i guess the one on the right is DX9. the texture on the burgundy part in the top right looks a bit more blurry.
i hope to god Scawen improves the lateral force / slip angle fidelity before we see reflections and dynamic shadows
although driving in the rain in netKar is really cool and the reflections do play a part of that .. but after i'm done with netKar i always come back to lfs
i did a few student placements at IBM in tech support. during my first work term i was often asked to convert documents from newer MS Office formats to the previous version so they could be opened by employees using OS/2. when i suggested people could just install windows 98 i was told it would be a cold day in hell when IBM installed 98 on its employees computers.
the next year i spent a good part of my student placement installing 98 on IBM employees' computers.
Scawen i agree with you 100% and i think that in recent years we've seen that although Microsoft's approach to business worked very well for them there has been some push back from the public.
actually even old hardware is blindingly fast. i have a commodore 64 that only runs at 1 megahertz but it can do a lot of calculations in a second.
the key is that even though computers got faster in the last 10 years the core MHz hasn't changed that much. for lfs to have a faster physics engine we either need new technology that will give Scawen 8x the processing power (per core) or he needs to come up with a new trick to get more done with the same processing power.
Doom is an example of a good trick that got more done with existing technology. it did that by leveraging a data structure to precompile expensive calculations.
oh man i feel like i missed a party here but seriously there is about as much agreement going on here as there is between republicans and democrats in the US congress.
where to begin?
fact: we like XP so much because it is a version of NT. in fact windows NT, XP, Vista, 7 and 8 all have the same operating system kernel developed by IBM for OS/2 (which was supposed to be a joint venture between IBM and MS). NT features the HAL (hardware abstraction layer) which is a bridge between the applications and all the nice hardware in our machines. newer versions of windows add extras that may or may not be useful and that WE DON'T NECESSARILY NEED. we need NT and the HAL to use LFS on windows.
fact: once compiled, our applications are just processor instructions. LFS running on XP is no different from LFS running on windows 8 or linux.
dustin has a point that there are a ton of ongoing security issues in microsoft software and that XP is no longer being patched. where i respectfully disagree with him is that almost all bugs that have been patched in the last half a decade are related to malicious websites.
as has been pointed out the average user gets infected mostly from that. my grandmother repeatedly got infected (while using fully patched vista) and one of my neighbors got the conduit malware too under fully patched windows 7. (source: i cleaned their computers). my XP machine never got a virus. it all depends on how you use your machine and how much you know about technology. patches are a good idea to limit how easy it is for malware writers to do their evil, but the malware writers are always one step ahead.
i'm comfortable stating i'm not an idiot for using XP. i have a dedicated machine hooked up to a simulator cockpit and use it only for that. if i were to connect it to the internet it would be so LFS.exe could do multiplayer connections.
unlike MS code, i'm pretty sure there aren't a zillion buffer overflow vulnerabilities in LFS.
paying for a windows 7 license for this sim cockpit would give me zero benefit at this moment. i literally would not see the difference as i did laps around aston in an fz5.
as has been mentioned, developing on XP in DX9 is only better for the user base. we can use our win 8 DX 11 machine to run this code if we want. no one is being forced to be less secure by this design decision!
one note to make about shader support in DX9 is that shaders do not benefit from FSAA! just look at euro truck sim which does a lot of heavy shader stuff and you'll see people with $500 video cards who complain the framerate isn't good enough. there are cases where shaders are great, mostly for kids games where cool looking graphics will increase sales. personally i don't like movies that rely on special effects instead of plot (say, X-MEN).
but seriously there is about as much agreement going on here as there is between republicans and democrats in the US congress.where to begin?
fact: we like XP so much because it is a version of NT. in fact windows NT, XP, Vista, 7 and 8 all have the same operating system kernel developed by IBM for OS/2 (which was supposed to be a joint venture between IBM and MS). NT features the HAL (hardware abstraction layer) which is a bridge between the applications and all the nice hardware in our machines. newer versions of windows add extras that may or may not be useful and that WE DON'T NECESSARILY NEED. we need NT and the HAL to use LFS on windows.
