Poll: rF1 vs. rF2 tyre-model

[lordpantsington]

What was the reason why ISI changed the tire model? What specifically did changing the model address? Has this ever been disclosed publicly?

For those pushing for the rf1 model, how many tires have you made in rf1?

 

[Lazza]

I think that's going to be impossible, rF2 tyre model is about tyre construction, wheras rF1 tyre model is a smart mathematical approach where you can match data rather easily, as you have direct control over the output.

I personally don't mind if rF1 TM can be used, however I wonder if it's really worth. Plus how's it going to work under changeable weather conditions?

I don't see any reason the rF1 model would be especially limited in relation to weather. The rF2 model, at the moment, only has simplified weather dependency itself.

Easy to overplay how 'easy' it is to tune the rF1 tyres to fit data too, but they are certainly more output/data based.

I would say not good idea. I'd say make it easier to bring rf1 tire data and use it for rf2. Let's move forward, not backwards.

The rF2 model is based on construction, which then produces behaviour. The rF1 model is much closer to being behaviour based. You can't automagically go from behaviour to construction, which is part of what makes the rF2 model more challenging to mod in the first place. If you want to use rF1 data in rF2, you need the old model enabled.

What was the reason why ISI changed the tire model? What specifically did changing the model address? Has this ever been disclosed publicly?

For those pushing for the rf1 model, how many tires have you made in rf1?

From a practical viewpoint it allows the game to model the behaviour across many many many scenarios, whereas rF1 gives you one shot at defining behaviour and uses that the whole time. All tyres, once defined, change in the same way in each scenario.

 

[Domi]

I think both (all) models have similar issues at the end... for example physical tyre models like rF2's one can model the behaviour in more scenarios, however, how do we know that their output matches the real tyre?

 

[sdancourt]

I think that go back to rf1 tire model because Modders troubles with rf2 ones is like killing the patient to overcome the desease. I guess ISI can, slowly if they want, build a kind of tire library open to anybody so modders can access and do their job, creating cars and mods. How many tire types works on rf1? The heart of RF2 is its outstanding feedback from the tire model. Killing it will be the end of the best Sim since GPL.

 

[Saabjock]

I don't know the first thing about modding.
I do know about cars and car handling.
If the original RF tires are used in AMS as a base, then it's not such a bad idea to merge them in RF2 to help modders.
That sim feel absolutely the most 'natural' of any out-of-the-box on cold tires.
Come to think of it...it feels great on ALL tires.

 

[blakboks]

I'm don't know the first thing about modding.
I do know about cars and car handling.
If the original RF tires are used in AMS as a base, then it's not such a bad idea to merge them in RF2 to help modders.
That sim feel absolutely the most 'natural' of any out-of-the-box on cold tires.
Come to think of it...it feels great on ALL tires.

Right?! I don't know the credentials of most of the people responding to this thread. But it sure 'seems' like we have a bunch of people who don't know anything about modding, and they're seeing the number '2' and assuming it's better. I'm curious if people really would notice the difference in a blind test; and be able to pick the right one.

Seems like people are a lot like this:

 

[stonec]

What was the reason why ISI changed the tire model? What specifically did changing the model address? Has this ever been disclosed publicly?

For those pushing for the rf1 model, how many tires have you made in rf1?

With a physical tire model as in rF2, it allows more detailed simulation of certain aspects as you're not modeling the behaviour but the whole tire itself. Things like flatspots and tire deformation at any place of the tire would be very difficult to implement using an rF1 type model, because deformation and flatspots are so closely related to the tire structure itself.

Almost anyone with basic modding knowledge can work with the rF1 tire model, it doesn't require you to "make" tires as such, you can get started just by adjusting a couple of parameters and a few slip curves. Basically plugging numbers into equations.

 

[pitkin]

Right?! I don't know the credentials of most of the people responding to this thread. But it sure 'seems' like we have a bunch of people who don't know anything about modding, and they're seeing the number '2' and assuming it's better. I'm curious if people really would notice the difference in a blind test; and be able to pick the right one.

A blind test would be interesting. Some rF2 cars feel brilliant, some rF1 cars feel brilliant, I doubt the average casual sim racer could tell the difference most of the time.

 

[Guimengo]

A blind test would be interesting. Some rF2 cars feel brilliant, some rF1 cars feel brilliant, I doubt the average casual sim racer could tell the difference most of the time.

Most interpret the FFB as the actual representation of a car's physics and parameters. Without checking ingame HUD for data, just from driving a rF2 car slotted with either rF1 or rF2 tires, I doubt more than an extremely small population would have any idea on telling them apart.


FYI I do agree with Wiggin.

