rFactor 2: The Big Interview (Part 3)

Paul Jeffrey

Premium
rF2 4.jpg

Part 3 of our Studio 397 / rFactor 2 interview with Marcel Offermans is now live.


If you missed our the first two instalments of our new interview, you can check out Part 1 here, and Part 2 here.

Part 3...

RD: (question from @Martin Fiala) - There's been a lot of talk about the upcoming UI, but very few mentions (if any) lately of the new plugin system that was supposed to solve the issue of plugins not being able to draw to the screen in the DX11 version. Is that still planned? Is it considered part of the new UI? If not, are there at least some in-game replacements for the most popular plugins planned, like proper track map or the very badly needed FFB meter?

MO: Allowing plugins direct access to DX9 was causing a lot of issues, which is why we took out this ability with DX11, with no plans to bring it back in the same way. A new way over overlaying information will be part of the new UI, although we are not yet confident that this will make it into the first (beta) release. That said, you probably have seen already that we can render track maps in-game in the new UI, so I’m confident we will see that feature return. Same probably goes for the FFB and pedal meters.

RD: (question from @Martin Fiala) - Are there any plans to improve the sound engine that seems to be basically limited to a very simple on/off reverb and to stereo only sound?

MO: Yes. The current plan is to look at this starting probably Q4 of 2019. I anticipate this work to go well into 2020 though, so you will have to be patient!

RD: (question from @Peter Koslowski) - All I need is 3D grass. Will all new and old tracks (Zandvoort already has it) be equipped with it?

MO: It is highly unlikely that we will retrofit the type of 3D grass that was used at Zandvoort to all other tracks. There are a few reasons for that. First of all, placing this type of grass right now is quite time-consuming. The second reason is that we are not yet satisfied with the overall effect. The third reason is a more technical one, we lack a few bits and pieces to really end up with nice looking 3D grass. So I guess the good news is that this is something our artists and graphics developers actively discuss. We have not decided on an exact implementation yet.

RD: (question from @leclettico) - There's a plan to manage a custom championship with the AI (so, with the same AI drivers) without having to create copy of liveries and drivers and using LogAnalyzer to looking into results and standings?

MO: This is planned as a “second stage” of our new competition system, the first stage being online races against humans. That said, I do think LogAnalyzer did an amazing job on their system.

rF2 1.jpg


RD: (question from @leclettico) - Rain and AI. There will be improvements in the managing of pit in case of rain?

MO: Yes. A few of your fellow community members asked a similar question and it’s certainly one of the areas where our AI can and should improve.

RD: (question from @Jonny Austin) - Many sim racers praise the handling in RF2. What are the other unique features that should make me buy RF2?

MO: I think by now we are the only “open” racing simulation that is still under active development, meaning you can build your own cars and tracks and customize a lot of aspects of the simulation. Also, we have fairly robust on-line play with intelligent prediction technology to ensure that even with some latency, we can still allow very close racing without (m)any surprises. Couple that to a dynamic track and weather system and I’m sure you will keep finding new racing challenges for a long time.

RD: (question from @Alexandr Meshkov) - Is there Portland International Raceway still in development? No news for 1.5 years about it in Roadmaps

MO: It was in development as we anticipated using it for some specific project. That fell through, and we also were not yet happy with how the track looked at the time so we decided to put it on hold and focus our resources on other projects. Does that mean we won’t ever finish it? Probably not. It’s just that right now there is no news about it.

RD: (question from @Alexandr Meshkov) - What is the dirt maps and how it works? I remember that was a word about that feature when Sebring releases but no explanation since that.

MO: Yes, I spoke too soon. We implemented dirt maps for Sebring and I think it really helped the look of the track, but when we evaluated the technology and how we used it after building that track, our artists and graphics developers agreed that we wanted to make further changes to the system before “handing it over” to modders. Which is why you did not hear much about it yet, as we’re still working on a better version.

RD: (question from @The Walker) – When do you expect to release the Tatuus Pack?

MO: Due to my lack of speed in answering all your questions, this pack by now has been released. Do you like it?

rF2 2.jpg


RD: (question from @The Walker) - Also, but this is just curiosity, I read somewhere like in 2017 that rF2 had the license for the Reynard 95i car. They never got released, right?

MO: Correct, and I don’t think the car was ever modeled either. It was one of the licenses we “inherited” from ISI and one of the issues we had with those at the time was that we were really getting a lot of feedback from our community to go for series or classes of cars instead of individual ones (unless those were “cup” classes of course). So we decided not to do anything with some of those licenses.

