Is this from a skidpad test because its much more focused? Could you show the longitudinal drag Force in relation to lateral tire force (Flon/Flat) over slip angle please?
@mechanic Yes modding is good hope source. It is so awesome that rF2 is moddable, possibly the last one of this kind. In my personal (still rather amateurish) physics modding experience and knowledge rF2 (despite few points in question) allows very satisfying and beautiful possibilities in recreating RL vehicle handling. Aero and tires are most tricky as they should be, but rF2 allow lots of controll to get most 4wheel vehicles to behave principally correct in the most part of whole picture for what we know that is simulated in rF2.
Fascinating conversation here. I agree that the tires need to be seriously looked at, whether it's air-pressure, heat related, or something else. Also, it seems a bit disingenuous (in simulation terms) to not have a proper physical drive-train model in place. It is after all, a very fundamental aspect of the vehicles being simulated. In some cases, drive-train behavior can impact the driving dynamics in significant ways. Just a quick point regarding the slow pace of updates. Yes, rF2 progress is slow, especially when considering some long standing issues but, consider that iRacing has not improved their ffb significantly from the start, despite many requests and having much, much better funding. It's been reported that improving iRacing ffb would require a major re-write of the code with potentially, adverse affect on other aspects of the simulation. It seems some issues are much harder to solve than we realize otherwise, they would already be solved. I imagine the same is true for the car physics, drive-train and much more. Still, the sooner such issues are addressed, the sooner the title can reach it's best "Simulation Value".
When posting in a particular sim forum, my policy is not to confirm the name of another sim for which I am showing comparative telemetry. That's because the plots will sometimes show something negative about the other sim and I prefer to be fair to the other developer. I happily post the telemetry in their own forum and make them aware of a problem when I think I have found one. Race vehicle simulation is a very hard software engineering challenge and I think all developers need all the testing help they can get. Many struggle to get sufficient beta testing participation, for example. So I try to play a small part in helping them, even though I don't participate formally in alpha or beta testing for any sim. I'm interested in too wide a range of sims and the spare time I have for this hobby is too limited to commit to anything like that. I am happy to talk in general terms about the relative strengths and weaknesses of particular sims I have sufficient experience with, hence my comments about iRacing specifically. I have spent a lot of time studying the physics of that sim and providing IRacing with feedback about physics issues I think I have found, including something resembling what's being discussed in this thread. Yes, very happy to provide details of the functions used for the plot I have shown. See the attached text file. I have tried to provide sufficient information to replicate my implementation, which is for McLaren ATLAS Express (my preferred telemetry analysis tool), in other tools such as Motec. ATLAS has a tough learning curve, but it is very powerful. I also develop custom telemetry tools using C++ and Python, for example. The function content documented in the text file is part of a bigger project to develop a telemetry toolset for multiple current-generation sims. Using each sim's telemetry API, I write a telemetry capture utility which writes to a generic CSV file format. The file format preserves channel name compatibility with iRacing (for common channels supported by the other sims' telemetry APIs). This enables me to use workbooks originally developed for use in iRacing for its native binary telemetry (IBT) files with all the other sims as well, using the CSV capability of ATLAS. I am aiming to support four other current generation sims, additional to iRacing. I am currently developing support for rFactor 2. Apologies, missed the Car_Yaw_Angle function, so re-attaching. I should also mention that my toolset uses the iRacing vehicle axis system convention: X = forward, Y = left, Z = up, a right-handed coordinate system. For other sims, I translate to that system, so that I can easily compare test results across sims. The rF2 axis system is: X = left, Y = up, Z = back, a left-handed coordinate system. One challenge when building this sort of toolset is that some sims do not provide the X, Y and Z components of the vehicle velocity vector expressed in the vehicle (local) coordinate system, but only in the world coordinate system. That's sometimes a problem, because the local velocities are required for computation of slip angle estimates for sims that do not report slip angle data directly; iRacing is one such sim. However, if the sim provides channels for the Euler angles (Pitch, Roll and Yaw) they can be used to derive the local velocities. I've attached an example C++ function implementation for conversion of world vectors to local vectors using the general rotation matrix. Alternatively, if the sim telemetry API reports the vehicle orientation matrix, the conversion can be achieved directly - the sample function effectively builds the general rotation matrix from the known Euler angles. For more details on this, see this web page: https://en.wikipedia.org/wiki/Rotation_matrix#General_rotations
Interesting, I am myself a (very lazy) dev and of course noticed as well that most of the tools I can create could be adapted to almost any simulation given they provide an API, all i have to do is make my tools "plugin aware" and everything runs fine, just create each sim's plugin to be compatible (not to mention when the simulation changes its API, it's much easier to maintain). Maybe we should discuss further this in private discussion, I'm pretty sure most ppl here aren't interested in all this stuff. Math channels (ATLAS or MOTEC) in another hand are surely way more interesting for the vast majority of us. There are several math channels I've learned here (thanx to @Lazza for example) that helped me a lot.