 

[gpfan]

Let´s keep the current model! All we need is good documentation and user friendly modding tools thats all!!!!

 

[mschreiner]

Remember everyone the New tire model is still being developed and tweaked by the Devs.
the rF1 tire model still is a very good tire model to use as a temporary model in my opinion until there are more new tires available to use for modders. Right now it can take up to 24 hours just to calculate the model

 

[Lazza]

For the 600th time, no one is suggesting the new model be dropped.

If both were available, some mods would be made with the old one. And some people would download a mod, find it uses the old model, and stop there. Flatspotting would be a giveaway if nothing else.

People used to the rF1 model might try an rF2 mod and notice ways in which it's better, more realistic, less tolerant of abuse, as I think most of us agree.

So generally there will be a push to the rF2 model. But, if a modder (or group) can't find/develop an rF2 tyre that works as well in their mod as an rF1 tyre they have, using the old tyre allows them to release the mod. Anyone willing is then able to use it, instead of staying away from rF2 completely and their fanbase being under the impression rF2 is 'impossible' to mod properly and somehow see the rF1 model as superior because of it. (a mismatched or poorly tweaked rF2 tyre will feel worse than a good rF1 tyre)

There's no negative to this. We won't have all mods with rF1 tyres just because it's easier to make - it's worth finding or making suitable rF2 tyres (for the modder) because they are better, which is why people here are worried about the rF1 model making an appearance in the first place. Demand will shape development, as always.

 

[woochoo]

From Marcel's interview on RD a few days ago, I think he was suggesting that enabling the rF1 model alongside the rF2 model would be a good way to get some reluctant modders to at least engage with rF2.
It seems the rF2 model is a barrier to entry for a number of people, and removing that barrier (not the model) sounds like a good thing to me.

Once a car is in rF2 with the 'easy' rF1 tyre model, then the modder might feel more confident investing time into learning the rF2 model.
Having to learn the rF2 model just to get a car to work in rF2 may be too daunting of a task for some (such as myself, but I've never even worked with the rF1 model either). Having their car work with the rF1 model is a good basis to move to the rF2 model. A stepped approach to development can be a good thing.

I think Marcel also suggested that the process of modders converting their cars from the rF1 model to rF2 model could be the basis for working out the requirements of a tool to make the construction of an rF2 tyre easier. A tool to convert an rF1 tyre to rF2 is what a lot of people are asking for. So if that turned out to be the case, giving the game compatibility with the rF1 AND rF2 tyre models could eventually make the creation of rF2 tyres easier.

And, I don't think there would be any reason why some basic tyre lockup/flatspot simulation couldn't be added to an upgraded rF1 model ('rF1+' model??). As well as grass/dirt pickup? Does Reiza have this? I think 397 are thinking about adding pickup support to the rF2 model (?) I'm no programmer, but I can imagine it would be possible to make an rF1+ tyre model a parallel, yet simpler, version of the rF2 model. With that in mind, it could be possible to race against a car with the other model, and there be no major 'cheating' advantage.

tl;dr
- rF1 model won't be replacing rF2 model in game. They would both be available.
- rF2 tyre model could be a barrier to entry for some modders (difficult to understand without prior knowledge), and removing that barrier would be a good place to start for those modders.
- Modders that add a car with rF1 model could then work on making rF2 model, and that learning/comparison process could provide knowledge to help 397 develop a rF1 -> rF2 tyre tool (probably not one-click, but I'm sure quite helpful).
- Perhaps 397 could add basic support for flatspots and pickup to the rf1 tyre model (rF1+ ?), to balance competition between rF1+ and rF2 tyres.

RaceDepartment interview with Marcel of 397:
part 1 - http://www.racedepartment.com/threa...studio-397-rfactor-2-interview-part-1.126534/
part 2 - http://www.racedepartment.com/threa...studio-397-rfactor-2-interview-part-2.126535/

 

[peterchen]

I think for the modders who throws out a mod with rF1-model, studio 397 could go and suuport them by doing a rF2-model for them.
That would make sense and would be a good way to help modders around.

 

[David Wright]

I am familiar with the rF1 tyre model. It models the tyre in terms of its grip characteristics (e.g. slip curve shape, load sensitivity, peak slip angle with load) which the modder defines. My understanding of the rF2 model is models the tyre in terms of its physical construction (e.g. ply material, density and angle of plies, rubber thickness and properties, etc). I think its unwise to assume that its possible to make a tool than converts an rF1 tyre to an rF2 tyre. If there was I think Studio 397 would propose this rather than suggesting the rF1 tyre model to be an option for player physics.

And while I'm sure more support for modders is possible, I'm not sure Studio 397 would be in a position to know the tyre construction properties for every or even most mods.

