The online racing simulator
iRacing
(13603 posts, closed, started )
When they test on these rigs e.g. CALSPAN, do they get data for over the limit behaviour/sliding, and also that takes into account how quickly or slowly that situation develops etc. So many possible variables, I do wonder about data.
Quote from StableX :yup... my thought is that the NTM currently is about ready to go from Dave's hands out to the beta testers to start getting right and developing before putting out to the paying customers once better.

Too bad the beta testers can't get it right either. Remember the HPD cold tires fiasco?

Speaking of NTM, go do Daytona Road with the MX5 and the HPD. Do a "normal lap" by driving on the banking of the oval. Okay now on your next lap drive on the apron all the way around. Do the same in the Tri-oval if you wish.

And to think the MX5 is going to do a WT event there soon, and the HPD is in the Proto/GT series which is proposing running Daytona the coming season(s).
Quote from Postman Pat :When they test on these rigs e.g. CALSPAN, do they get data for over the limit behaviour/sliding, and also that takes into account how quickly or slowly that situation develops etc. So many possible variables, I do wonder about data.

The guy iRacing contacted to arrange the tests at CALSPAN is a friend of mine. iRacing expressed interest in testing beyond the peak lateral force slip angle. I don't know specifically what was done though because this was a few years ago, before iRacing was released.

CALSPAN can go up to 30 degrees slip angle. What seems likely to me is they might have tested one tire out to that angle at the highest. It's expensive to do this and 30 degrees eats up a tire in a hurry... Anything higher than that would still be a guess on their end.

It's not hard to find out roughly how much drop off iRacing's various tires have in their force curves with some quick skidpad tests. In Atlas watch the vector sum of lateral/longitudinal acceleration at a steady state cornering peak (all lateral), then kick the car out as far sideways as you can, release throttle/brake steering, and see what the percentage difference is.

If the car pushes hard at the limit in steady state cornering, you should be seeing an INCREASE in lateral acceleration when you slide. Instead, on the cars I tested, there's a rather large decrease. I want to say 15-25+%, but don't quote me on that. I remember that on at least two of the cars lateral acceleration during steady state cornering would drop 7% during the first 20 seconds of running due to the rapid rise in tire surface temperature. So to test the drop off you have to get the tires to that point first or else it'll look like there's a lot more drop off then there really is. Circle around for 30 seconds or so before doing the test to get passed most of the temperature effect.

Skip Barber: I can't stand this car now. When the NTM came out on several cars besides the FGT and HPD, the Skip was the best one. They tweaked it the next patch and now I can't bear to drive that thing anymore.
Quote from Postman Pat :But I still say we don't know how much of that is lack of physical forces telling us what's about to happen - only eyes and ears. What *should* that be like to drive?

An hour of so of tweaking tires oneself in a sim would probably convince most people that the role of a lack of physical forces in all this is highly overrated. When you get a bit sideways and do something, say release the throttle, what does the car do? Does it spin, do nothing, or straighten up on its own? That's very easy to see without feeling g-forces. People routinely tweak their setups in sims and feel the differences right away, even very subtle ones. We see the same thing in RC car racing, both real and simulated. Yet there are accelerations involved in neither.

I'd argue that drifting is probably made harder by g-forces rather than the other way around. One of the reasons drifting on slicks might be a little harder than it is on street tires is the accelerations are higher and response times are lower. In a sim the accelerations are missing completely, making this easier to do rather than harder. You wouldn't know that from driving most sims outside of LFS though, where real drivers routinely complain about how difficult the sim version of the car is to catch a slide in. Operating exactly on the slip angle peaks all the time is probably helped a lot by feel, where you can more immediately sense small variations in rotational acceleration for example, but many of the behaviors we're talking about here are well outside this region and are plenty easy to spot without any g-forces whatsoever.