RD: (question from @SLimJim70 ) - As someone who works constantly changing shift patterns, do they ever plan to incorporate 'drop-in/drop-out' online racing?

MO: Yes, that is planned as part of our competition system, where we intend to host events that contain multiple races where you can attend one or more of them in a period of multiple days, which should accommodate people with changing or unpredictable schedules.

RD: (question from @VernWozza) - Are there any plans to do a thing with the damage model in the sim. Loving what iRacing have planned, I would love to see something similar in rf2.

MO: Right now, no. I like what iRacing is doing to their damage model, so we’ll definitely keep an eye on that development but for now we have nothing planned in that area.

RD: (question from @Cristian Luis) - Will the new UI allow the possibility to listen to music while in the main menu? Just like in rF1?

MO: I don’t know yet. We have the technology to play music files, but the tricky part is choosing music everybody likes (and keeps liking even if he plays rFactor 2 for a long time). That is hard, and possibly expensive, considering we would need to license that music too. So would that really be the best way for us to invest our money? I’m just not sure. Nothing stops you from listening to music while you’re driving and/or clicking around in the menu system. Our current UI allows you to add music. We just don’t see a lot of people doing that.

RD: (question from @tswest) - Will there be more tutorials and assistance to allow third party builders to build tracks for rF2 like it has been for rF1. I am a track builder and would like to see more Building tracks in rF2 for DUMMIES tutorials. At the moment there is very little info and I get the impression Studio 397 wants to keep everything for themselves.

MO: If we gave you that impression, that is certainly not correct. The tools we provide to track and car builders did not fundamentally change much. We did upgrade them to 64 bits and newer versions of 3D Studio Max. With those in place, I think building tracks is mainly a matter of experience in 3D modeling in general and, on top of that, a lot of hard work. As engines become more advanced and faster, the level of detail for tracks raises, which has the downside that it becomes harder for people to build a whole track from scratch. We now use a team of 3 to 4 people, all seasoned professionals, to build a track and still it takes us many months, so I can only imagine how long it takes someone by himself who has a daytime job on top of that. But let me reverse the question, what kind of tutorial are you looking for, as there are certainly a ton of tutorials out there already about all aspects of 3D modeling and a lot of those apply to track too, so I guess the type of information we would need to add is specific information about our tools? Or am I overlooking something here?

RD: (question from @Rui Santos) - Any plans to update the wet weather physics, racing line, water on track, etc.?

MO: We are aware that our grip levels on a wet track are not always correct, so we do intend to tweak those. Water, rubber and marbles on the track are already quite accurately modeled through our RealRoad system so I think we’re in a good place there and we’ll keep expanding it.

rF2 3.jpg


RD: (question from John Eric-Saxen) - Across all industries we have recently seen a lot of buzz around AI and machine learning based on big data. However, in racing simulators the development of AI has stagnated at a very basic level. Could we possibly see a form of "neural AI" based on learning in some form in rF2 in the future? This could reduce the time and effort required to tweak the AI to work well with various car classes?

MO: This to us is certainly a very interesting research topic. As you might know our parent company, Luminis, is involved in cloud technology and we also have specialists in AI and machine learning. What’s holding us back right now is that we first need to finish our competition infrastructure so we can more easily collect enough data for such algorithms to work effectively. But this is without a doubt the future of AI in simracing (and a very nice question, thank you)!

RD: (question from @Travis) - will we have a return of the dev blogs? Michael Borda's blog was excellent; are these to be continued at some point?

MO: I certainly hope so! With the Brabham BT44B finished, I guess we need to find a new topic for a series of blogs. And maybe some of our other physics engineers can chime in too, as it’s not just Michael doing all the great work in that department.

RD: (question from @pcraenme) - Are you thinking of introducing something like a season pass/one year membership, so you pay all the upcoming DLC content for a certain period (e.g. a whole year) in advance, but for a cheaper price then having to buy all the individual DLC's when they come out?

MO: Such a system is currently impossible to implement with the features Steam offers us. We will keep aggressively pricing bundles and participating in Steam sales. That model works a lot better. That said, Steam might add features in the future that help with “season passes” so please also relay such requests to Valve as they tend to listen to their users (if enough of them want a certain feature).

RD: (question from @Terry Rock) - Will there be any emphasis placed on setting up any RF2 dedicated servers, strictly for pick-up racing with original content?

MO: It’s our intention to host races for all kinds of people through the competition system, so I think you’ll find both pick-up style races and championships there. And obviously we’ll keep a close eye on what our community wants to drive and adjust our schedule accordingly.

