rFactor 2 | Taking Stock And What Comes Next

Paul Jeffrey

Premium
Studio 397 have given fans an interesting insight into the ongoing development of the simulation, as they continue to investigate and identify bugs and issues whilst expanding the core experience.
  • Bug tracking and fixing overview
  • Advice for driver swap workaround
  • Release documentation for modders

As many of you will be well aware, Studio 397 are the development team behind rFactor 2, and thanks to the changing world landscape brought about by the Covid-19 situation, recently found themselves thrust into the spotlight as the simulation of choice for the virtual Le Mans 24 Hours event held a couple of weeks ago.

rF2 Le Mans 2.jpg


Although the event was a huge success for the studio and sim racing itself, 24 hours of racing and 200 + drivers did bring up a few issues with endurance racing within the sim, and the team are actively looking to recreate and rectify these problems - with work having already started to yield some positive results:
"A few weeks ago, after the 24 hour race, we promised to give you regular updates of our efforts to track down and fix the issues that occurred during that race, and a few races before that. In fact, this effort is not something we started two weeks ago, it is an ongoing process. But before we look at the specific issues, let us first explain a bit more about software development and fixing bugs" said Marcel Offermans of Studio 397.​

For starters, let’s give you a bit of background on how we develop code, the processes we have in place to ensure that our code is correct as well as a brief introduction on what kind of bugs there are in code and how to fix them.
Every change we make to the codebase, either to fix a bug or to implement a new feature, is developed in isolation. As soon as the developer working on it is confident the change is correct, he or she will do two things. The first is to ask at least two colleagues to review the changes. By getting others to look at the code, we typically catch mistakes that the original author would miss. I’m sure you have had cases where you are blind to your own spelling mistakes when writing some text. For code it’s typically no different. The second is to create a build on Steam that can be tested by others. If the change passes both checks, it gets integrated into the next update. That goes to our group of beta testers who again test the change to ensure it works as designed. If that passes, you will find the change in the next public update.
So that’s how we deal with changes. What about the extensive codebase we already have? Here the process starts by identifying a bug and being able to find a series of steps to reproduce it. If we can reproduce it, we typically have no problem fixing the underlying issue.
There are two types of bugs in code. Bugs that come from mistakes in the logic of the code. Once identified they are typically easy to reproduce and fix. The second type is timing and threading related, also known as concurrency bugs, and in recent years these tend to happen more often as processors get more and more cores and a lot of things happen simultaneously. Reproducing this type of bug is usually a lot more tricky as the slightest change in the execution timing might cause it to appear or disappear. Finding them requires a combination of luck, lots of testing and in some cases rigorous code reviews. Given that the rFactor 2 codebase consists of millions of lines of code (to compare, a typical novel will be around 15.000 lines) you can probably figure out that going over it front to back is going to be a very time consuming process. And you might still glance over the actual issue (do you still remember what the name of the street was that the main protagonist was crossing on page 34?).
Now that we’ve explained the process, let’s go back to the 24 hour event. Directly after the race we started analyzing and categorizing all the different reports we got and pretty soon we were convinced that we were looking at one or more concurrency issues. Reproducing these was going to be tricky. We have a framework that we can use to setup fully automated tests, so our first step was to try and reproduce the exact conditions of this 24 hour race. Specifically we started designing test scenarios that resembled the reports. Two weeks later, we are now seeing the first results of those tests, with a reproducible scenario that we are investigating further. I expect us to be able to explain more in the next roadmap, but the good news is that we found something.
As a result of finding something, we also have some preliminary advice for those running endurance races. Contrary to what most series have been doing, making sure that the replacement driver joins shortly before the intended driver swap and then having the original driver disconnect soon after, it for now is probably better for all drivers to join the server before the race starts and stay on the server throughout the whole event. Our code in theory can handle up to 104 drivers and another 104 spectators, so if your races have less than that, try out this advice while we continue to track down these issues.

rF2 Le Mans.jpg


Of course rFactor 2 is very much a platform favoured by endurance drivers, and any improvements to the stability of the simulation in long distance racing and driver swapping situations is likely to be well received by the vast majority of the playerbase. However, another key aspect of rFactor 2 is the ability for community members to create and release mods for the software, and this is something that the Dutch studio have long since mentioned they want to improve and increase the levels of documentation and tools for established and aspirin modders alike.