The trouble is a lot of times people will see something like this and comment on it, and even though they're probably dead right, somebody else will dismiss the observation in one fell swoop by simply saying "you can't feel the g-forces," and wins by default. After spending way too much time in my hobby sim over years and years playing with this, and seeing how easy it is to make a car that spins uncontrollably, or self corrects in a slide and won't spin unless you try really hard to do so, or anywhere inbetween, these "lack of real feel" and "it's something in the FFB" arguments irritate me to no end. These people haven't experienced and played with these things themselves at all and really don't have a clue what they're talking about.

Sometimes the g-force argument is just plain silly. If you stomped on the brakes in a straight line and the car took off 1000 feet into the air, would you say "well maybe the real car does that. After all, I can't feel the g-forces?" I see this being just about as ridiculous as that would be to anybody here.
Quote from Postman Pat :This makes no sense if it's a complete pile of shit as you're saying. DaveK's held lap records and won races in the real car. He ought to know himself if it's this far off.

Do we know if Dave does the cars himself? I don't think he does, actually. My impression is that Eric Hudec and maybe one or two other guys do most, if not all, of this.

Dave's one of my heroes so I don't want to bash him too much, but come on.. Even if Dave does this himself, look at GPL. GPL was awesome, but the slip angles in steady cornering were around DOUBLE what they were supposed to be. Some people still hail that as the most realistic sim ever, even with 50% accuracy in probably the most important area of tire behavior. They don't know what they don't know, but thanks to a healthy dose of Male Answer Syndrome, will go on and on about it as though they do, gleefully throwing out the opinions of real drivers in the same cars because they can't feel the g-forces or something like that.

Knowing how the real car feels and then getting the sim version to do that is harder than it seems. I even have troubles doing this in VRC. I know what I want the response of the cars to be, but yet can't seem to get them exactly as I want them to be most of the time. I doubt Dave is much better at that than I am. I wouldn't assume the iRacing version of any particular car handles exactly as Dave would want it to.
Quote from jtw62074 :An hour of so of tweaking tires oneself in a sim would probably convince most people that the role of a lack of physical forces in all this is highly overrated. When you get a bit sideways and do something, say release the throttle, what does the car do? Does it spin, do nothing, or straighten up on its own?

Maybe I'm missing your point. But how can you check how that corresponds with a real car which it has a driver in it making corrections? Obviously there are extremes of behaviour. But real racing drivers (champions) and instructors - I can think of a couple who've also sim raced at a very high level - have said that in RL they make continuous adjustments with their inputs based on feeling the balance and accelerations of the car, that these are completely instinctual and unconscious, and that the eyes wouldn't be fast enough to do this. By the time you see most of it, it's too late. And the fastest drivers, are the ones who can feel and interpret this stuff best.

You can do fast lap in the iRacing skip and not experience any of this sliding and snap-back horribleness as long as you detect the yaw early enough so no sliding. But there are places and situations where it's visually difficult to spot e.g. certain corners where everything is already rotating fast and there's bumps and humps and cambers etc. to throw into the mental calculation. If you could physically feel the changes of balance and rotation, even I a non-RL racing driver, can see it's obvious it'd be completely different.

It could be (to play devil's advocate) that the over the limit behaviour is actually too manageable, and made so in sims for this reason. We take it for granted that that's the way to drive and live there all the time, moaning about difficulties and non-continuities. Whereas, in RL, you'd start crapping yourself the moment it started to feel such situations where about to develop, and correct them fast. I don't know. I do think the cars drive wrong, but I'm just saying most of us are criticising lack of similarity with what we haven't experienced.

I just assumed DK was 'the physics person' and if the cars do wrong stuff like this, then it's his fault.
The physics engine can be really good but if the data that goes in is bad horrible terrible then the final results will always suck. Case in point rfactor. When it came out it was pretty unrealistic. But when people started putting better data in and started using more scientific approach instead of fudging numbers based on feel the physics started improving tons and tons.

Also I'd argue that race driver feedback is pretty unreliable by nature. All of it is based on how human beings use memory. Instead of storing precise information what happens we tend to remember more about the "main storyline", adjust it by our ideals and thinking and then just fill the empty spots to remember more about it later. Basically racecar driver feedback is eyewitness report about the handling of the car and should be treated as such.