 

[woochoo]

I think its unwise to assume that its possible to make a tool than converts an rF1 tyre to an rF2 tyre. If there was I think Studio 397 would propose this rather than suggesting the rF1 tyre model to be an option for player physics.

I didn't mean to suggest it would be a magic tool to do a complete conversion, but maybe I accidentally did.
Perhaps it would be extremely difficult (or even more difficult than that...) to 'convert' tyres with a program. Maybe 397 could produce some analysis and tuning tools, or something that can produce a starter for an rF2 tyre with the input of a few standard/known/basic construction characteristics, and then add the rF1 data, and then produce a good starting point for further development of an rF2 tyre. Maybe that could be more like an 'rF2-lite' tyre model, more than an 'rF1+' model.

It might be that the best they can do is produce a tool that makes it easy to compare the performance characteristics of a car makers existing rF1 tyres with rF2 tyres. That could be comparison of performance graphs in a separate application, or an instant A/B test function in dev mode, or a combination of both. Or something else

In any case, in the interview Marcel said it's a process that hasn't started yet, so I guess all we can do at the moment is speculate for our own amusement.
My post is full of 'perhaps' and 'maybe', but i can confirm, for sure, from an inside source, that i have amused myself. Nothing more to it than that

Part 2 of the above linked interview is where it's discussed, for anyone who hasn't read it yet

 

[Lazza]

That was certainly a part of what made the rF2 model seem fairly inaccessible when it first appeared, if you were used to the rF1 tyre files - ttool lets you see the tyre geometry, what it does with various loads and cambers, but unless I missed something (or something has changed more recently) you can't easily run a few tests to see how it behaves in terms of the rF1 reference (slip curves etc). The hours building a tyre (often quoted as 24 hours, but that's a full suite you probably wouldn't need to do until you're just about finished) would be easier to put up with if you could more easily run some representative tests that give you an idea how the tyre behaves.

You could imagine ttool (or another tool if needed) being able to run some some scripted performance tests on rF1 and rF2 tyres to help modders match the rF2 tyre to an existing old model tyre (or some tyre data they might have on hand, in the absence of tyre construction data).

But for sure a 'conversion tool' wouldn't really work, because going from rF1 to rF2 is actually working backwards.

 

[woochoo]

The hours building a tyre (often quoted as 24 hours, but that's a full suite you probably wouldn't need to do until you're just about finished) would be easier to put up with if you could more easily run some representative tests that give you an idea how the tyre behaves.

Maybe Luminis could rig up a cloud-based number cruncher for people to compile their tyres.
It seems they're all about the cloud, but I'm not sure if that includes cloud processing

 

[D.Painter]

I'd consider myself as a modder. Being a modder doesn't mean I know about physics.
Those out there that release a car mod that they've built the car, done all the textures, Done the in-game and also think they've done the physics are kidding themselves!
3D Modeling.
Building good 3D models are a art. One that takes a long time to get right. Remember I said "good" models. There are a lot of crap out there. In fact as bad as it is there's some far better models in rF1 then there are being built for rF2. This takes dedication and years to learn.

Texturing.
This is also a art. Again one that takes a long time to perfect. Not to say perfection is truly achievable but it's not hard to pick the good over the bad.
People that are good at this are also dedicated to only this. Of cause this means all aspects from the drivers seat to the liveries on the outside.

In-game.
This is seen as a simple job. Most guy's that do this are quite quick at it. Does this mean they've done a good job? No! I've been working on mods and around mod groups sins SimBin GT-R and found this job alone requires a far more thought out approach then we see from to many in-game people. Someone who has a good eye. A knowledge of texturing and most impotently a good knowledge of how Alpha channel works! So many don't have a clue about this but will argue they know it all. Of cause the proof is in the result.

This comes to the point.
PHYSICS>
All physics concerning the car. We see people saying this isn't hard. The reality is if you think it's easy strait of the bat your kidding yourself.
Those that are good at this are dedicated to it! Spent years working on it.
They can't built 3D models. They can't or are average at texturing. And have some knowledge of in'game.
These people are the ones and only ones I'd listen to when talking about physics. The rest of you are just standing in the shallow end of the pool and "think" you know it all.

For me, Physics are over my head. I understand it in layman's terms and am not stupid enough to think anything else. Being called a modder doesn't mean you know it all and can do it all.
Why do you think we go into modding and call it a "Team"? It takes a team to do it right. Anything that's done by one guy and release isn't worth the download.

Tyre Physics.
I was shown in layman's terms by a dedicated physics guy how and what is required to build a rFactor2 tyre model tyre.
Now remembering that I was totally against going back to the rF1 tyre model.

