FIA Formula 2 coming up? Hoover spilled the beans :)

I've been a firm believer in a future where people who really know how to do tyres right (which unfortunately seems to exclude at least 95-99.9% of rF1 modders) can do a repository of sorts for different basic types of tyres which can more easily be adapted to mods without having to reinvent the wheel (see what I almost did there?) every time.

But rF2 isn't really a spreadsheet sim anymore, at least not totally. The tyres combine two basic models: A real-time dynamic brush model for the actual surface and contact patch simulation, and a pre-calculated (so yeah, spreadsheet) carcass model for tyre deformation, which naturally affects the contact patch size and shape. The upside of pre-calculating the carcass is that they can do it much with a much higher resolution and make it more accurate in the range it's been designed for. The downside is that the range it's designed for needs to be properly defined to work properly in situations that arise while racing. Though considering that Niels has gone on record saying that the test iteration loop which calculates the values runs for about 4 hours, so the conditions it's calculated should be pretty darn vast, and more accurate than any dynamic system running at say 500 Hz could hope to achieve. Ideally a dynamic system and a static system should produce the same values given the same inputs, so as long as the testing ranges are solid the rF2 system should provide unrivaled accuracy at almost no cost.

If information about turbos is as easy to gather as you imply in your last paragraph then I welcome a very accurate turbo model too. I'm just worried that it's going to be almost impossible for people to do for example the Mitsubishi Evolution X FQ-320, FQ-360 and FQ-400 models, because the main differences are the turbos and information is so damned hard to find.

@LesiU: I don't think that the tyre model is too accurate or complex, but it's slowly starting to push the limits. If you need to run an iteration cycle of 4 hours to see what differences your adjustments made it's going to make tweaking tyres an immense chore. Not to mention that while it's not easy to get good slipcurve data for different tyres, it's next to impossible to actually get some real info on how they're built, with regards to specifics at least. Physically modeling the carcass means you have to know the carcass' properties. How the different layers are arranged, what kind of rubber blends are they using, how thick the plies are, how they're arranged, how many of them are, how stiff the sidewalls are, etc. etc. And if you do make a change, it'll take 4 hours before you can try it (not for every change but at least for structure changes) to see what exactly happened.

Like Niels said:

EDIT: Back to you Turbo, I don't feel that the tyres are "built in" to the physics of cars as such, even in rF1, because I've had great success with using tyres from mods where I know they're good (Niels' Corvette and Supra mods spring to mind) on other cars, basically only editing tyre and rim sizes to match properly, and tweaking pressure settings if necessary.

The real reason the tyres tend to be edited on a car by car basis in many sims is most likely due to the fact that simulation isn't perfect, you can't just plug-in real world numbers from everything and get totally realistic results, there is always an art to it when working with finite computational resources. Another reason is that the necessary data simply isn't available. Slipcurves aren't usually measured NEARLY as comprehensively as we'd need (we'd really need full 360 degree slipcurves for comprehensive results, at varying speeds and loads), tyre structure data is a closely guarded manufacturing secret, and so on. It's easier to get many other measurements from the car, and when all of those match up, you basically have to tune the tyres to make the car work in the simulation like it does in real life.

 

Actually,, the colour red contains (or used to contain) lead, so in reality, red paint makes (made) the car slower.

 

You guys should have given us the rF3 from rF1 instead.

 

I am no coder, nor do I wish to be, but my thoughts on this make sense to me so I will put them out there. If ISI is capable of working out the turbo model to where they achieve the accuracy they want, then a tool to modify the parts we want to modify shouldn't be that tough. The reason I feel this way is because the tools don't need to modify the science behind the turbo, to anything that is directly related to that doesn't need changing.

The tool should only need to worry about values that would normally be changed in real life, such as length of down pipe, diameter of any given section of the turbo, size of the inter-cooler etc. Yes I know that this still leaves quite a bit to mess with, but with a bit of research most modders should be able to comprehend what was on the cars they want to mod. The science of how the turbo works isn't going to change.