Also in car feedback in games through eyes is different than the visual feedback when sitting in car. I don't think you can directly compare those two because the actual environment is so different.
Quote from Hyperactive :
Also in car feedback in games through eyes is different than the visual feedback when sitting in car. I don't think you can directly compare those two because the actual environment is so different.

Well yes, unless you've got at least 3x24" screens, you can't see a f***ing thing through most corners! And then there's the fact, according to research I've seen, that you need a variable FOV to accurately represent speed and distance in a 2D sim environment: high FOV gives more accurate low speed representation (optical flow), and low FOV gives better high speed representation when tested on subjects who can accurately judge and produce speed in real cars. That's one thing - ever noticed you've spun out at around 10mph (what??) and then realised computer says it was something like 30mph?
Quote from Postman Pat :...represent speed and distance in a 2D sim environment

Easily solved

http://www.nvidia.com/object/3d-vision-main.html

iRacing and LFS both work flawlessly with it, and when I first got it a couple years ago I couldn't believe how much it changed my perception of the track, especially elevation changes. I'd rather race in 3D on a 22" screen than in 2D on 3 24" screens any day. Of course, 3D with 3 24" screens would be optimal , and the product does that too (3D Vision Surround)
Quote from Postman Pat :Maybe I'm missing your point. But how can you check how that corresponds with a real car which it has a driver in it making corrections?

A couple ways to check are to watch videos or study vehicle response plots of real vehicles via telemetry, cross checked against simulation. In both you can see what the driver is doing and how the vehicle responds, and why it does so. When a simulation does the exact opposite of what the real car does in a given situation, there's a problem, and nobody needs to feel any g-forces to see that.

Quote :Obviously there are extremes of behaviour. But real racing drivers (champions) and instructors - I can think of a couple who've also sim raced at a very high level - have said that in RL they make continuous adjustments with their inputs based on feeling the balance and accelerations of the car, that these are completely instinctual and unconscious, and that the eyes wouldn't be fast enough to do this. By the time you see most of it, it's too late. And the fastest drivers, are the ones who can feel and interpret this stuff best.

I'm talking about things like you get in iRacing's oval cars, where if you slide the car, the easiest and quickest way to recover is to simply steer into the spin. Most of the iRacing cars will straighten up immediately. Real cars do not do this. If they did, you'd see it in Nascar and other series all the time. Seeing things like this does not require any g-forces to be felt. My example about having somebody sit down and play with tires in a simulation for an hour or so in order to see what drop-off in tire force curves after the peak does was meant to illustrate another example where the g-force argument falls on its face. Some things are obvious without feeling anything. If you turned the wheel and the car did a back flip, you would know immediately something was wrong. Sure, with some g-force feeling you could have anticipated the backflip coming sooner and been more likely to prevent it, but that doesn't mean the backflip is not just fundamentally a wrong reaction. I'm talking about the backflip while other people are talking about how they could anticipate backflips earlier if only they had g-forces like the real drivers have.

For me it's more subtle than this. I see steering into a spin immediately straightening up a car right away, and other similar things, within minutes (sometimes seconds) of trying a new sim. I don't need g-forces to see things like this, and I doubt your racing buddies who were talking about feel would either. Just tell them to slide the car and steer the opposite of what they would normally do and watch their reaction. It'll be priceless.

My point is that people use the "missing g-forces" argument everywhere for everything, even in cases where it really doesn't apply such as in the example of steering into a spin straightening up a car immediately, or countersteering a little bit causing the car to spin out twice as quickly as it would have if you had done nothing with the steering at all, or simply steered a few more degrees into the turn. This type of thing is blantantly obvious visually even without any g-forces or force feedback. I don't think your friends in your example were talking about the same thing I am.


