rFactor 2 | New Build Update and Indianapolis PBR Released

[Paul Jeffrey]

Studio 397 have today released a new build to the rFactor 2 racing simulation, alongside a new PBR shader update and refresh of the popular Indianapolis Motor Speedway circuit within the sim.

• Indianapolis 2020 V1.06 released.
• New build update addresses the first pass of fixes and optimizations.
• Build update available as opt in release candidate.

Studio 397 have today released both a new build update to the core rFactor 2 simulation, alongside the much anticipated refresh of the classic Indianapolis Motor Speedway Steam Workshop circuit for the title.

In terms of the build update, the new release today adds a number of fixes and quality of life improvements to the title - the most recent build in what the Dutch development team have promised is the start of a concerted effort to address many of the outstanding issues within the simulation.

Build Update Notes

• We’ve removed the Max 2012 plugins as they no longer work, and provided the last working Max 2017 plugins. We have no intention to upgrade them any longer as we can’t get these versions of Max any more to test. Furthermore, we urge the modding community to upgrade to current Max versions.
• Fixed: Reflection mapper in Scene Viewer, so it renders from the same point as the user sees.
• Fixed: Bind a mouse click to any input control.
• Fixed: Issues with showroom performance.
• Fixed: Missing error prompt when textures are missing from content.
• Fixed: Issue with a white screen you get when you try to assign a control.
• Fixed: Visible damage sometimes corrupting car bodies, to properly fix the hotfix that temporarily disabled the visible damage.
• Fixed: Issue where clouds could show similar corruption as car bodies.
• Fixed: Showroom Upgrades not displaying which require MAS Files.
• Fixed: Glitch where “the lights would suddenly go off” when transitioning from light to dark at dusk.
• Fixed: Exploit where you could still set the FE car to use 250kW in the setup in a race when that setting is reserved for “attack mode”.
• Fixed: Overlays would not correctly follow the attack mode countdown timing.
• AIW Editor: Added new line smoothing option, and added soft selection for manual adjustments.
• Added an internal option to the player.json to turn off multi-threading in the UI. Defaults to “on”.
• Added code support so we can detect a disabled Steam overlay, to warn you about why a shopping cart won’t show up.
• Changed how a dedicated server shuts down when asked so it no longer crashes with a non-zero return code.
• Pausing sessions in a dedicated server will now only pause the first one.
• Increased an internal timeout when talking to Steam to give it more time before we give up or retry.
• Scene Viewer Changes: Moved Web UI Port to 5396 by default and added support for up to 128 scenes in configuration file

In a change from the normal procedures with build releases, this new update will be available to players as a special 'Release Candidate' branch of rFactor 2 - meaning the current public version of the simulation remains unchanged, and platers can choose to 'opt in' to this new update right-clicking rFactor 2 in your Steam library, and selecting ‘Properties – betas – release-candidate’.

Indianapolis Motor Speedway Update

In other rFactor 2 news, the studio have also deployed a new update to the free Indianapolis Motor Speedway circuit, adding a wide range of improvements to one of the older circuits within the simulation.

You can check out the changes in the latest build notes below:

• Update of 2013 Release to 2020.
• Replaced 2013 layout with modern equivalent
• Full PBR Conversion of all assets
• Replaced Crowds with new assets
• Replaced Marshals with new assets
• Added Pitlane Props (Pit Stands, Fuelling Rigs, Pit Markers)
• Added accurate background nightlighting
• Added Advertisements
• Added building at Road Course S1
• Improved various assets such as armco barriers
• AI Updates to 2007 GP and Modern GP layouts.
• Various other improvements and tweaks

Original Source: Studio 397

rFactor 2 is available now exclusively on PC.

Got questions? Our community have answers! Check out the rFactor 2 sub forum here at RaceDepartment for a great way to engage with your fellow fans of the simulation.

 

[hotak]

@mantasisg i mostly agree with you, but i think the core of the problem is that most simracers are racing fans, but know next to nothing about cars also because no one cares anymore about mechanical details of real cars too! Reviews all focus on subjective opinions and confronting laptimes (often done in different weather conditions), no one cares anymore about suspension architecture or advanced solutions.
Also very little of the simracing base actually "knows" the different physics engines, what they do or don't simulate, but still everyone feels like being able to judge developers and call their games "arcade" or "best sim ever".

I think this discussion would even deserve its own thread, just to put some objectivity back into the discussion:
All sime engines today are VERY realistic, by this i don't mean that they're all perfect, there are some nuances and in some cases they're straight wrong (IIRC rfactor2's differential coast behaviour and pCars2 FFB during braking are litterally wrong), but they're all just better than, say, the gMotor2 that everyone has been in love with.

let's tryto be more precise about what is actually simulated, how complicated it is and what approximations most sims use.

source: I'm not a mechanical engineer, but i am an AI & robotics one and did an 11 DoF (degrees of freedom, to compare rF2 has 190+) motorbike model with a brush tire model, 4 DoF active aero and trajectory-following control, and always followed sim engines developement.

