Punishment for a small mistake is a slight slide, loss of speed, wrong trajectory...and less frequently a spin or crash. It makes racing more enjoyable. (of couse we tend to go faster in sims but going slower does not solve the problem)
Recently there was another mod inspired by your work, the new version of caterham. It seems the author has worked a lot on the tyres curves using your tools and I enjoy it very much. It manages to give this large choice of possibilities even at high speed, like iRacing (from videos I have seen). For me this is the way to go.
(Yesterday I saw the curve of friction vs slip angle of a real tyre...after the peak, it just falls less than 10% and then remains flat! Self aligning torque falls, but certainly not friction...so many sims get this wrong)
I just think there is a problem with physics models in sims. Developers sometimes seem to think realistic means unforgiving. Because of that, catastrophic events like spin are too frequent in sims, and driving sims fast often means "finding the way to drive fast without spinning", whereas in reality it is more a question about proper trajectory, timing...etc.
I do not know if I manage to explain well. In real racing, spinning and crashing is far less present, and even at very high speed there is still a large choice of trajectories, possibilities, because tyres are more forgiving.
Punishment for a small mistake is a slight slide, loss of speed, wrong trajectory...and less frequently a spin or crash. It makes racing more enjoyable. (of couse we tend to go faster in sims but going slower does not solve the problem)
Recently there was another mod inspired by your work, the new version of caterham. It seems the author has worked a lot on the tyres curves using your tools and I enjoy it very much. It manages to give this large choice of possibilities even at high speed, like iRacing (from videos I have seen). For me this is the way to go.
(Yesterday I saw the curve of friction vs slip angle of a real tyre...after the peak, it just falls less than 10% and then remains flat! Self aligning torque falls, but certainly not friction...so many sims get this wrong)
Yes, and he writes he got rid of Pacejka model. There are not too many ways to model a tyre in a PC sim. It's rather a Pacejka or a model which divides the tyre in several parts and check how these parts grip or slide or stretch...like LFS.
So maybe now nkPro has a very close tyre model from what LFS uses. At least in concept.
So maybe now nkPro has a very close tyre model from what LFS uses. At least in concept.
This is not negative, but positive complaining. People complain because they see the huge potential in LFS.
In France they say: "Qui aime bien chatie bien"...something like: "Who likes well punishes well". The day there will be no more people complaining, then you have to worry
This is software world, game industry. Slow progress means quick death.
Last edited by Juls, .
Picture this: 150 kph, on a very very very light turn, like a highway sloooooowly smoooothly turning. You are far far away from the grip limit, very safe...german autobahn. 5th gear, a bit more than half throttle.
Then your exit lane is not far anymore. You lift the throttle to slow down...and you are dead, because you forgot opposite lock. Haha. Opposite lock at 150 kph on highway. Are you sure? Is it not a tad too sensitive even for old 911s ?
And not all 911s from historix mod have this lift off oversteer. Why? And these 911s with lift off oversteer are almost unable to oversteer pushing throttle. Why? People are satisfied because it looks like 911 and sounds like 911. I can't help seeing what feels weird sorry.
Anyway, I am not a LFS fanboy. Most people will realise how wrong tyre model is in rFactor (and FFB in Simbin sims) when they try a new generation sim. iRacing, updated nkpro, and I hope Blimey sims will be good too. I think historix mod is a very good mod, and gets the best out of rFactor tyre model. But of course they can not solve the tyre model flaws.
New sims have a more stable tyre model. Small driving mistakes do not lead to catastrophic events like a total loss of grip, but simply lead to a loss of speed, different trajectory...etc. The tyre itself will handle small mistakes through flex, twist...etc.You have to make stronger mistake to reach the diverging area.
Old sims tyre models have a narrower stable area. With a very little mistake you reach the diverging area and you can't recover.
If you move quickly your wheel on a highway and come back quickly to center, the tyres contact patch will twist and absorb the entire move, and the car trajectory will not be changed at all (a friend car mechanic engineer show me that). And this is even more true with old tyres like these 911 had.
Try this in rfactor or a Simbin sim and you will understand that tyre models they use is unstable, diverging. The older the sim, the more unforgiving the tyres.
Last edited by Juls, .
Sure, anyway, any improvement is worthwile.
If the game was chinese and intended for international use, it would be entirely written in english first.
Non-english speaking people rarely choose a game because it supports their native language....they choose a game because of the game itself. Most of us learn english because we want to play a game that is english only. Same with movies.
I have a multilanguage website with no visits. Maybe I should have done something else before internationalization
Non-english speaking people rarely choose a game because it supports their native language....they choose a game because of the game itself. Most of us learn english because we want to play a game that is english only. Same with movies.