fact: once compiled, our applications are just processor instructions. LFS running on XP is no different from LFS running on windows 8 or linux.
dustin has a point that there are a ton of ongoing security issues in microsoft software and that XP is no longer being patched. where i respectfully disagree with him is that almost all bugs that have been patched in the last half a decade are related to malicious websites.
as has been pointed out the average user gets infected mostly from that. my grandmother repeatedly got infected (while using fully patched vista) and one of my neighbors got the conduit malware too under fully patched windows 7. (source: i cleaned their computers). my XP machine never got a virus. it all depends on how you use your machine and how much you know about technology. patches are a good idea to limit how easy it is for malware writers to do their evil, but the malware writers are always one step ahead.
i'm comfortable stating i'm not an idiot for using XP. i have a dedicated machine hooked up to a simulator cockpit and use it only for that. if i were to connect it to the internet it would be so LFS.exe could do multiplayer connections.
unlike MS code, i'm pretty sure there aren't a zillion buffer overflow vulnerabilities in LFS.
paying for a windows 7 license for this sim cockpit would give me zero benefit at this moment. i literally would not see the difference as i did laps around aston in an fz5.
as has been mentioned, developing on XP in DX9 is only better for the user base. we can use our win 8 DX 11 machine to run this code if we want. no one is being forced to be less secure by this design decision!
one note to make about shader support in DX9 is that shaders do not benefit from FSAA! just look at euro truck sim which does a lot of heavy shader stuff and you'll see people with $500 video cards who complain the framerate isn't good enough. there are cases where shaders are great, mostly for kids games where cool looking graphics will increase sales. personally i don't like movies that rely on special effects instead of plot (say, X-MEN).
would you think using a lookup table instead of doing any calculations is something a serious sim programmer would never consider? yes it is a huge shortcut but i get the feeling some sims do yaw = steer angle * reduction factor where the reduction factor is like tire heat, how much it's locked, etc.
XP alive and well in my sim cockpit! i could have bought a license for windows 7 but i bought two power leather seats out of a volvo S80 instead. difference of OS in LFS: did not notice any yet. difference of cockpit with new seat:
that's good to know notanillusion. if you can notice what i mean with the mouse then what i'm experiencing is not an illusion
that's a good point. you would get some delay in steering feel with a low pressure tire like the ones we use a lot in lfs.
i don't think LFS has any delay in the controller to speak of. i'm running the sim at a steady 144fps and if i sit in a parked f1 car with the brake on and turn the wheel i feel the fronts gripping and see them moving pretty much instantly.
the effect i mean is really the yaw rate of the car.
LFS Physics approach vs other sims
after spending some time in netkar and going back and forth with LFS i found the distinct impression that in other sims the vehicle yaws as soon as you turn the steering wheel.
in LFS it seems like the sim wants the front tires to prove the car should be yawing and it waits for the feedback to come on a later update cycle. (i'm talking in terms of program design and how in the physics loop the steering angle affects the car.)
i think it's the reason LFS feels more organic: the forces that are produced by the tires are calculated and then applied to the inertia of the car (including rotational inertia of the tires which at speed damp the yaw as well). other sims are happy to yaw the car as a result of the steering angle directly.
what do you think? do you agree there is a delay in LFS compared to other sims? do you think there is an extra fidelity that is gained in exchange?
in LFS it seems like the sim wants the front tires to prove the car should be yawing and it waits for the feedback to come on a later update cycle. (i'm talking in terms of program design and how in the physics loop the steering angle affects the car.)
i think it's the reason LFS feels more organic: the forces that are produced by the tires are calculated and then applied to the inertia of the car (including rotational inertia of the tires which at speed damp the yaw as well). other sims are happy to yaw the car as a result of the steering angle directly.
what do you think? do you agree there is a delay in LFS compared to other sims? do you think there is an extra fidelity that is gained in exchange?