Thankfully, it looks like Studio 397 are on the road towards developing some much needed further documentation around these very subject, as Marcel continues:

"With all the recent updates to the game visuals, we are now preparing to release documentation to help modders take full advantage of these updates. For now this will focus on track content, with car updates coming at a later date. We have been working hard to upgrade our tools and documentation to help make that possible. Tools will be made available including 3ds Max 2021 plugins and an independent Material Editor, and full documentation will be released for the shaders, as well as some guides and tips for getting setup. This will represent a significant step forward in the way content can be created and the options available to modders, with all the latest features, including improving night-time effects, tree shading and road and terrain blending options. To support this, we will release an updated Loch Drummond that will be included in ModDev by default. This track is a small fantasy circuit that has been updated to make full use of the various options from our work on recent tracks.
In recent updates, we made adjustments to old content to help it be compatible with lighting updates. We will add a few options here to help modders, including the ability to disable these changes on a per-material basis if required. Also with this update, we will allow modders to customize the atmospherics in tracks again. This will now be in the form of scalars to adjust the density of haze in different conditions and also to set the amount of air pollution. This should result in an easier to use system and help us keep effects consistent across the game.
This information will all be made available on our Developers Guide: https://docs.studio-397.com/developers-guide/

rf2 portland.jpg


So that's the immediate future of rFactor 2 covered, just left to remind you that the studio recently deployed a new update to the simulation that includes the liveries of the Le Mans 24 Hours for those who own the cars, the recent free Portland track, various BOP improvements, announced the Ferrari 488 GT3 and a further, as yet unannounced GT3 car alongside a nice update coming the way of the existing models, Silverstone updates and plenty more besides.

A good time to be a sim racer and rFactor 2 fan it seems....

Original Source: Studio 397

rFactor 2 is available exclusively on PC.

Want to know how to get the best from the sim? Start a thread in the rFactor 2 sub forum and let our community offer you the benefit of their massive combined experience.

rF2 Footer.jpg
 
Last edited:
How long are they "still in development"? Thats the thing.
You know as well as I do, that bugs from rF1 are still in there and haven't been addressed.
You also know, the way they have closed it now, kept silent of new entries in the files and such; I will never drive a toilet bowl down Mulsanne straight :)

Sure, they should develop, but with every bug that is 10 years old, you can't use that line.

Good reading is also a thing ... Marcel says that the technology does not stand still, which always involves changes. In short .. I am very happy with every update that comes from Rfactor 2
 
Just a thought, in light of what Reiza, a studio of similar size to S397, has been able to do with the SMS graphic engine, or Kunos with the Unreal Graphic engine.
Should S397 just ditch the ISI graphic engine and apply the physic code that makes RF2 such a pleasure to drive to more modern and robust graphic engine instead of trying to fix legacy issues that are plaguing RF2.
RF2 is a bitter sweet experience at best, despite a best in class physic and FFB driving feeling. After all those years, it might be time to ditch the old shell and as Kunos and Reiza have recently demonstrated possible, just start a new chapter.
Why ditch the best foundation in terms of physics on the market while offering a decent graphics engine and keeping everything in your hands? I would hate seeing rF going the Madness engine or UE4 engine route just for a bit more eye candy. If people still think that the new content looks bad then they maybe should reconsider their standards. The new cars on well done tracks look sublime and I can't see that much different to sims like ACC, AC or PC2 when it comes to offering a fine sharp image from the cockpit. The sky could be rendered a bit more beautiful with more types of clouds but that's the only thing that stands out.

Considering that Reiza inherited the bland FFB and everything that I don't like about PC2, there is no guarantee that a new engine won't offer a new set of challenges.
 
Why ditch the best foundation in terms of physics

I hear this so often from the fanboi's, but it's just not true. It may be the most advanced, but noone, not even S397, can use it right. There are always glaring issues in the physics of new cars that never seem to get resolved (just see this thread for examples).

The proof to me that RF2 isn't any good to me is with my professional clients. Of (mostly australian) ones I have, 50% are with AMS, 30% are with RF1 and 20% are Assetto Corsa. I've had 2 try rF2 and just go back to rF1 after a few month's, it's just not good enough.
 
Why ditch the best foundation in terms of physics on the market while offering a decent graphics engine and keeping everything in your hands?

I agree with you, it is not what I am saying. My tough is to keep the great physic and FFB and to match it to a better graphic engine and not just for "eye candy".
better graphic , yes, but also better UI and more robust interface.

My Physic and FFB in AMS2 is nothing like PCars2, it is actually very good. if yours is different, you might want to spend a little time with some the custom files available and tailor it to your taste.

I think there is some confusion about the code that govern physic and FFB and what an engine does.

UE4, SMS, ISI can all drive good or bad, it is the physic and FFB code from a knowledgeable developer that makes a difference.

S397 could probably apply the best of class physic & FFB code to a better engine and we could end up with a SIM that could be better than anything out there as a total racing simulation experience, instead of trying endlessly to patch a sinking boat with dubious results.
 