I have a multilanguage website with no visits. Maybe I should have done something else before internationalization
Every time I scroll down a thread, I see only your ugly avatars 
This one is a bit less visible...but definitely the worst haha
This one is a bit less visible...but definitely the worst haha
I was wrong, wheel is pulling exactly as it should on banked tracks in LFS. 
For example, the oval turn is banked 18 degrees, and the car receives 0.3 lateral G....the wheel pulls exactly as much as on the skid pad when you reach 0.3 lateral G.
LFS models FFB through caster (unlike ISI engine based sims) and that should be enough to feel very well the bumps. Maybe the problems comes when LFS engine has to estimate what lateral forces are caused by the bumps.
When I watch this vid, bumps give very strong lateral force (the bump in t1):
http://de.youtube.com/watch?v= ... _b&lang=deu&id=20
And this lateral force, through caster, should give far enough FFB. I really would like to know how LFS feels with a track covered with narrow, strong bumps like IRL.
For example, the oval turn is banked 18 degrees, and the car receives 0.3 lateral G....the wheel pulls exactly as much as on the skid pad when you reach 0.3 lateral G.
LFS models FFB through caster (unlike ISI engine based sims) and that should be enough to feel very well the bumps. Maybe the problems comes when LFS engine has to estimate what lateral forces are caused by the bumps.
When I watch this vid, bumps give very strong lateral force (the bump in t1):
http://de.youtube.com/watch?v= ... _b&lang=deu&id=20
And this lateral force, through caster, should give far enough FFB. I really would like to know how LFS feels with a track covered with narrow, strong bumps like IRL.
Last edited by Juls, .
It concerns multiplayer. Look at the developer point of view. A browser is the perfect tool to make a multiplayer lobby. It supports chat, forums, can display documentation, display statistics stored in database in a very nice way. Everyone wanting to play a multiplayer game has a browser. Browsers are reliable, have been tested extensively by millions users.
Whenever a developer has to implement a multiplayer lobby, most of things he has to painfuly implement and test are already better done in a browser.
I asked a friend to try rfactor and LFS.
In rfactor he was spinning all the time, but he wanted to learn how to drive it...not because of graphics, but because it pretends to be a real car on a real track.
In LFS he started driving and did a few turns witout spinning, going a lot faster than in rFactor....and after three laps he spun and :" I can't drive that it is too difficult!".
Conclusion....if you give them pretended real tracks and cars, people are ready to accept everything. If you give them fictionnal cars and tracks, they will not forgive anything, they will not see how the physics feel better, they will get nervous at the first crash and lose patience. Amazing how few letters like M O N A C O or G T 4 0 can totally change people mind.
For example, I see people are fascinated with the new historic GTL mod for rFactor. I am sorry, handling is horrible...I took a porsche 911 and it was impossible to spin it with full lock and full throttle in 2nd gear....but every time I released the throttle in 5th gear almost in straight line it was spinning. If real Porsche were handling like that and spinning at 150 kph every time you lift your foot 2 centimeters up, there would be no driver alive...such throttle movement happens very often when the car shakes on bumps.
This and FFB spikes you can feel every track polygon with some cars...it ruins everything but people are hypnotized by few letters.
Really a pity LFS did not want to include real tracks.
In rfactor he was spinning all the time, but he wanted to learn how to drive it...not because of graphics, but because it pretends to be a real car on a real track.
In LFS he started driving and did a few turns witout spinning, going a lot faster than in rFactor....and after three laps he spun and :" I can't drive that it is too difficult!".
Conclusion....if you give them pretended real tracks and cars, people are ready to accept everything. If you give them fictionnal cars and tracks, they will not forgive anything, they will not see how the physics feel better, they will get nervous at the first crash and lose patience. Amazing how few letters like M O N A C O or G T 4 0 can totally change people mind.
For example, I see people are fascinated with the new historic GTL mod for rFactor. I am sorry, handling is horrible...I took a porsche 911 and it was impossible to spin it with full lock and full throttle in 2nd gear....but every time I released the throttle in 5th gear almost in straight line it was spinning. If real Porsche were handling like that and spinning at 150 kph every time you lift your foot 2 centimeters up, there would be no driver alive...such throttle movement happens very often when the car shakes on bumps.
This and FFB spikes you can feel every track polygon with some cars...it ruins everything but people are hypnotized by few letters.
Really a pity LFS did not want to include real tracks.
My problem with Vista is that horizontal span feature was removed (allows to stretch desktop on two monitors, and softwares will believe you have only one very large monitor and will display on both screens).
You can not play anymore on two monitors with Vista...unless you buy a matrox box...which is stupid because it was working very well with XP.
Did it come back with SP1?
You can not play anymore on two monitors with Vista...unless you buy a matrox box...which is stupid because it was working very well with XP.