maybe when ipv6 is fully rolled out it will be a better time for S3
so this became a thread about iracing
to be fair they aimed for something and got it, online racing without as much crashing and bad driving once you get out of the rookie license. their content is freaking great and everything else is minimal effort. they couldn't make all that content with their expensive and detailed processes if they didn't charge what they do. as stated the server costs are not that big for sure.
it's totally an opinion thing, but i can't stand to listen to iRacing for long. you can hear the lack of engine inertia modelling on every shift. that and the way other people's cars move relative to yours give the iRacing world a feel like paper cutouts with no weight.
certain cars in LFS can give you a headache from a monotonous drone, but the whole sound environment is organic. LFS just has more suspension of disbelief for me.
we can't really say one is better than they other. LFS aims for great physics. iracing sort of mentions physics and they're not that bad really depending on what you're looking for, but they aren't out there saying they aimed for physics as top priority and they have the best physics and their physics physics physics. we are the ones making that battle.
LFS almost overdoes the "pushing to the limit" thing. you're always on the limit without much effort. if you want you can set up your car to be more dangerous and fast, then you have to have more skill but you get extra tenths. kind of like tightrope walking.
in iracing you're guaranteed a nice stream of content. v8s at bathhurst wasn't there when they launched. and why should it stop now?
if you want the hardest challenge and you want to find your inner racer, LFS is for you. if you want to have "fun", then maybe iracing is better.
to be fair they aimed for something and got it, online racing without as much crashing and bad driving once you get out of the rookie license. their content is freaking great and everything else is minimal effort. they couldn't make all that content with their expensive and detailed processes if they didn't charge what they do. as stated the server costs are not that big for sure.
it's totally an opinion thing, but i can't stand to listen to iRacing for long. you can hear the lack of engine inertia modelling on every shift. that and the way other people's cars move relative to yours give the iRacing world a feel like paper cutouts with no weight.
certain cars in LFS can give you a headache from a monotonous drone, but the whole sound environment is organic. LFS just has more suspension of disbelief for me.
we can't really say one is better than they other. LFS aims for great physics. iracing sort of mentions physics and they're not that bad really depending on what you're looking for, but they aren't out there saying they aimed for physics as top priority and they have the best physics and their physics physics physics. we are the ones making that battle.
LFS almost overdoes the "pushing to the limit" thing. you're always on the limit without much effort. if you want you can set up your car to be more dangerous and fast, then you have to have more skill but you get extra tenths. kind of like tightrope walking.
in iracing you're guaranteed a nice stream of content. v8s at bathhurst wasn't there when they launched. and why should it stop now?
if you want the hardest challenge and you want to find your inner racer, LFS is for you. if you want to have "fun", then maybe iracing is better.
a tiny bit off-topic but reading about different people's controller experiences is cool.
if you use a force-feedback joystick like a MS sidewinder, you do your steering against the force coming back from the wheel. it's hard to describe but you have the fast lock-to-lock advantage a mouse has only with more smoothness and a lot of help from feeling the grip.
it also has an analog thingy you would use for the airplane throttle that works for gas and brake.
still using a wheel is so much more immersive any disadvantages don't matter.
--
on topic, i would say the physics engine allows some things that could give an advantage. you can get real-time information about the car's handling from insim ... say longitudinal slip? and since we're driving with computers, the inputs given to LFS could be modified based on that ... say the clutch?
an enterprising person could do the equivalent of gaining a tiny advantage kind of like a race engineer can get a few extra hp by reprogramming the motor's ECU. all part of the game then? would that really count as an exploit?
if you use a force-feedback joystick like a MS sidewinder, you do your steering against the force coming back from the wheel. it's hard to describe but you have the fast lock-to-lock advantage a mouse has only with more smoothness and a lot of help from feeling the grip.
it also has an analog thingy you would use for the airplane throttle that works for gas and brake.
still using a wheel is so much more immersive any disadvantages don't matter.