RD: (question from @mantasisg) - Better wet surfaces simulation, viscous and dynamic hydroplaning

MO: I think there are two things we can improve when it comes to wet surfaces. The first thing are the grip levels of parts of the track that have a lot of rubber and marbles on them when the track gets wet. We are aware that those are not totally right yet. The second thing is, as you mention, aquaplaning, which is currently missing. We are still studying the latter to see how we can best implement that in our model, because if we do it, we want to make sure it’s physically accurate.

rF2 6.jpg


RD: (question from @mantasisg) - Better gearboxes, drivetrain simulation. Will it happen ?

MO: Right now we are still prioritizing the UI and competition system, but once that is in place we will certainly free up some resources to work on these aspects. Some of you might know that, before forming Studio 397, some of us were involved in the development of Automobilista for Reiza Studios, where we implemented some of these features in the rFactor 1 engine. Of course we want to take the simulation of the drivetrain and gearbox to the next level as the rFactor 2 engine offers a lot more possibilities to model such things.

RD: (question from @mantasisg) - More graphics (besides PBR) improvements. Will it happen ?

MO: Yes, certainly, for us it is very important to get the basics right first, which means improving our lighting model and using PBR based rendering. Once that is in place, it opens the door for more consistent improvements. Most likely you will see us improving cockpits, a very important aspect since that’s your daily office as a racing driver, but also address things like shadows and our particle systems.

RD: (question from @mantasisg) - More variety in cars content, more classics, more range in years, more range in vehicle types. Will it happen ?

MO: I would say it is already happening, but it will also be fair to say that we won’t ever be able to cover every form of racing out there, so there’s a nice task for our community as well to fill in some of the gaps. And if those gaps turn out to be popular, we can even look at working with modders to go after official licenses, etc.

RD: (question from @Boby Kim) - Nowadays the community wants graphics instead of physics as graphics are directly visible, and thus understandable, and physics are not: a direct outcome of Reiza going Madness=they will receive a huge chunk of the Project Cars community who are not interested in physics layers. How do you stand up against this change with rF2 in the future as more and more people choose for graphics and content instead of true simulation?

MO: Interesting question. Obviously I can’t answer on behalf of Reiza Studios, but I’m sure they had good reasons to make the choices they did. I think honestly with our new DX11 engine, if you compare it to the DX9 engine we started out with, we’ve already improved a lot. In some cases even well beyond what the competition has, such as our car paints that feature up to six different unique material regions that can be different on each individual car. We are far from done though, graphically, so I expect us to stand our ground. That said, from experience, a good physics engine is what keeps people coming back in the long run and I think we are really doing well in that department.

RD: (question from @DriftManiaX) - Will they ever fix all the bugs instead of releasing new paid content?

MO: I like to believe we’re doing both. No software is free of bugs though, so I guess the formal answer is no, we won’t ever be able to fix them all. We’ll definitely give it our best shot though, balancing both adding new stuff (features, paid and free content) and fixing bugs.

rF2 5.jpg


RD: (question from @FS7) - Will rF2 have an option to enable trackmap on the hud?

MO: As you might have seen, we already have a trackmap that we can enable in our new broadcast overlays, so the technology is there to also add it to the HUD. If you ask me personally I don’t think a trackmap is a very “sim” like feature, but I guess a lot of people do ask about it so I’ll answer this with a firm “yes”.

RD: (question from @formidable) - Is expected the release from a classic Le Mans layout (70-80-90), plus some Group C or LMP1 Cars?

MO: Right now we don’t have plans to release classic versions of Le Mans, doing the current one was already a huge project and we intend to provide a few more updates for that one. Group C cars are a personal favorite, so who knows. LMP1 cars are unlikely at this point as they are slowly dying and not yet “classics”, like Group C.

RD: (question from @Dean Hayes) - Given that your currently working on laser scanned versions for both Lemans and Nurburgring is it safe to say that Spa is next?

MO: It’s a lovely question, but one I can’t answer right now.

RD: (question from @Will Mazeo) - I'd like to know about more pit stop work… Mandatory pits with pit window (u got 2 official GT3 packs, this is a must have) / Min pit timer stop (like ADAC GT) / Larger pit stalls as people requested in the forums / Pit stop animations at some point?