- Suspension model:
independent suspensions are "easy" to model if you consider them rigid, the most complicated part is simulating non-independent architectures, so most older sims approximate them with "equivalent" independent architectures, that said all modern engines should do multilink (the most complicated independent architecture) correctly.
The limit is that - for example - mac-pherson is not really good because suspensions aren't actually rigid. I'm not sure how various sims deal with this, i think is mostly up to the car-physics team to find work-arounds, for example by adding "fake" very rigid spring+dampers to simulate suspension flex.

-Transmission model:
On surface is very easy to do: you have a force coming from the engine that gets split according to wheelspeed. That's if you do it in an approximated speed-based way. To do it correctly, you'd have to do it in a force-based way, add brakes, inertias and flex to the model and then model differentials in a more physics-based way.
Easier to do than suspension flex since all flexximility is in only one direction (the "link" direction), but still not implemented in every sim.

-FFB:
this would be a VERY long paragraph, so i just resume it saying the the reason FFB is not as straightforward as taking front wheel forces and pass them to the wheel, is that in real life those forces pass trough a flexible link and contrast forces at the wheel, but our physical wheel only reads positions, it has no load cell, and given non-infinite update speed in the sim-wheel connections, flex is also very difficult to implement.

-Aero model:
This is also very complicated to get right, also because RL data is not really directly misurable and real-time aero simulation is just not feasible.
Still most modern sims do a good job by using multiple wings across the car with different sensibilities to yaw/pitch/roll angles.

-Tyres:
Here's where various sims differs the most. I want to set this straight: all modern "sims", (yeah, even pCars3) have extremely complex tyre models, and i'm just gratefull that somehow they can get those to run at like 600hz frequency, i doubt any programmer on this forum could even get close, i'd like to see more people appreciating this instead of constantly bashing developers.
What's the difference and why some people sometimes like less advanced ones?
we have various kinds of models:
- pacejka: used on very old sims, it uses a "magic formula" not physics based in any way, but that gives tyre curves that are very close to experimental real ones for "normal driving conditions" (that's why it used to be so difficult to drift in sims)
-lookup-table based: very detailed tables associating variables (temp/pressure/wheelspeed/slip/etc.) to grip values are generated offline, either by a physics-based model (rF2) or manually trying to replicate RL curves. The sim then may just interpolate between the table values in real-time or do some extra real-time physics-based processing on them.
-realtime: this is the most computationally expensive and most difficult to do. I think only the madness engine has gone this way, but i'm not sure about ACC. The most basik kind of real-time is brush: you basically treat the tire as multiple "brushes" that can flex independently and that have a contact patch each. The tire grip is then calculated based on how much contact the tire is making and how much each of the the brushes's flex is "pulling" the tire in a direction. That's what was used in NFS: shift series, the SETA tyre model builds up on this considering lots more variables so that the brushes only accounts for tire carcass, flex is also based on temperature, nearby brushes position, etc, and the contact patch is then a model itself that takes carcass model output and builds on it and surface conditions to output the actual grip.

Obviously real-time model have to take more approximations than offline models, but they also should react better to extreme conditions and can account for more surface nuances.
All this wall of thext was just to say that what we have is VERY good, it's simply a matter of quality of the single cars physics and of how much each tyre model is understood (expecially when a physics-based model is used, either online or offline)

 

[Jan Larsen]

Great post! ^ But...

...slightly

I dont know why. but I sense a Mod doing facepalms somewhere

 

[gc24]

So, is Indy supposed to have that much vibration with the Indy cars?

 

[Kenny Paton]

I dont know why. but I sense a Mod doing facepalms somewhere

Not at all. I think this is a fascinating subject, but yes it can't be allowed to derail the thread so please start another to continue the discussion and I will move the posts in here to it.

 

[Cote Dazur]

Not at all. I think this is a fascinating subject, but yes it can't be allowed to derail the thread so please start another to continue the discussion and I will move the posts in here to it.

A thread/conversation, on all the points addressed in @hotak post, away from a specific SIM title, would be lovely.
Looking forward to it.

 

[Tr0nic]

It's there for me now without RF2 running, the store page was updated very recently to re-add VR support so i guess that's fixed this particular issue as well.

Awesome!! They rly fixed it... THANKS S397!