--
on topic, i would say the physics engine allows some things that could give an advantage. you can get real-time information about the car's handling from insim ... say longitudinal slip? and since we're driving with computers, the inputs given to LFS could be modified based on that ... say the clutch?
an enterprising person could do the equivalent of gaining a tiny advantage kind of like a race engineer can get a few extra hp by reprogramming the motor's ECU. all part of the game then? would that really count as an exploit?
in the patch test thread there some crash report. apparently if lfs is running in the background and you switch back and forth from fullscreen in asseto corsa, lfs crashes.
maybe Scawen is scouring directx documentation to find out what steps a program has to do to satisfy the internal requirements of a microsoft DLL. who knows.
maybe Scawen is scouring directx documentation to find out what steps a program has to do to satisfy the internal requirements of a microsoft DLL. who knows.
i was really surprised to find a video debunking that only iracing has the 2FMSH. it makes some sense if the rear contact patches are bigger than the fronts, but i thought you could do that in any car in iracing. too tired to google more
rf2 does have a really sweet tire model. i might have to buy it to be able to drive the nurburgring in more cars
ya man, you said "he just didn't work..." after Scawen said "I'm working...". basically you called him a liar. it makes sense based on what you know.. a long time without updates. i am a programmer and math/science/physics amateur and i believe you can spend years working and really get some things done and understood without reaching your goal yet. it depends on the kind of work.
it's what they say when they crown a new king:
The king is dead! long live the king!
it means the person can die but there is still a king. i was saying that to be funny, because people say lfs is dead but it will never die. hell i made a new track for 4d driving a few weeks ago (after all these years i never knew about shift-f1!!!!!!! damn!!!! (to edit terrain too))
ahh.. i still use windows XP on my racing cockpit computer
S.E.T.H. i found your post friendly and not insulting. you insulted Scawen with an assumption, so i'm not saying i like you, but anyway, we have a saying in french that it takes all kinds of people to make a world. at least you have dialog sometimes
but you didn't answer my question.the most you said was "you can have good races in nfs too".. as if need for speed is a platform two people can use to test their skills against each other.. wait a minute, where is that...
there.AC has no multiplayer, iracing has that.. thing.. and the guy passing me is holding the gas and brake at the same time to magically straighten his car out. i tried to like it! i felt like i was racing against the sim instead of the other drivers. what else is there.. rfactor. in the words of peter griffin COME ONNNNNNNNNNN.
you serious? i tried the demo of rf2 again for fun. i couldn't get on the track without a black flag in practice. even with the AI steering for me it still got a black flag. when it decides to enter "slide mode", it's like i'm in a big SWITCH statement. people who insist it's good say you need a car with better physics. the physics is different for every car. the physics is different for every car.now. please, be nice to molocco! start your own thread please! he's maybe not right and you can make fun of the concept of "gentlemanly" but he does have a little point. and you will sleep better tonight if you start a thread instead of trololol.
back to topic, no one would hate to see a patch man! and your plea isn't a bad idea. fordman's thread was so cool. can i just say your approach isn't as sensitive as fordie's was.
if i was a dev and read "you used to do this stuff..." i would feel bad.
of course it's a matter of opinion and trust, but i personally see no reason to doubt Scawen when he says he's working on the tyre model. everything about the work he did for me personally leads me to believe he wouldn't lie. LFS is a piece of software with integrity written all over it.
no one wants to please the community more than this guy. this is the topic right? a patch would be nice? then you have to convince me he could make a patch and just doesn't feel like it.
ps sorry about the off-topic. i replied to seth because i don't want someone who doesn't know lfs to read his post without a counter argument.
@S.E.T.H. your bashing of the LFS tire model is purely opinion.
you said it yourself, there will never be a perfect simulation of reality, so what do we do then? Scawen came up with something to allow us to race online. i would argue what makes the tire model so good is how organic it is. the same tire model applies to all cars, all tire sizes, all surfaces, and reacts well in all situations.
just to give an example, i did a 180 spin on asphalt at 5% throttle in a miata in iracing (at 2mph). it's only my opinion, but i don't like a physics model that has some situations where completely impossible things happen. like in netkar pro when you slide 500m across sand at low speeds.