MO: So that’s a few questions. I think our plugin system is powerful enough for leagues that want to follow the rules of specific series to implement those (mandatory pit window, minimum time in the pitlane). The larger pit stalls I am not sure about. Why would you want those? I hear a lot of people commenting they would like more pit stalls rather than less so I’m not sure if that makes sense. Finally the pit stop animations, those are hard to do for arbitrary types of cars and series regulations and I have not seen a single game that has done this for all the cars we have. If you limit yourself to a single series, you might be able to pull this off, if you need to solve all the cases, this becomes a very expensive project. So I don’t see this happening until there are better animation systems we could easily integrate that make it possible to implement “smart” pit crews that don’t require extensive motion capture animations.


RD: (question from @ilema) - Drift content + drift tracks + drift scoring system? Possible?

MO: I don’t believe there is anything fundamentally blocking people from making cars or tracks for drifting. Now I’m not super familiar with the drift scoring system. From the few races I’ve seen it’s mostly jury based, right? I am not sure how we would translate that into code. If you have an idea, you can certainly write a plugin that does that. If the question is, will we make drift cars in the near future? Probably not.

RD: (question from @pcraenme) - Can track loading times be improved, since rF2 is the slowest by far when it comes to loading a track?


MO: Our loading is currently done on a single core, and I think the reasoning was that loading times will mostly be I/O (disk speed) bound anyway. Now if that is still true in these times, remains to be seen. It’s an area we have not investigated a lot, but you’re not the first to suggest this so I’m sure at some point we at least will want to analyze better what the slowdown is. One reason I know and that is that our tracks are probably some of the largest ones around.

RD: (question from @A man with a Harmonica) - Do you intend to fix the erratic and often bizarre AI behaviour on wet track where they pit every lap and are either too fast or excruciating slow


MO: Yes, we would like to improve those situations, making AI more aware of the cost of changing tyres all the time versus running on a track with a certain wetness using the wrong. Some of these issues are triggered by starting in the rain but on a dry track. That confuses them. It’s raining so they start on wets, and when they start driving they think the track is too dry for them and they go to slicks only to go back to wets a short while later. Stuff like that I am sure we can improve on.


Part 3 is done! Stay tuned for our fourth and final instalment soon!

rFactor 2 is a PC exclusive racing sim from Studio 397 - Available now.


Check out the rFactor 2 sub forum here at RaceDepartment for the latest news and discussion regarding this excellent sim. Like your racing hard and fair? Join in with our rFactor 2 Racing Club for all your eSport racing fun! Oh, don't forget we like mods to, with our own rFactor 2 Modding Forum for you to enjoy!

Like what we do at RaceDepartment? Follow us on Social Media!


 
 
Not, that I’m interested in music while racing, but playing in the pits from the pit speaker system sounds interesting, could there be a way to add your own music files, there by sidestepping the licence issues?

Personally i think adding a simple music player is possible (Euro Truck Simulator 2 has one for example), but for me this is wasted energy in a racing simulator and relatively pointless. By that i mean everyone has Windows media player, or winamp (legend of a music player that!) or some other one they can use in the background so is it worth the Devs time and effort to add an in game version? Not for me.

Whilst having music played over tannoy's at the track would be "cool", i just don't think the small number of users to use it would justify the time to code something like that.

There's far more important areas already to concentrate on and if we look at RF2 and other Sims they're all missing "something" from the actual simulation experience of driving, that's long before we consider in game music.

Nice idea, but nope, not for me.
 
I'd imagine the reason for the super slow loading times is a combination of old tech, complex data structures, spaghetti code and big data files with lots of duplicate information. Rf2 also has multiple layers of packaging and each 3d mesh has more data than other games that is used less efficiently (or is basically useless duplicate). Just alone the fact that rf2 uses 4 uvmaps per mesh despite all them being identical in most cases already adds a 4x multiplier to some operations. Add in a mixture of various old file formats like bmp, tga and it all just keeps adding up. Although it is not straightforward to compare rf2 to any other game directly because you'd need to make sure the files going into the games need to be identical to get a comparable result and I have 0 interest doing that comparison. Regardless I'm sure rf2 3d files are heavier than the 3d files other games use. There is not really any more useful data in them. The data is just stored, loaded and used inefficiently.

I think rf2 doesn't have any optimizations (or can not have due to its spaghetti code and packaging nature) so it loads each content one at a time from start to finish. I think only ai files for the cars are loaded together for all cars in the end. Because how rf2 has veh files that can deeply alter the cars I'd imagine the way rf2 loads cars is that each skin is effectively loaded as full car from start to finish. This probably means rf2 goes through each car's folders and packages from start to finish loading decryption, data files, sounds, physics, car textures, pit crew textures and 3d objects one after other (not in that order obviously) per each car and then starts again from 0 for the next car whereas modern games tend to do all those things in batches. I think due to the rf2 loading code being really old being written when computers had something like 2gb of memory and with ssds not used yet so going through each skin individually from start to finish is easier to fit into the memory than working with with fewer larger sets of data. Maybe even reading and writing to the disk constantly to avoid using too much ram.
 