You need to know how the real tyre is constructed. (No tyre company is going to give you this)
You need to build the 3D tyre model match the verts with the construction of the real tyre so the deform works properly.
You need to know heat transfer values. Hot to cold.
How the heat transfer gose to other component around it, brake disc, rims and so on.
Heat transfer from the brakes to the tyre.
All values you need to know.
Tyre pressure, How this affects the heat in the tyre and heat transfer to other components.
Ones you have all this in place you need to allow a program to bake the tyre for you. This process can and does take hours.

Then you test it.
If it's wrong you start again.

There's a lot more to it then that of cause but I for one understand why modders, Physics guy's are finding the rF2 tyre model so difficult to deal with.
The first hurdle is getting the tyres real world construction. This alone is impossible in most cases so a brick wall before you even start.


"If" rF1 method can be improved and worked with by the dedicated physics people out there then I don't see we really have a choice.
Rieza have done it. Used the rF1 model as a base so with time and people much smarter then me in this field I'm sure rF2 can be still seen as how the 2 imply.

End of the day. I've changed me view on it. If going back to rF1 tyre method will help physics guy's to get more mods into rF2 I'm for it.

 

[Lgel]

I'd consider myself as a modder. Being a modder doesn't mean I know about physics.
Those out there that release a car mod that they've built the car, done all the textures, Done the in-game and also think they've done the physics are kidding themselves!
3D Modeling.
Building good 3D models are a art. One that takes a long time to get right. Remember I said "good" models. There are a lot of crap out there. In fact as bad as it is there's some far better models in rF1 then there are being built for rF2. This takes dedication and years to learn.

Texturing.
This is also a art. Again one that takes a long time to perfect. Not to say perfection is truly achievable but it's not hard to pick the good over the bad.
People that are good at this are also dedicated to only this. Of cause this means all aspects from the drivers seat to the liveries on the outside.

In-game.
This is seen as a simple job. Most guy's that do this are quite quick at it. Does this mean they've done a good job? No! I've been working on mods and around mod groups sins SimBin GT-R and found this job alone requires a far more thought out approach then we see from to many in-game people. Someone who has a good eye. A knowledge of texturing and most impotently a good knowledge of how Alpha channel works! So many don't have a clue about this but will argue they know it all. Of cause the proof is in the result.

This comes to the point.
PHYSICS>
All physics concerning the car. We see people saying this isn't hard. The reality is if you think it's easy strait of the bat your kidding yourself.
Those that are good at this are dedicated to it! Spent years working on it.
They can't built 3D models. They can't or are average at texturing. And have some knowledge of in'game.
These people are the ones and only ones I'd listen to when talking about physics. The rest of you are just standing in the shallow end of the pool and "think" you know it all.

For me, Physics are over my head. I understand it in layman's terms and am not stupid enough to think anything else. Being called a modder doesn't mean you know it all and can do it all.
Why do you think we go into modding and call it a "Team"? It takes a team to do it right. Anything that's done by one guy and release isn't worth the download.

Tyre Physics.
I was shown in layman's terms by a dedicated physics guy how and what is required to build a rFactor2 tyre model tyre.
Now remembering that I was totally against going back to the rF1 tyre model.

You need to know how the real tyre is constructed. (No tyre company is going to give you this)
You need to build the 3D tyre model match the verts with the construction of the real tyre so the deform works properly.
You need to know heat transfer values. Hot to cold.
How the heat transfer gose to other component around it, brake disc, rims and so on.
Heat transfer from the brakes to the tyre.
All values you need to know.
Tyre pressure, How this affects the heat in the tyre and heat transfer to other components.
Ones you have all this in place you need to allow a program to bake the tyre for you. This process can and does take hours.

Then you test it.
If it's wrong you start again.

There's a lot more to it then that of cause but I for one understand why modders, Physics guy's are finding the rF2 tyre model so difficult to deal with.
The first hurdle is getting the tyres real world construction. This alone is impossible in most cases so a brick wall before you even start.


"If" rF1 method can be improved and worked with by the dedicated physics people out there then I don't see we really have a choice.
Rieza have done it. Used the rF1 model as a base so with time and people much smarter then me in this field I'm sure rF2 can be still seen as how the 2 imply.

End of the day. I've changed me view on it. If going back to rF1 tyre method will help physics guy's to get more mods into rF2 I'm for it.

I fully agree with your post.

The physics of RF2 require the colaboration of a team of experts mechanical engineers, one of them being a tire specialist. Those guys are not easy to find, and if they are good at it, they have or quickly find, a very demanding full time job in real life and quit modding.
It also requires obtaining very detailed data of a real life team that is not available just by browsing on Internet.

Cheers.