Tires on the other hand are quite a bit more difficult to model for sims as there are so many variables that do change on a regular basis. Tire compound, ply material as well as all the rest of the stuff mentioned earlier can all have an effect on how the tire reacts to certain conditions and how long the tire will last. Every tire designed in real life is different from any other in some respects and there are literally thousands of examples out there from street tires to racing tires, and from quite a number of different manufacturers.

My point to this is that if modders are willing to put the effort into trying to get their tires right, then doing turbo's is a piece of cake compared to them.

 

Does anyone have a clue when we will get a car in these cars? They're obviously testing them so soon hopefully?
Think it was about a week after the test with the Clio's?

 

Once ISI have got there work done with the turbo maybe they could do a little piece in the technology section on how they did it.

 

I'm not too 'up' on turbos and how they'll be modelled in the files but has anyone noticed the new parameters in the the Clio's engine file.

 

Yes, few weeks ago...

 

Haven't checked, what parameters would those be?

 

They are for fuelmixture. Depending on wheater(temp, humidity) in Clio you have slightly more or less HP

 

Hopefully they'll also model air pressure at some point, since NA engines will start struggling on high altitude tracks, especially compared to turbos. =)

 

If they start getting into air pressure and air density then I can imagine how fun it would be for a Pike's peak run

 

I'd love to, but that is one dev I really don't want to take time away from.

The track guys are building a track at the moment, and actively taking notes, renders and screens as they go, so that I can use them for a track tech page, and minimize the time I take from them. The tire model and other things like the turbo might have to actually wait until we're out of testing before I can put them together.

 

I believe most important is that it is planned to be done, even for some later point. Some members are also quite good at guessing as they have made mods for SCGT, F1C etc, so some aspects might be figured out by community too as did happen with less moddable products, so no rush, take time needed and make it good, that is what I would like to say, oh I did say it

 

.... update 1094 or 1100 with FIA formula 2 championship mod are at the lobby now. for testing i think.

 

This reminds me from Richard burns rally. Many say how incredible it's physics are, but I remember very clearly maker of those physics saying how they were calculated and then tables were used, can't remember if it was tires, turbos or what (it was at original race sim central if I'm not mistaken), but similar way there was software that calculated situations and tables were used realtime for performance reasons.

With enough accuracy, ie. small steps between table cells, one really can't tell the difference, but also when enough large table is made, finding correct spot from table because taxing. With todays hardware however I believe resolution of table can be so good that error marginal is very small, probably more error comes from delays of monitor, wheel etc.

I guess it is bit similar to 3d models, one can model even pistons to engine, but does it make better experience to know that there are actually pistons in engine that is still static 3d model?

 

Tables are going to be faster in almost all circumstances than calculating things dynamically, but they do have the problem that the table needs to be designed properly so that you don't run into situations where the values aren't calculated properly.

But yeah, if you have a dynamic system and a table based system done with the same resolution and the table is properly defined, they should produce identical results. With the table based approach though you can go for higher accuracy without destroying performance.

 

Hi guys don't mean to sound pushy or anything just wondering roughly when we think we will get to drive the Formula 2 car. It's probably the car I am most exited about so what do we think a week, a month, 2 months? Thanks

 

About half year range. Beginning of 2013 or Christmas of 2012.

 

Fairly soon. If everything went to plan it would be week(s).

I went to bed (or rather laid in bed, unable to sleep) last night with numbers pouring through my head, as I worked out the stats for the AI last night. Basically so that you will have the right people at the front in your races against the AI.

I know some might figure it's a tease, but I hope it'll be taken as a catalyst for genuine discussion: The F2 is actually close to being my favorite car in sim racing, which surprised me a lot, because it has wings. It is probably the perfect balance for me between speed and downforce. I'm someone that normally prefers low/no downforce cars. With the public specs there are, any debate about why that is? I certainly wouldn't have expected so...