 

[MetalMania]

Just ran a quick 15 lap race at Indianapolis "Brickyard" version, stock car vs 25 AI. I like the updates to the track, but the AI were useless. Almost every one of them hit the wall either on entry or exit of every corner, every lap, for the entire race. This was also the first time I drove the "new" stock cars. I hadn't run the Indy oval in a long time, but I seem to remember the AI having trouble there before too. I haven't tried any of the Indy road courses yet since the update, but I've had good races vs AI on the old track (2014 layout in particular). Really hope they can fix the AI on the oval, it's bad.

I've also had a strange thing happen in recent weeks with AI on some mod tracks (Road America 2019 and COTA, running S397 GTE). After a lap or two, about half of them pit even if it's only a 10 lap race, and the rest just start pulling off to the side of the track and retire. I figured maybe it was because they're mod tracks, and I suppose it could be, but I don't remember this happening on those tracks before and it doesn't seem to be all the time. Maybe it's just coincidence, and I know S397 can't be held accountable for issues that occur with mod content, but I do worry every time there are build updates what is going to end up getting broken - unfortunately including things that are native to RF2.

 

[davehenrie]

This may sound crazy, but try renaming your player.json to player.jsXXXon . MarcG has found this cure several 'odd' behaviors. Then verify the integrity in Steam and you should get a new fresh player.json installed.

 

[sbabb]

the person who was making the Stockcars said he will look into the issue these cars now have on the new Indy, they didnt had that before fyi

 

[Bob Gnarley]

Indy GP circuit
New player.json - deleted cbash
23 ai - BMW CS Cup - Str 100% - Lim 33% - Agg 44%
Grip: saturated - temps 77 - 82f

In practice i noticed the AI losing the plot at several turns. Here we go again, usual junk ai from 397.

Started the race from the back to see how far the ai would make it and...pile up at T1. Restart, same again. Restart, same again.

At Spa the GT3's slide around like wet fish and endurance pack 1 & 2 start conking out after a lap as per MetalMania's post.

The ignorance that 397 reserves for its offline users is the same as i shall be applying to any future content.

 

[mantasisg]

@hotak Thats a super post, amazing knowledge, I'd like to reply to it in separate thread, maybe someday, would need time for it.

I agree that underlying physics might be proper simulation, and then by making individual content physics it could be "arcadified", or if in the other way - made unrealistically hardcore.

In my opinion, for example, if game has serious underlying simulation, but all cars are off and arcadified, then to me that sim counts as an arcade or simcade.

It would be indeed amazing if many simracers knew technical differences in between different simulation engines independently from content. Maybe some interactive tables with comments could be arranged someday. Experienced physics modders could get on that, not sure about non-moddable simulations... Believing ? Speaking of rF2, I think people don't know lots of its strengths, but there are some limitations that aren't very much known either.

 

[Bram Hengeveld]

There are no authorities here in RaceDepartment who would speak physics and realism, would create educating and interesting articles to help to keep simacers brain from death.

May I remind you that you are posting on a community forum. This means the community itself is responsible for the content. So do please climb in your pen and start writing awesome content about physics, when you are done please tag @Bram Hengeveld, @Paul Jeffrey or @VernWozza and we'll make it appear on the frontpage with a magic click of a button so the entire simracing world can see it.

This is an open offer to every other sim racing (physics) experts, we honestly like to hear from you

 

[davehenrie]

Just ran a quick 15 lap race at Indianapolis "Brickyard" version, stock car vs 25 AI. I like the updates to the track, but the AI were useless. Almost every one of them hit the wall either on entry or exit of every corner, every lap, for the entire race. This was also the first time I drove the "new" stock cars. I hadn't run the Indy oval in a long time, but I seem to remember the AI having trouble there before too. I haven't tried any of the Indy road courses yet since the update, but I've had good races vs AI on the old track (2014 layout in particular). Really hope they can fix the AI on the oval, it's bad.

I've also had a strange thing happen in recent weeks with AI on some mod tracks (Road America 2019 and COTA, running S397 GTE). After a lap or two, about half of them pit even if it's only a 10 lap race, and the rest just start pulling off to the side of the track and retire. I figured maybe it was because they're mod tracks, and I suppose it could be, but I don't remember this happening on those tracks before and it doesn't seem to be all the time. Maybe it's just coincidence, and I know S397 can't be held accountable for issues that occur with mod content, but I do worry every time there are build updates what is going to end up getting broken - unfortunately including things that are native to RF2.

rename your player.json file to player.jsXXXon.(advice from MarcG) Validate the files in Steam. This should download a fresh copy of the player.json file. (caution, any edits made to the old file will have to be re-edited.)
If you are running a race based on laps, try time. or vice versa. This has been an issue occaisionally at even the DLC tracks like Sebring. I was trying to help a guy with this, but my AI completed every lap. I would also suggest turning off any time acceleration while racing. The faster time=of=day computations are advanced, the more rounding-up errors can garble things.