I think due to the rf2 loading code being really old being written when computers had something like 2gb of memory and with ssds not used yet so going through each skin individually from start to finish is easier to fit into the memory than working with with fewer larger sets of data. Maybe even reading and writing to the disk constantly to avoid using too much ram.
This ancient way of loading (and holding) files could also explain the way rF2 does handle the 3 urgent JSON files(controller, player, multiplayer) when the game crashes.
In a more modern game the files are loaded into memory and then closed on the HD/SSD while the game are running.
But in rF2 the files are not closed - so if the game crashes the files on the HD are (very) often damaged and cannot be accessed after the crash.
After my install of rF2 under Steam I had 3-4 crashes which completely destroyed the 3 mentioned JSONs - again and again!
And eventhough I consider the rF2 TM/physics one of the best out there I got fed up with the mandatory reconstruction of the files and the following wheel calibration and graphic setting reconstruction.
So I finally deleted rF2 from my HD.:rolleyes:
And use instead mainly Raceroom (or to a lesser degree ACC and pCars2) - because none of these 3 racing games have such laughable problems. :whistling:
https://forum.studio-397.com/index.php?threads/the-game-dont-start.63297/#post-989146
 
This ancient way of loading (and holding) files could also explain the way rF2 does handle the 3 urgent JSON files(controller, player, multiplayer) when the game crashes.
In a more modern game the files are loaded into memory and then closed on the HD/SSD while the game are running.
But in rF2 the files are not closed - so if the game crashes the files on the HD are (very) often damaged and cannot be accessed after the crash.
After my install of rF2 under Steam I had 3-4 crashes which completely destroyed the 3 mentioned JSONs - again and again!
And eventhough I consider the rF2 TM/physics one of the best out there I got fed up with the mandatory reconstruction of the files and the following wheel calibration and graphic setting reconstruction.
So I finally deleted rF2 from my HD.:rolleyes:
And use instead mainly Raceroom (or to a lesser degree ACC and pCars2) - because none of these 3 racing games have such laughable problems. :whistling:
https://forum.studio-397.com/index.php?threads/the-game-dont-start.63297/#post-989146

I smell big BS right there. Mind you, I had a few crashes aswell in the past few years while playing rF2 - mostly while loading a track or while closing the sim and fortunately never while being on track, but I never ever had to deal with the controller.json, multiplayer.json or controller files after a crash. Before writing such nonesense I would make sure that I have an idea of what I am dealing with. And those other mentioned sims have other laughable issues too that I won't mention here, because neither ACC nor PCars 2 are of interest in this topic.
 
I'd imagine the way rf2 loads cars is that each skin is effectively loaded as full car from start to finish. This probably means rf2 goes through each car's folders and packages from start to finish loading decryption, data files, sounds, physics, car textures, pit crew textures and 3d objects one after other (not in that order obviously) per each car and then starts again from 0 for the next car whereas modern games tend to do all those things in batches.
Indeed. If you load a track with 104 identical cars, the same car will be loaded into memory 104 times... when just the once should suffice.
 
I smell big BS right there. Mind you, I had a few crashes aswell in the past few years while playing rF2 - mostly while loading a track or while closing the sim and fortunately never while being on track, but I never ever had to deal with the controller.json, multiplayer.json or controller files after a crash. Before writing such nonesense I would make sure that I have an idea of what I am dealing with. And those other mentioned sims have other laughable issues too that I won't mention here, because neither ACC nor PCars 2 are of interest in this topic.
Calling others posts bullshit and being overly aggressive often points to problems with one self.
So I will refrain from going into discussion on that level.
Have a nice day.:roflmao:
 
I'd imagine the reason for the super slow loading times is a combination of old tech, complex data structures, spaghetti code and big data files with lots of duplicate information. Rf2 also has multiple layers of packaging and each 3d mesh has more data than other games that is used less efficiently (or is basically useless duplicate). Just alone the fact that rf2 uses 4 uvmaps per mesh despite all them being identical in most cases already adds a 4x multiplier to some operations. Add in a mixture of various old file formats like bmp, tga and it all just keeps adding up. Although it is not straightforward to compare rf2 to any other game directly because you'd need to make sure the files going into the games need to be identical to get a comparable result and I have 0 interest doing that comparison. Regardless I'm sure rf2 3d files are heavier than the 3d files other games use. There is not really any more useful data in them. The data is just stored, loaded and used inefficiently.