Quote :
You can do fast lap in the iRacing skip and not experience any of this sliding and snap-back horribleness as long as you detect the yaw early enough so no sliding. But there are places and situations where it's visually difficult to spot e.g. certain corners where everything is already rotating fast and there's bumps and humps and cambers etc. to throw into the mental calculation. If you could physically feel the changes of balance and rotation, even I a non-RL racing driver, can see it's obvious it'd be completely different.

That's fine.

Quote :
It could be (to play devil's advocate) that the over the limit behaviour is actually too manageable, and made so in sims for this reason.

I don't buy that for a second. If anything, I'd argue the exact opposite. If the car's "too easy to drive and drift" the quickest way to "fix that problem" is to add a bunch of drop off after the peak. If you really want the car to whip around quickly, add more drop off at the rear than the front, like the FGT and HPD had when the NTM first went on them. Then dial it back a bit when people complain it's too hard and the cars are doing really funny things. This is what Hyperactive was arguing and I agree 100% with that.

I've seen the tire force curves used in rFactor, and in iRacing you can get a pretty good idea what the force curves are doing after the peak quantitatively through the telemetry. Like I said in another post, they come in somewhere between LFS and rFactor. LFS is closest of the three in this area which is why you don't have these problems in LFS, and it's a drifters dream in comparison to any other sim. An unfortunate result of that is that a great deal of simmers think that's wrong and it's too easy to be right. The trouble is, like you seemed to be suggesting, they're comparing it to other sims rather than to real cars.

Most sims these days are quite good up to the limit, so guys that are really fantastic drivers and ride the limit all the time without going very far over it are happy as can be. They're likely to have no idea what I'm talking about. Dancing around right on the slip angle limit is indeed helped by having g-forces and physical feedback. This is what your friends were talking about I think and I agree with that. But passed that where you're sliding a lot, most of the sims don't do a very good job, and in this area you don't need physical feel to see something wrong.

LFS handles this area better than any other sim I'm aware of, besides my own and two others that friends have made that aren't public. The original Hard Drivin' back in the 80's didn't have this problem either, because Doug Milliken knew what tire force curves looked like out in the slidey area when he made it. This was way before GPL.

If you were here at my house I could show you exactly what I mean by playing with force curves and letting you try all these things yourself as I've done probably hundreds of times over the last eleven or twelve years or so. There are vehicle behaviors that you get when the force curves drop off significantly that you do not get when they don't. It's really that simple. And in many cases these are the very same vehicle behaviors people end up having a realism debate over. These types of things don't require g-forces, or even force feedback for that matter, to see. To me they're as obvious as the proverbial back flipping car.

A g-force argument is fine, but it has its place. More often then not it finds its way into arguments that do not require g-forces to settle the issues, which can be seen visually quite easily if one knows what he's looking for.

I can see where some people might have a tough time swallowing all this because they haven't written their own car simulators to play around with these things in and see what does what. Even a lot of people who have written car simulators fail to see connections between many of their inputs and the resulting outputs, as evidenced by just about every sim ever made.
Quote from jtw62074 :
LFS handles this area better than any other sim I'm aware of, besides my own and two others that friends have made that aren't public. The original Hard Drivin' back in the 80's didn't have this problem either, because Doug Milliken knew what tire force curves looked like out in the slidey area when he made it. This was way before GPL.

:jawdrop:

Hard Drivin' was what I cut "sim racing teeth" on. I used to get to that arcade as much as I could to mess with that game when I was very young (dating myself a bit?). I can't believe you even have connections to that game, that's awesome!

I enjoyed Race Drivin' the sequel equally so; getting that corkskrew done right on the stunt track was an accomplishment. Would be fun to get ahold of either game for a keepsake one say. Odd thing is they were never lined up , such is the life of a simracer I guess
I haven't played iRacing in months. So can I just get a quick survey?

on a scale of 1 to 10, 1 being extremely disappointed and 10 being very satisfied.
How happy are you with the NTM?
Quote from jtw62074 :

LFS handles this area better than any other sim I'm aware of


Sorry for butting in to this interesting conversation (not sarcasm), but I'd like to say that as an iRacing subscriber, even after forcing myself as much as I can and allowing others to force me to believe that iRacing is most realistic, LFS still does feel more natural to me when driving it and I do actually enjoy driving on LFS more despite it being less pleasing on the eyes and ears. I use iRacing purely because it now has the community that I wish LFS had and the variety of races that I wish I could take part in, in LFS.