LFS is made to do one thing and does it really great. you can bump against other people online and if you lose control of the car you never feel like it wasn't your fault. lag spikes are the exception but it's a problem impossible to resolve with the communication delays we have.
about your point that LFS has become our reality and we compare everything against it. maybe you have too little LFS experience to give it credit? how many hours have you spent trying to make a setup that suits your style of driving? can that even be done in other sims, or are their suspension calculations all based on one example of a car, with one particular setup, meant to be driven in one particular style?
personally i've tried colin mcrae rally 1 and 2, 1nsane, Screamer 4x4, GP3, GPL, iracing, rfactor, netkar pro, lfs and a lot more that are too far from simulation to count. nothing else feels so much like driving, so i choose to drive that mostly and because of that i'm unable to tell other tire simulations are better?
no, it's your definition of what makes a tire simulation good. yes, agreed the tires heat up too fast and cool off too slow. what about iracing has no flatspots? the tire is the same average temperature on all points. what about rfactor... well never mind. what about netkar where the tires screech at 0.5 mph?
i ask you a question: what is it about the lfs tire model that stops you from having a good race against other people? that is what its design priorities were for.
you said it yourself, there will never be a perfect simulation of reality, so what do we do then? Scawen came up with something to allow us to race online. i would argue what makes the tire model so good is how organic it is. the same tire model applies to all cars, all tire sizes, all surfaces, and reacts well in all situations.
just to give an example, i did a 180 spin on asphalt at 5% throttle in a miata in iracing (at 2mph). it's only my opinion, but i don't like a physics model that has some situations where completely impossible things happen. like in netkar pro when you slide 500m across sand at low speeds.
LFS is made to do one thing and does it really great. you can bump against other people online and if you lose control of the car you never feel like it wasn't your fault. lag spikes are the exception but it's a problem impossible to resolve with the communication delays we have.
about your point that LFS has become our reality and we compare everything against it. maybe you have too little LFS experience to give it credit? how many hours have you spent trying to make a setup that suits your style of driving? can that even be done in other sims, or are their suspension calculations all based on one example of a car, with one particular setup, meant to be driven in one particular style?
personally i've tried colin mcrae rally 1 and 2, 1nsane, Screamer 4x4, GP3, GPL, iracing, rfactor, netkar pro, lfs and a lot more that are too far from simulation to count. nothing else feels so much like driving, so i choose to drive that mostly and because of that i'm unable to tell other tire simulations are better?
no, it's your definition of what makes a tire simulation good. yes, agreed the tires heat up too fast and cool off too slow. what about iracing has no flatspots? the tire is the same average temperature on all points. what about rfactor... well never mind. what about netkar where the tires screech at 0.5 mph?
i ask you a question: what is it about the lfs tire model that stops you from having a good race against other people? that is what its design priorities were for.
LFS is dead! long live LFS!
if i knew everything i know about S2 but didn't have the license, then i would think $46 canadian dollars is very cheap.
to give you an idea i ordered the Milliken "Race Car Vehicle Dynamics" book in order to get the most out of live for speed. the book cost more than the software.
there are some very good points made about how many people you can find online. assuming you are racing in Japan, then like me you are in a timezone without many racers online in convenient hours. at least for my experience i can tell you i had some very good racing online recently.
that is not to say i could do that right this second if i wanted to. it's just a shame. the best experience lately is the cargame server on saturday with league racing also very good.
just the ability to drive a car like the FZR on all the tracks S2 has to offer is worth $50 for me, even if it is mostly against the AI.
to give you an idea i ordered the Milliken "Race Car Vehicle Dynamics" book in order to get the most out of live for speed. the book cost more than the software.
there are some very good points made about how many people you can find online. assuming you are racing in Japan, then like me you are in a timezone without many racers online in convenient hours. at least for my experience i can tell you i had some very good racing online recently.
that is not to say i could do that right this second if i wanted to. it's just a shame. the best experience lately is the cargame server on saturday with league racing also very good.
just the ability to drive a car like the FZR on all the tracks S2 has to offer is worth $50 for me, even if it is mostly against the AI.
hmm it's possible, but you haven't truly convinced me. there has to be some explanation for this trend of easy console games being so popular.
on the other hand you have COD and such that aren't cakewalks for a noob for sure, but even a beginner gets to kill stuff once in a while.
of course this doesn't apply to EVERY american. no insults intended! it's a reflection on what american gaming companies have chosen as their offering based on what they thought would profit the most. i think there may be a bit of a vicious cycle here.