Did it come back with SP1?
Only one person knows what is missing in LFS for the moment concerning bumps
What I can see, is that car trajectory is almost not changed in LFS when the car is on a banked road or receives lateral wind. And wheel is not pulling in such conditions.
I suppose this is somehow related to this lack of feel on bumps. Banking, lateral wind, bumps...all these things give some additional lateral forces on the tyre:
- on banked road the gravity is no more vertical (from driver point of view) and pulls laterally the car->suspension->tyre. Lateral wind does the same thing.
- when one wheel is on a bump, the road surface is no more flat under this wheel and pushes the wheel laterally.
It looks like in LFS tyres react "too well" to such lateral forces. They counter these lateral forces without needing any help, and as a result no steering torque is transmitted to the wheel, and car trajectory keeps almost straight and no correction is needed. IRL, such forces twist the tyre, and change the contact patch shape. As a result, you can feel a steering torque, and the car is not moving anymore in the same direction than rims, but slightly sideways. We all know how feels a strong lateral wind on a highway...
There are many bumps in SO tracks, my favorite ones. Very detailed road surface, with details like sewer...tramway rails. Maybe some more narrow bumps (affecting only one wheel) would be nice, but there is already everything needed to give a very good "bump experience" if FFB is updated.
I think new sims like iRacing are a very good thing...they push the envelope and force other sims to reach higher accuracy. I can bet next sims to be released will make a lot of efforts to better render bumps
What I can see, is that car trajectory is almost not changed in LFS when the car is on a banked road or receives lateral wind. And wheel is not pulling in such conditions.
I suppose this is somehow related to this lack of feel on bumps. Banking, lateral wind, bumps...all these things give some additional lateral forces on the tyre:
- on banked road the gravity is no more vertical (from driver point of view) and pulls laterally the car->suspension->tyre. Lateral wind does the same thing.
- when one wheel is on a bump, the road surface is no more flat under this wheel and pushes the wheel laterally.
It looks like in LFS tyres react "too well" to such lateral forces. They counter these lateral forces without needing any help, and as a result no steering torque is transmitted to the wheel, and car trajectory keeps almost straight and no correction is needed. IRL, such forces twist the tyre, and change the contact patch shape. As a result, you can feel a steering torque, and the car is not moving anymore in the same direction than rims, but slightly sideways. We all know how feels a strong lateral wind on a highway...
There are many bumps in SO tracks, my favorite ones. Very detailed road surface, with details like sewer...tramway rails. Maybe some more narrow bumps (affecting only one wheel) would be nice, but there is already everything needed to give a very good "bump experience" if FFB is updated.
I think new sims like iRacing are a very good thing...they push the envelope and force other sims to reach higher accuracy. I can bet next sims to be released will make a lot of efforts to better render bumps
Last edited by Juls, .
Yes I understood that
What would be interesting is to analyse what are exactly the differences, it may be interesting information.LFS tracks lack bumps, and maybe this could be added easily (it seems the engine supports quite detailed road surface). But I suppose this is not the only difference.
I have a few questions for iRacers:
- when you are on a track with a strong slope, like an oval, and you keep your wheel centered...does the car follow a bit the slope or continues perfectly straight?
- when you are on a bumpy track, with your wheel centered, does your steering wheel move with the bumps?
- when you have one wheel on the track and the other on the concrete stuff they put inside/outside corners(forgot how to say in english), can you feel your steering wheel is pulling in one direction?
Last edited by Juls, .
In fact I can feel bumps in LFS too when the wheel is centered, but it is a bit weak when the FFB is at 50%. At 200% it is a lot better but then the self aligning torque is too strong for my wheel
I was thinking about something else...not only one should feel the bumps more, but bumps should have an impact on car trajectory and require correction.
In LFS when I take a strong bump only on one side (cutting a chicane), with the steering wheel centered, I can feel it a bit, but it does not seem the car trajectory changes. Or if the car trajectory changes it is too little, because I do not need to correct at all.
I do not have iRacing to test, but I suppose you feel bumps very strong and bumps change car trajectory too, and that's why users say they have to correct their line all the time. Is it right SpaceMarineITA ?
I was thinking about something else...not only one should feel the bumps more, but bumps should have an impact on car trajectory and require correction.
In LFS when I take a strong bump only on one side (cutting a chicane), with the steering wheel centered, I can feel it a bit, but it does not seem the car trajectory changes. Or if the car trajectory changes it is too little, because I do not need to correct at all.
I do not have iRacing to test, but I suppose you feel bumps very strong and bumps change car trajectory too, and that's why users say they have to correct their line all the time. Is it right SpaceMarineITA ?
Think about people who use a force dynamics 301. I doubt they will miss this bug.
FGED GREDG RDFGDR GSFDG