Quote from Shotglass :

id be surprised if netbike models gyro forces at all from what ive seen so far

There is something I don't understand here. What do you need to add to model gyro forces once you have rigid body dynamics (as described in some of these papers http://www.cs.cmu.edu/~baraff/sigcourse/index.html) implemented?

Quote from Shotglass :

but didnt they also say somwhere that the ferrari game would be based entirely on isi and theyre working on their own physics engine for the game after that ?

I'm not sure, but it seems unlikely they would work with two different engines at the same time. There is interview with Ian Bell in the latest AutoSimSport. He said their Falcon project is separate from Ferrari Project (and arcade Kart Attack). There is no info on Falcon on their website yet, I guess that is because they keep information on game concept and content secret as it's supposed to be something very innovative. In the middle of 2006 they were planning that Falcon would take them 2 years, so it may actually be released before Ferrari Project. Falcon for sure is multi-platform project (PC and consoles) and it has multi-threaded engine, so even if it is heavily based on ISI physics engine they had to take a deeper look at ISI code while porting it to consoles and making sure it works in multi-threaded environment. In the process, they might have been tempted to improve the physics engine significantly.

OK OK
RK said Richard Burns Rally is one of his favourite games. He prefers rally simulator over racing simulators because he can race in real life all the time but cannot drive in real rallies. In his opinion RBR is the best rally simulator and one of the best motorsport simulators. RK is rally fan, he likes rallying, and RBR allows to some extent to see how it feels to drive a rally car. Rallying and racing are very different and RK grew up on racing tracks, so he continues racing. He drove in Barborka Rally (in Poland, including Karowa Street SS) three years ago, but now he is not allowed to do it again. He treats simulators as fun and relaxing way of spending time. Although the simulators are developing in a good direction, they will never be as good as driving in real life.

Quote from deggis :

Kubica + RBR this for real and not some stupid rumours?

Here is RK interview text and video (in Polish). He speaks about RBR in it.
http://www.pcworld.pl/news/122239.html

I watched some short (7 laps) AI races in 3 slowest cars on all rallycross configurations, with X33 version. They all used default rallycross setups.

BL2:
- AI make violent turns at the entry to the pits (their path is just strange), as a result they almost lose control and crash sometimes.
- AI in XFG very easily damage their suspension on the jumps. They go into the pits to repair it even when it seems they could continue and finish the short race.

BL2R:
- Suspension damage in XFG, similar to BL2 situation.
- AI in XRG and XFG (less frequently) tend to hit the inside wall of T1.

FE5 and FE5R:
OK

FE6 and FE6R:
- Their behaviour after the race (I've seen it reported before) is probably due to lack of proper garages.

All configurations:
- They sometimes stop after the race but before reaching the pits. I'm not sure if it is caused by running out of fuel. If that's the case then maybe fuel calculations are too optimistic for gravel driving.
- AI are relatively not as quick as they are now on all-tarmac tracks. I guess it's not a bug though.

Last edited by w126, Mon, 17 Dec 2007 21:24.

Quote from March Hare :

There isn't any difference.

Well, one is an inertial reference frame, the other is not.

Quote from hugoluis :

The development team of Live for Speed could be hundreds of people!

Then three people less (LFS Devs) wouldn't make any difference. Why don't you just leave LFS alone, take the remaining hundreds of people and establish your own project? Oh, wait, there already is a racing simulation with hundreds of developers http://forum.racesimcentral.com/showthread.php?t=125970

Quote from CXC_Simulations :

Looks can be very deceiving. Just because one is much bigger and moves to greater ranges does not necessarily mean it will feel more real to you in the cockpit. I realize finding 2 of these simulators near you is going to be difficult, but if you get the chance, let us know what you think.

Rather then looks, my opinion is more a result of reckoning that the seat does not usually move relative to wheel and pedal base in a real car when cornering, accelerating or braking. On the other hand, finding a car that actually does that (seat moves; not that its constructor wanted it) may be easier near me.

I am looking forward to trying any motion simulator, although getting access to only SimConMOTION seems most likely here. Not sure if it would be any indication of the level of experience provided by the more expensive systems.

Quote from CXC_Simulations :

The FD 310 - Plain and simple, it is just another interpretation of a motion system. We're different. I've put many hours in an FD 301 and for me, this approach was better. But, that being said, other people may like their approach better. I think we do it better in a smaller cheaper package.

You must admit you are a little biased here. For me your product is really strange, an inferior solution (wheel and pedals not moving) still in the price range of FD 301, i.e. not affordable for a hobbyist. At the same time a device similar to yours can be built several times cheaper (SimConMOTION).

Quote from Forbin :

However, the reality is, the AI actually drives the same car as you in LFS, with the same physics. Other games cheat in that, as JTbo mentioned, the AI behaves how the track and car makers intend it to, rather than how it should.