MacFox, your point about timezones being a problem is so true. but it just confirms my speculation. if americans loved challenges, i mean really truly sought out the greatest challenge then they would pick LFS as their (edit: multiplayer) driving sim and we wouldn't have a timezone problem on this continent.
by the way when i say american i'm including canadians in that bunch. we're more of the same, on average.
a few days ago i showed the racing cockpit i built to some friends who were over at my house. neither of them had ever heard of LFS. basically they were stunned by their experience and were very surprised to hear the software only cost $50 and was several years old. they spent hours on the sim and had to tear themselves away when it got late.
basically, the average north american knows nothing of LFS and when introduced to it can find no flaw in the handling or physics. it's as real as it gets.
one friend mentioned the latest forza has a feature where you can rewind time if you screw something up, so you can redo the last few seconds of your driving to get it right. he thought that was pretty sad.
i pointed out that LFS wasn't very popular in north america because the average american just wants to win. they get online in LFS and lose repeatedly to faster europeans, then quit the game and go play gran turismo.
any new releases by scavier aren't going to change this. basically LFS will never be massively popular in the united states unless the US culture changes to one of modesty and honest hard work. there may be business models that will lead to success.
iRacing seems to have done something by becoming a status symbol and by encouraging slow driving. where else can you stay under the limit and win a race? people buy in to iracing because their laptime is within 0.1s of their idol's time on the oval (no limit problem there :P) hell even i can get with a few 0.1% of the wr time on ky1...
basically, the average north american knows nothing of LFS and when introduced to it can find no flaw in the handling or physics. it's as real as it gets.
one friend mentioned the latest forza has a feature where you can rewind time if you screw something up, so you can redo the last few seconds of your driving to get it right. he thought that was pretty sad.
i pointed out that LFS wasn't very popular in north america because the average american just wants to win. they get online in LFS and lose repeatedly to faster europeans, then quit the game and go play gran turismo.
any new releases by scavier aren't going to change this. basically LFS will never be massively popular in the united states unless the US culture changes to one of modesty and honest hard work. there may be business models that will lead to success.
iRacing seems to have done something by becoming a status symbol and by encouraging slow driving. where else can you stay under the limit and win a race? people buy in to iracing because their laptime is within 0.1s of their idol's time on the oval (no limit problem there :P) hell even i can get with a few 0.1% of the wr time on ky1...
by that reckoning, fans should stop going to Talladega real soon, because that's a stale track.
sounds like a problem that makes more sense when coming back to it spontaneously after a pause.
on a side note i believe this may be Scawen's approach to development and the reason why he achieves such elegant results. we can consider two major categories of development style, what my profs called "hacking" and proper analysis. hacking in that sense is making a rapid prototype and then attempting small modifications to it in order to move the behaviour forward in the direction desired. things like looking at a formula and saying "maybe if i put +1 at the end it will give me the right result". in comparison analysis is a deduction from the problem and solution as to what the correct function is and then implementation of that function.
hacking works well and has the illusion of moving forward quickly, but at later stages of a complex project you get diminishing returns sometimes equating to painting yourself into a corner. proper analysis is a nightmare when going into new terrain and development can appear stopped for ages since no temporary or intermediate versions of the function are developed or tried. once the puzzle all clicks into place, though, the result is smooth and worth its weight in gold.