  • Deleted member 526227

Looking really forward to the next steps of S397. It has been an incredible year so far with all of these stayhomeevents, big achievements.
Glad to read there will be a new GT3 soon plus the Ferrari. Current Porsche would be a good choice too, especially on an new iconic European track :)
 
Guys, we must understand how tough it is for programmers to find bugs and update anything
on our beloved Sims, especially on rFactor 2 since they are a small company...

Studio 397 are doing a great job lately with the visual updates, new tracks and cars, so we
should be happy, and more happy will be in the very near future because they are working
hard to improve the best Sim - so we have to support them and not being too picky, please.

Long Live rFactor 2!
 
Just a thought, in light of what Reiza, a studio of similar size to S397, has been able to do with the SMS graphic engine, or Kunos with the Unreal Graphic engine.
Should S397 just ditch the ISI graphic engine and apply the physic code that makes RF2 such a pleasure to drive to more modern and robust graphic engine instead of trying to fix legacy issues that are plaguing RF2.
RF2 is a bitter sweet experience at best, despite a best in class physic and FFB driving feeling. After all those years, it might be time to ditch the old shell and as Kunos and Reiza have recently demonstrated possible, just start a new chapter.

Highly unlikely to happen, for starters S397 had to make their own DX11 Engine to work with RF2 (the game was DX9 before they took over) and they've done a massive amount of work on the Lighting & Graphics over the past few years and are continuing to do so, not to mention the Performance Improvements (of which there have been many) during that time as well.

So to suddenly drop all that in favour of another "Graphics" Engine which they'd have to learn from scratch, nah just cannot see it happening and quite frankly I hope they don't, they've got a full grasp on this one and will only continue to improve it into the future.
 
I am a user of rfactor 2 and I am happy as a roadmap to a roadmap is improving in graphic and physical terms, since studio 397 came, improvements have been made and I know that there are still some to be done but over time they will. After the 24 hour event, they updated the cars physically and visually. No simulator is perfect I had iracing and online is the best but the ffb is bad and the tire can't hit the nail on the head, until not long ago it had no day / night transition. There are fansboy of all simulators waiting to destroy the rival
 
Agile working? Seems like the product backlog items are either constantly changing or prioritised (hopefully nothing is deleted).

Bug fixes will always be there and that’s totally understandable for any software but there’s always a target or planned date for the fix(es) [for at least the ones that have high prio]

So is the outcome
- polished stable UI?
- new online Ranking system?
- New visuals for cars, tracks...etc?
- GT3 bop?
- New cars?
Anything more?

it would be nice to know the below to have the overall picture.
- What’s committed on the road map for at least 2020.
- What’s in development (at least short term).
- What’s got postponed if any.

(From a sim racer fan)
 
BRB just grabbing some popcorn

Jokes aside, I respect all the work S397/S371 (Discord users will know this joke) has done. After various updates the official content is looking very good. Just hopeful that the new UI won't take another 3 years to finish up... Other issues are starting to press on also, so taking that long would be not good.

Otherwise, it's good to see that they're working on fixing the issues that occured. The only thing which doesn't sound too good is that "they found something". I like the honesty, but I think saying that the issue was "under proper investigation" would work better.

And as a friendly reminder, if you encounter any bugs, it would be better to proper report them (including what exactly happened, what were you doing (what are the steps to encounter it), can you reproduce it yourself, any other specific details - giving as many details as you can can help the devs a lot sometimes!), but you know, you can do whatever, this is just my honest suggestion...
 
So to suddenly drop all that in favour of another "Graphics" Engine which they'd have to learn from scratch, nah just cannot see it happening and quite frankly I hope they don't, they've got a full grasp on this one and will only continue to improve it into the future

You are most probably correct, it is not going to happen, might not even be a good idea.
I am not totally convince they have "full grasp", but they most likely put an honest effort into improving it.

I am just a little sad that other developers, similar in size, come up with a brand new shinny very functional SIM in less than 12 month when RF2 is still in dire need on improvements.

Of course if the Physic and FFB was not so interesting, no one would care, but it is excellent in that respect, hence the concern.
 
You are most probably correct, it is not going to happen, might not even be a good idea.
I am not totally convince they have "full grasp", but they most likely put an honest effort into improving it.

I am just a little sad that other developers, similar in size, come up with a brand new shinny very functional SIM in less than 12 month when RF2 is still in dire need on improvements.

Of course if the Physic and FFB was not so interesting, no one would care, but it is excellent in that respect, hence the concern.

Seeing as some ISI Employees/Contactors moved to S397 I'd say they have a much better grasp on the engine hence the "full" quote :)
 
Seeing as some ISI Employees/Contactors moved to S397 I'd say they have a much better grasp on the engine hence the "full" quote :)
Probably true, I was mostly facetious, as there is progress, but at the same time that engine must be super hard to work with as the progress are not coming as fast as every ones hope.:):)
 
Back
Top