However, I'd like to know your views on Netkar Pro and whereabouts you think that is in comparison, if you have tried it of course. Whenever I drive NkPro the driving feeling always extremely impresses me. It's just lacking in entertainment to be honest, once you grow weary of hot lapping with a cheesy grin on your face that is.
I feel the same way about LFS. It feels that most natural, the car behaves logically and everything just seems to "make sense" dynamically. Obviously the production values are a few generations of ideas behind and the visceral feeling of driving isn't there. ( And the road cars ARE too easy to slide... the tires lack 'bite'. The FBM is a better road car analog when it comes to feeling the load up of the tires leading up to breaking free to oversteer and the snappy consequences. )

NKP and the console SIMs do the visceral feeling thing well.
Is the SRF still on OTM, cause it feels alot better than the others. A slight skid doesn't mean instaspin and horrible death.
Quote from atledreier :Is the SRF still on OTM, cause it feels alot better than the others. A slight skid doesn't mean instaspin and horrible death.

NTM on that one.

I haven't had much time with it but when I took it out for a few laps it seemed much more manageable than it used to be.
Quote from Ball Bearing Turbo ::jawdrop:

Hard Drivin' was what I cut "sim racing teeth" on. I used to get to that arcade as much as I could to mess with that game when I was very young (dating myself a bit?). I can't believe you even have connections to that game, that's awesome!


Haha, thanks..

When Doug Milliken and I were working together on the tire testing for VRC Pro, he told me a lot about Hard Drivin' and the guys that worked on it with him. Doug is listed in the credits only as a "test driver." The way he explained it to me is that since they were the first people to make a video game using a real vehicle and tire model, they didn't want to give any hints as to how it was done. So Doug's name was slipped in at the bottom of the credits as a "test driver" so competitors wouldn't catch on. They apparently never did, either. No other games came out by competitors that were remotely similar.

By the way, have you ever played San Francisco Rush that came out years later? That one to me felt very much like Hard Drivin'/Race Drivin' did and I often wondered if it was done by the same people. Years and years later after meeting Doug, he confirmed that it was done by the same core group of guys.

I think I was 15 when Hard Drivin' came out. I too was absolutely floored by it after a friend dragged me to an arcade insisting I'd love it. That and Race Drivin' are what got me interested in vehicle modelling and got me thinking about physics modelling (aside from the first MS Flight Simulator which got me interested in 3D graphics). Everything I've done since was inspired by this. To "meet" Doug Milliken so many years later in a newsgroup (rec.autos.simulators) and have fascinating talks with him over so many years and find out he was the reason Hard Drivin' and Race Drivin' drove like it did, was wonderful. To later work with him personally and call him a friend has been a privilege.

I learned to drive a stick shift on Hard Drivin' and is really where I learned to drive a car. I learned to slide a car and drive fast on that. It translated directly to real driving too which made it all the more mesmerizing to me. I'd say it launched me into this whole vehicle and physics modelling stuff to a large degree...