--
reading this thread started a thought experiment in my head about the minimum mathematical description of the driver and environment needed for the ai to decide which inputs were best in a racing application. i made a list of variables that turned out so much longer than i thought at first haha but my intuition tells me there is a clear and elegant solution to this problem. let me sketch what i was thinking without any promises of rigor.
define a circuit as a series of corners
define a corner as these data points:
-a (mostly) straight lead-up vector to the turn-in point
-a vector describing the outside limit of the track near the turn-in point
-an apex point
-a vector for the edge of the track that defines the track-out limit
this part is what makes most sense, the idea being that optimum turn-in and corner exit curves can be calculated based on the expected speed at the turn-in point and the length of track to the following corner. this allows the "mid-point" of the corner to not be in the geometric middle but in the race-optimal middle. example: for a corner you approach at high speed but that has only a short distance following it before the next corner, you want to break later and apex later, so more than the geometric half of the corner is used to turn-in. this allows deduction of a vector starting at the apex and going toward the outside of the corner, say the "corner normal vector".
these points and vectors would be the static track data. adding some uncertainty or error about the exact locations of these points would give a realistic human-like performance to the ai since we describe near-perfect turn-in and corner exit arcs but have to struggle to see exactly where turn-in point and apex are as we bump along the circuit toward a corner.
simplified modifications to these static data would work to respond to "enemy" vehicles ( ;-) ). for example, if another car is on the inside when approaching a corner, the effective apex would be a point one car width along the "corner normal vector" from the static apex.
the only other data needed (aside from the steady-state cornering limit) is about the turn-in and corner exit capabilities of the car at the limit. my gut tells me this is as simple as retaining a factor that describes how rapidly the cornering radius can be tightened on turn-in and another factor describing how much the increase in cornering radius can be restrained under full acceleration on corner exit. this factor would have to be a learned dynamic value to respond to vehicle, setup, tire wear, damage, etc.
the idea is that for a race application you are only interested in the limit. that means that given the static corner data: effective turn-in wall, effective track-out wall, effective apex and corner normal, you can apply turn-in factor to calculate the correct speed at the turn-in point, the turn-in arc as a bezier and the corner-exit arc, and update the throttle steering and brake as progress along those arcs is observed.
if in the middle of following these arcs something unexpected happened, new arcs could be calculated and followed. a stopped car appearing at the corner exit could imply a new effective track-out wall.
by making this minimal model, the points and vectors (and thus arcs) can move in real time and reduce the inputs to the calculations needed to achieve a complex racing line down to a few parameters.
on a side note i believe this may be Scawen's approach to development and the reason why he achieves such elegant results. we can consider two major categories of development style, what my profs called "hacking" and proper analysis. hacking in that sense is making a rapid prototype and then attempting small modifications to it in order to move the behaviour forward in the direction desired. things like looking at a formula and saying "maybe if i put +1 at the end it will give me the right result". in comparison analysis is a deduction from the problem and solution as to what the correct function is and then implementation of that function.
hacking works well and has the illusion of moving forward quickly, but at later stages of a complex project you get diminishing returns sometimes equating to painting yourself into a corner. proper analysis is a nightmare when going into new terrain and development can appear stopped for ages since no temporary or intermediate versions of the function are developed or tried. once the puzzle all clicks into place, though, the result is smooth and worth its weight in gold.
--
reading this thread started a thought experiment in my head about the minimum mathematical description of the driver and environment needed for the ai to decide which inputs were best in a racing application. i made a list of variables that turned out so much longer than i thought at first haha but my intuition tells me there is a clear and elegant solution to this problem. let me sketch what i was thinking without any promises of rigor.
define a circuit as a series of corners
define a corner as these data points:
-a (mostly) straight lead-up vector to the turn-in point
-a vector describing the outside limit of the track near the turn-in point
-an apex point
-a vector for the edge of the track that defines the track-out limit
this part is what makes most sense, the idea being that optimum turn-in and corner exit curves can be calculated based on the expected speed at the turn-in point and the length of track to the following corner. this allows the "mid-point" of the corner to not be in the geometric middle but in the race-optimal middle. example: for a corner you approach at high speed but that has only a short distance following it before the next corner, you want to break later and apex later, so more than the geometric half of the corner is used to turn-in. this allows deduction of a vector starting at the apex and going toward the outside of the corner, say the "corner normal vector".