Forza Motorsport 2 is similar to LFS in its approach to AI, at least that's what the creators say. http://forzamotorsport.net/new ... ssreports/pitpass36-2.htm

It's very simple:
tyre longitudinal grip = (tyre longitudinal force) / (tyre normal load)
tyre lateral grip = (tyre lateral force) / (tyre normal load)

What you describe is correct (on tarmac at least), it's just load sensitivity, although whether this effect should be stronger or weaker is arguable.

Quote from AndroidXP :

The RAFTyreExtract.exe limited me in my data gathering.

I'll just post my current ('WIP') versions of both programs here, while there is some interest in them. The new RAFTyreExtract should not have the problem mentioned (just checked with over 12 MB RAF). It also works with all 4 wheels and hopefully is more correct then the previous version, though I still know how to improve it further. The new version of MultiDim has two little new features: dimension listing option and double-click centering.

Quote from Niels Heusinkveld :

I think I can get longitudinal and lateral forces, loads.. I suppose you also need to know slipangle and slip ratio?

The longitudinal force is not available in rFactor internals plugin API (http://forum.racesimcentral.co ... php?p=3150984#post3150984). Unless there is another source for it, you would have to try deriving it from lateral force, steering arm force and suspension/steering geometry, but pneumatic trail seems to be also necessary.

Last edited by w126, Wed, 14 Nov 2007 15:40.

Did you try changing the FF strength setting? It seems it should influence the result of the experiments with the wheel catching the slides on its own. When the FF is not strong enough the wheel has problems overcoming its internal resistance and turning quickly enough to keep up. Obviously, this is more likely to happen with higher angle modes because the wheel has to rotate quicker in such cases.

Quote from Dissident :

Here's an idea: create the device as an open helmet which covers the user's head and shoulders.

Whenever you need to apply forces, motors inside the device apply the forces on the user's head using the shoulders as leverage point - you can then simulate front, back, left and right forces.

I was thinking of building such a device and using components of a strong FF joystick. All its electronics could be used and the steering software (or rather a driver or a bridge between a racing sim and a device) would be just a DirectInput application.

Apart from that, we can still hope that some mass market device using galvanic vestibular stimulation will be released one day. http://www.forbes.com/personal ... x_lh_0804remotehuman.html

I think car action in "Death Proof" is nice. However, the dialogs are pretentious and boring. I guess I'm too old.

Quote from Niels Heusinkveld :

I've not been that interested in rFactor after they basically said the force combining is not realistic.

Did they (ISI) say that? Where? Could you post a link?

Quote from ajp71 :

Can any GTL/RACE owners confirm if it looks any different to the previous GP track versions or is it a straight copy and paste?

So far all we have is this single screenshot and it is very similar to GTL track.
http://forum.racesimcentral.co ... =3534394&postcount=59

Real suspensions springs may be progressive (i.e. have slightly more lb/in when being compressed more) instead of being totally linear.

Quote from atledreier :

Let's hope Scawen reads these threads

Do you think he's been getting complete lines of code from Yoda in his sleep?

Quote from StewartFisher :

The selt-aligning torque should be a factor in calculating the feedback torque, but there should also be a component from the product of the longitudinal wheel force and the scrub radius.

There is also another component originating from the lateral force and the mechanical trail, which is usually the most significant one, because the mechanical trail is usually bigger then the pneumatic trail.

Quote from AndroidXP :

While I still can't completely wrap my head around it, one thing I question myself is why nKP and LFS are so different in that regard. Up to the apparent SAT limit they feel quite similar, but then they start to do completely different things. nKP goes limp while LFS doesn't really do anything at all at understeer (if it's not FWD throttle induced). Who is right? Or is neither, and we should get a little bit of easing up but not as exaggerated as in nKP?

Maybe the open wheel cars in nKP have smaller mechanical trail values and the SAT component is more visible as a result?

Quote from Sueycide_FD :

i think the AI in the next stages need to be a little less computer heavy. i can get about 30+ cars in rFactor and a full field in GTR2 with minimum lag but in LFS manage only 6 cars

The AI cars in rFactor and GTR2 use simplified physics simulation, therefore the AI algorithms may be simpler also. Both these things are less computationally intensive than in LFS, where the AI cars are fully simulated physically and the AI has to do the same steering that we do. The approaches of LFS and ISI engine are very different, so IMO their performance shouldn't be compared directly.

http://www.sold-out.co.uk/prod ... =225&productLabelID=1
£4.99 inc. free delivery to UK

Quote from vf1-xj220 :

the main problem here is that EVERYONE wants everything free...

This is not true. It's a market and people evaluate what they get for their money and compare it with other options. The current LFS pricing scheme is what the developers wanted, obviously while staying within the market constraints.

I wish Bugbear would make Rally Trophy 2 instead of these mindless games.