And of course, the corkscrew was fun. So was the tunnel which I always had to drive around the roof. Sooo much fun

The arcade my friends went to started with Hard Drivin'. The owner of the place thought it was the coolest thing in the world too and was usually there watching us play and sink endless quarters into it. Eventually Race Drivin' came out. The big, expensive cabinet version. The guy bought two of them and had them linked together. Yep, you could do multiplayer races on it. I remember at 15 or 16 years old (before I could drive a real car) beating a guy at the autocross course that did SCCA racing. My driving style was hugely drifting. It was really the fastest way to go around the last turn. I'd drop it down a gear to get the back end out and just sail around it. It's fun now to actually understand why what I was doing worked. I didn't know back then how I was "supposed to drive." I just did whatever was fast (and fun). Fun times

This is actually when I first started paying attention to yaw velocity and yaw acceleration, although back then of course I didn't know what those terms meant, but the concepts were solidified in my head way back then. At some point I noticed how the horizon moved as the car rotated and started paying a lot of attention to that. Eventually I learned to drift the car and not overcorrect it. Maybe that's partly why now I can jump in a sim and see things immediately that most others seem to never notice. Been paying attention to that kind of thing for 22 years now I guess. Oh dear, has it been that long already?
Quote from MadCat360 :NTM on that one.

I haven't had much time with it but when I took it out for a few laps it seemed much more manageable than it used to be.

Wow. I drove a race and been doing some practice in it, and I really thought it had the OTM, it felt somewhat believeable. I've even had some nice fourwheel drifts and some slides that I could catch. If you get the backend too far out there's no catching it, though. I accredited that to running out of steering lock, but I haven't analyzed what really happens.

Too bad the SRF is an understeery little bugger.
Quote from Ball Bearing Turbo :Easily solved

http://www.nvidia.com/object/3d-vision-main.html

iRacing and LFS both work flawlessly with it, and when I first got it a couple years ago I couldn't believe how much it changed my perception of the track, especially elevation changes. I'd rather race in 3D on a 22" screen than in 2D on 3 24" screens any day. Of course, 3D with 3 24" screens would be optimal , and the product does that too (3D Vision Surround)

So would you say 3d vision using a projector would be a worthwhile upgrade from a triple 2d setup?
That's the route I'm planning to go, I just want to see it first with my own eyes. Everyone on the nVidia forums who has a projector raves about it, I think even Shotglass here likes projectors if memory serves me right. Plus it takes a lot less GPU power - it's already dodgy rendering twice as many frames (one for each eye), you'd need a serious beast to run 3 monitors in 3D. No question, a projector/3D setup would be superior to a triple 2D setup for immersion IMO. The extra dimension adds more than I ever thought it would.

Also, there is no cross-talk when using a projector. I do get a bit on my monitor near the top/bottom if there is highly contrasted objects sometimes.
Please stop talking about upgrading from a triple 2d setup to a 3d projector setup, my wife will literally kill me if I spend any more money on my pc...



Ok, I'll give you the cold shower.

I've tried it on my cinema setup, and I much prefer the triplescreen setup. The 3D doesn't make up for the 35-40 degrees less FOV, imo. I have a 106" screen, around 7' viewing distance, soI get around 85 degrees FOV in my setup. The racing rig give me 130degrees and immersion is much better. The 'larger than life' feel of the projector setup ruin it for me.

You're welcome.

8)
So basically bigger image is better than smaller 3d image? Would you prefer non-3d over 3d on single monitor too or is it just fov/fps issue that makes 3d bad or not good enough for higher res/more monitors?
It's a matter of FOV alone. In my setting I can have EITHER wide FOV (eyefinity, 130deg FOV) OR 3D (projector, 85deg fov). I greatly prefer a wide FOV over 3D. If I could have both I'd take it.

Image size has nothing to do with this. A huge screen seen from farther away can have alot less FOV than a small display up close. Also, the triplescreen display give you some 'free' FOV since the side screens are at an angle.
Quote from atledreier :Ok, I'll give you the cold shower.

I've tried it on my cinema setup, and I much prefer the triplescreen setup. The 3D doesn't make up for the 35-40 degrees less FOV, imo. I have a 106" screen, around 7' viewing distance, soI get around 85 degrees FOV in my setup. The racing rig give me 130degrees and immersion is much better. The 'larger than life' feel of the projector setup ruin it for me.

You're welcome.

8)

Thanks
This thread is closed

iRacing
(13603 posts, closed, started )
FGED GREDG RDFGDR GSFDG