these points and vectors would be the static track data. adding some uncertainty or error about the exact locations of these points would give a realistic human-like performance to the ai since we describe near-perfect turn-in and corner exit arcs but have to struggle to see exactly where turn-in point and apex are as we bump along the circuit toward a corner.
simplified modifications to these static data would work to respond to "enemy" vehicles ( ;-) ). for example, if another car is on the inside when approaching a corner, the effective apex would be a point one car width along the "corner normal vector" from the static apex.
the only other data needed (aside from the steady-state cornering limit) is about the turn-in and corner exit capabilities of the car at the limit. my gut tells me this is as simple as retaining a factor that describes how rapidly the cornering radius can be tightened on turn-in and another factor describing how much the increase in cornering radius can be restrained under full acceleration on corner exit. this factor would have to be a learned dynamic value to respond to vehicle, setup, tire wear, damage, etc.
the idea is that for a race application you are only interested in the limit. that means that given the static corner data: effective turn-in wall, effective track-out wall, effective apex and corner normal, you can apply turn-in factor to calculate the correct speed at the turn-in point, the turn-in arc as a bezier and the corner-exit arc, and update the throttle steering and brake as progress along those arcs is observed.
if in the middle of following these arcs something unexpected happened, new arcs could be calculated and followed. a stopped car appearing at the corner exit could imply a new effective track-out wall.
by making this minimal model, the points and vectors (and thus arcs) can move in real time and reduce the inputs to the calculations needed to achieve a complex racing line down to a few parameters.
Last edited by CarlLefrancois, .
fact: S2 is so amazing already we just drool when we think of how good S3 could be.
think 3D bezier curves and then you have it...
does anyone really suggest LFS devs owe us something?
Rockingham would be nice as extra content, but really, i spent less than $50 to have an incredible physics and car setup engine and unlimited online play for many years. i can even make my own autocross layouts on all tracks and track configurations.
they did lead us to believe they would work on finishing S2. it's hard to believe they aren't doing that since they keep telling us they are
Rockingham would be nice as extra content, but really, i spent less than $50 to have an incredible physics and car setup engine and unlimited online play for many years. i can even make my own autocross layouts on all tracks and track configurations.
they did lead us to believe they would work on finishing S2. it's hard to believe they aren't doing that since they keep telling us they are
doh! you're right that plot doesn't cover the strangeness in the data mentioned in your article. i'd have to make another plot since i'm a visual type of person and would like to see if it's at least a smooth curve.
or in changing setups. talk about a challenging problem! i believe LFS ai just drives the same for all cars/setups since e.g. during passing a TBO ai driver "changes lanes" so fast he often sends his car into a skid
it must be assumed that rate of lane change is necessary so passing works in the BF1...it's so interesting to read this thread. sounds like a great approach and much fun to ponder so far.
with all due respect where is there room for response to other drivers on the track around the driver?
relying on path points around the circuit then implies either having multiple paths to allow passing which results in two ai cars on side by side rails or having basically a switch statement and two driving modes, one for normal and one for passing.
imo treating the car beside the driver, the apex and the runout limit of the corner all as equivalent points in the environment and calculating a response from that would be more organic.
maybe it's more accurate to say the resistance per square inch on the surface of the balloon decreases. of course this goes against what we commonly know from experience, BUT we don't have access to an infinite variety of balloons with varying rubber thicknesses and compounds.
except that a racing application sees extremes of load sometimes with tires off the ground...
i'm attaching a plot of front vs rear contact patch area per pound of load based on the data in your article. the curves seem reasonable and don't give the impression of totally wrong data.
it's also worth noting that the crossover characteristic of the front and rear contact patch areas is probably designed by the engineers to influence the handling. seeing as it is a formula car and all...
i didn't incorporate the patch pressure data though! does anyone know if available grip is directly proportional to patch area x patch pressure?
FGED GREDG RDFGDR GSFDG