Here's another example plot from the same sim and car as the first plot. This isn't showing a controlled dynamic scenario, but it's focussing on tyre physics and contrasting a sequence of corners driven in two laps: - On the left I took too much kerb in one of the corners, unsettling the car dramatically. With frantic countersteering I was able to save the car (just). In most cars in iRacing, recovery from that degree of dynamic upset would be extremely difficult and often impossible. The comparison sim is generally recognised as having very good countersteering responses, though some think it is too forgiving. My personal view, from countersteering in real life in fast RWD road cars and karts, is that it is closer to reality than many sims. - On the right, there is a normal transit through the corners. This sim provides a lot of telemetry channels, but overall rF2 provides more, assuming you can get all the required sensors enabled. From top to bottom, the plot areas are: - Slip angles (solid colours) and steering wheel angle (dotted orange). - Slip/Grip Fraction. Proportion of the tyre's cornering potential that is being exploited (0-255 at the time I plotted this, since converted to float 0-1 in my toolset). - Touching. Whether the tyre is in contact with the track surface (boolean, 0 or 1). - Tyre lateral forces. - Tyre vertical loads. The car doesn't have very sophisticated suspension, so there is a lot of wheel hop visible. Conversely, it's a very enjoyable and challenging drive.
Yes, you are surely right, but I write primarily for those who are already interested or becoming interested, and those who understand the value of telemetry and the scientific method. Happy to take PMs if you prefer for anything specific. @Lazza got me steered in the right direction with my rF2 telemetry capture implementation on several points, e.g. the need to enable sensors for some cars.
Telemetry is indeed the only tool that can describe minutely what we see in movement, then it still must be understood and interpreted, but at least remove most of the human factor. Vision lie to us all the time, as each magician well know, yet even numbers can lie, but is much harder.
Everything we see, hear, feel, touch, etc... is interpreted by our brain, which "fills in" the blanks with "shortcuts", basically our brain "lies" to us all the time, indeed.
i recently tried porsche RSR on iracing.I chose rsr specifically because i own it in rf2 and have good experience and muscle memory of rf2 RSR. First of all the excessive steering input is much better punished you get tons more understeer compared to rf2. Power on understeer also seems more "natural" there. in rf2 even with small throttle input cars became very yaw friendly. This doesn't mean i couldn't go sideways in iracing, i could but, it required more throttle,more aggression even though the setup at iracing is quite oversteery. speaking of setup the RSR in iracing had very very different setup philosophy. the rear springs rates are much higher than front and rake is way more than rf2 version. Rake difference seems very interesting. in rf2 12mm rake (while front is at lowest) seems very good balance wise. in iracing i was running over 30mm rake ! (while front is at lowest) that is huge difference and car was totally drivable. I tried different base setup but the high rake trend in iracing seems to be always present. Why so much setup difference when both cars are same ? I also tried to drift the car in iracing with tc off, it is believable UPTO a very small slip then after that threshold is reach even 5-10% throttle is enough to completely spin out. The drift in iracing doesn't maintain forward momentum.
Irating is totally worthless as a comparison in my book, at least for the physics department, while I can give a 9/10 on the competition aspect, I can't digest the fact that the car is on rail in a turn and the next lap you enter a bit faster (because you were on rails before) and you became a passenger of the car, unable to regain grip after interminable time, even if somehow you managed to remain on track, the grip does just don't come back. I even won a race in Iracing, will never come back.
I'm quite surprised nobody seemed to have linked Nicki Thiim's onboard lap of last year's Le Mans in this thread, considering the fact that I feel the GT handling is in the core of the issue. Purely looking at how he handles the car, I'd say his driving resembles how I drive in ACC more than how I drive in rF2. I always felt rF2's GT cars to feel too light.
Same. i also feel that GT cars in RF2 are quite "throwable" they seem so light. they lack momentum feel like ACC has. I was testing on test track it had hotwheel like structure->> (https://i.pinimg.com/originals/dc/98/9c/dc989cef432b46ee8620ff21a8e32a78.jpg) when my car fell from top the gravity seemed very weak. it fell to the ground so slowly. Could this be issue that RF2 gravity is weak ?
https://twitter.com/i/status/1355153728513437696 Last few seconds from this video. It looks like an R/C car. Or if the car has no weight. I don't believe this is possible in real life.
I will try to remember that. Recording isn't a problem. We need to know the distance it traveled from top to ground.record the time it took to cover that distance. With some math we can figure out gravity strength. With that said i don't know i'm able to find all this parameters/data. Plot twist: I'm spinelli
I'm no physicist, but I'd think the front would grip less with weak gravity. Lower gravity = lower friction = lower grip. Could be wrong. GT3/GTE cars in rF2 seem to have very low forward momentum when cornering, almost as if you're driving a 500kg car instead of a 1200-1300kg car.
If you want to get an idea of how light small car drives, you can take for a drive "not" Costin Nathan from 1967 endurance carpack (480kg with driver). I also recently watched this awesome video. Thats ~500kg car, with probably no downforce, maybe even with lift. As you can see with light car it's a lot easier to keep the momentum, also you can see how much less inertia it has. And it still has fair amount of grip even if it weights just that much and without aero downforce. I have heard people trying to explain physics in various simulations by using "feel of weight", or "weight transfer". And it usually is rather abstract, not explaining anything and mostly also false guesses/sensations. Last but not least gravity defines weight. rF2 has weight reading in garage of how much wheels are loaded (at least in the old UI). The rate of falling would be slightly different thing, it is not too common that race cars are falling down, and if they does, then aerodynamics are very much in effect. P.S. that FormulaE stuff was weird, but also must take in account that there sometimes seems to be broadcasting artifacts of how cars move, as broadcasting seems to be in desync with physics, sometimes even replays does appear to be suspicious as if not looking exactly how expected it would based on feel during real time driving.