 

[MarcG]

Running the Sim creates a fresh player.json after backup, no need to verify unless necessary.

 

[hotak]

May I remind you that you are posting on a community forum. This means the community itself is responsible for the content. So do please climb in your pen and start writing awesome content about physics, when you are done please tag @Bram Hengeveld, @Paul Jeffrey or @VernWozza and we'll make it appear on the frontpage with a magic click of a button so the entire simracing world can see it.

This is an open offer to every other sim racing (physics) experts, we honestly like to hear from you

I'll try to write something in the incoming weeks, but have lot to do for my thesis so iI cannot guaranteee anything.
Unfortunately i can mostly talk about madness angine and rfactor2 (and gMotor2, but it's old), don't really know much about AC/ACC or how much raceroom progressed since early days.
I just have a bit more insight than most people because i actually made some motorbike simulations, but i made them in simulink/matlab and while they're much "simpler" than most sims, they do not run in real time (mostly because they do division operations and those are very slow when numbers get small, partly because they become "unstable" when the bike starts falling)

 

[Bram Hengeveld]

I'll try to write something in the incoming weeks, but have lot to do for my thesis so iI cannot guaranteee anything.
Unfortunately i can mostly talk about madness angine and rfactor2 (and gMotor2, but it's old), don't really know much about AC/ACC or how much raceroom progressed since early days.
I just have a bit more insight than most people because i actually made some motorbike simulations, but i made them in simulink/matlab and while they're much "simpler" than most sims, they do not run in real time (mostly because they do division operations and those are very slow when numbers get small, partly because they become "unstable" when the bike starts falling)

What I meant to say is it doesn't need to be exclusive about physics. We are a community forum where we can showcase whatever we want about our great simracing hobby. We always like to feature awesome written content on the frontpage.

 

[jappygamer]

here you go, quad 3090s, might be able to play ACC in VR with this? Cant wait till the "new" release by kunos on UE5 comes out.. we can rent turns on ILM's supercomuter infrascructure to get 25fps..
View attachment 444550

Oh man, i dont know if you are some sultan or not but.... i can see your sweet home burning after 1 ACC hour session...no ventilation room between graphic cards!

 

[jappygamer]

I don't know how to combine posts, so here is post #2:
When it asks for a favorite server, click on MORE and get the list. Choose the server with the lowest ping. DO THIS EVEN IF YOU NEVER RACE ONLINE! Do this once and the time taken to scan the entire list of servers will be greatly shortened. Don't worry if the server is Passworded or for another series you don't race. Just choose the lowest ping server and forget it. Two minutes will become just a couple of seconds that you only have to skip past once as you navigate to offline racing.

Thanks for the free tip, but WHEN the hell is S397 adress that!? It is a shaming issue for a company like them, it should have been solved months ago..

And now i launch an idea to S397: may be if we pay for a dlc without these favourite online servers you would solve it sooner? I dont know... may be 9.99 Eur and 5.99 when there is some steam offer?..

Nahh I was just kidding S397, we can wait! take your time.
There is no hurry, it is said that sweet things come slowly .

 

[Mark Lawton]

My opinion, not everyones, but mine at this time.
And I reserve the option to change my opinion in the future, if things change.
I loved RF2, but it just became so unstable. After updates I found the new UI just crashed, after taking ages to load- even though I'm hitting it from a SSD. Or I get to do 2 or 3 races then another CTD.
Great physics yes, and I loved the Ai- but I uninstalled some time back. I had paid content in there that I also loved, but these days I just can't do it anymore sorry.
I have whenever I've tried again, downloaded, installed all the workshop content thats mandatory- even though half of it has issues, got through the UI snail to set a race, tried something simple... waiting waiting waiting oh look CTD.
Sorry I think this was a great sim, but now, for me, its too old, broken, and dated. Graphics, sound, just everything really. I no longer have a romanticized view, and I doubt I'm going to use my race time trying to find fixes anymore. I've just given up with this mess.
I will try again in the future, and my opinion may change then.
IF RF3 comes out, I'll probably be first inline to buy a copy. For now, as it stands -sorry guys this just aint happening for me.

 

[Hikonyan@rFS]

@hotak The views of those involved in the development of sims, such as those used in research and development, are very interesting. You mentioned that tire models fall into three categories. If you know the strengths and weaknesses of "Tame Tire" and "Fiala Tire", can you tell us? This is because rFactor2 may be used in Tame Tire (partial?) And Fiala Tire may be used in AC / ACC / iR.