I think rf2 doesn't have any optimizations (or can not have due to its spaghetti code and packaging nature) so it loads each content one at a time from start to finish. I think only ai files for the cars are loaded together for all cars in the end. Because how rf2 has veh files that can deeply alter the cars I'd imagine the way rf2 loads cars is that each skin is effectively loaded as full car from start to finish. This probably means rf2 goes through each car's folders and packages from start to finish loading decryption, data files, sounds, physics, car textures, pit crew textures and 3d objects one after other (not in that order obviously) per each car and then starts again from 0 for the next car whereas modern games tend to do all those things in batches. I think due to the rf2 loading code being really old being written when computers had something like 2gb of memory and with ssds not used yet so going through each skin individually from start to finish is easier to fit into the memory than working with with fewer larger sets of data. Maybe even reading and writing to the disk constantly to avoid using too much ram.

The game is based on F1 challenge 99-02, which is a game from 2003.

As for the AI pitting problems in wet races; the problem is that the AI picks wet tyres based on two parameters that don't correspond very well. At the start of a wet race the track is not yet wet (0%), but the rain factor is 100%. So, the AI starts out on wet tyres because the rain factor is 100%, then pits for hard tyres because the track is not yet wet, then pits again for wets because the track has gotten wet. If that makes any sense. It can easily be solved by having the track start out wet instead of dry at the start of the race. Then the AI does not pit for wets, then dries, then wets again.
 
I'd imagine the reason for the super slow loading times is a combination of old tech, complex data structures, spaghetti code and big data files with lots of duplicate information. Rf2 also has multiple layers of packaging and each 3d mesh has more data than other games that is used less efficiently (or is basically useless duplicate). Just alone the fact that rf2 uses 4 uvmaps per mesh despite all them being identical in most cases already adds a 4x multiplier to some operations. Add in a mixture of various old file formats like bmp, tga and it all just keeps adding up. Although it is not straightforward to compare rf2 to any other game directly because you'd need to make sure the files going into the games need to be identical to get a comparable result and I have 0 interest doing that comparison. Regardless I'm sure rf2 3d files are heavier than the 3d files other games use. There is not really any more useful data in them. The data is just stored, loaded and used inefficiently.

I think rf2 doesn't have any optimizations (or can not have due to its spaghetti code and packaging nature) so it loads each content one at a time from start to finish. I think only ai files for the cars are loaded together for all cars in the end. Because how rf2 has veh files that can deeply alter the cars I'd imagine the way rf2 loads cars is that each skin is effectively loaded as full car from start to finish. This probably means rf2 goes through each car's folders and packages from start to finish loading decryption, data files, sounds, physics, car textures, pit crew textures and 3d objects one after other (not in that order obviously) per each car and then starts again from 0 for the next car whereas modern games tend to do all those things in batches. I think due to the rf2 loading code being really old being written when computers had something like 2gb of memory and with ssds not used yet so going through each skin individually from start to finish is easier to fit into the memory than working with with fewer larger sets of data. Maybe even reading and writing to the disk constantly to avoid using too much ram.

BMP or tga are not used, except from dds, png is used for animations, otherwise i think ur on point.
Map layers from what i see are pretty much the same in AC at least for tracks idk about cars. An important difference i ve seen however is object nesting, i suspect AC (not even ACC & UE4 which is very optimized) fast loading times are because track loads 7-8 big kn5 files and then game compiles them into a track with various objects with separate properties, while in rf2 mesh grouping is minimal usually, eg. if u need a different LOD out u need a different mesh file as well, same with textures AC uses large multimaps a lot more than rf2 so the number of files need to be loaded and allocate memory for (along with memory pointers) is huge in comparison.

Also rf2 1st of all needs to extract all those files from largely compressed package file(s).... rf2 methodology is for benchmarking!
Since servers usually have their own mods u need to download i dont see the point for packing and un-packing all the time, and while compression saves disk space u also have duplicates in the steam folder.

On the other hand high bandwidth is common, i downloaded RRE ~32gb in what took me more than 1.30 hours only to delete it after a few minutes...

Rf2 still the best sim out there (i only haven't played iracing).
 
I'd imagine the reason for the super slow loading times is a combination of old tech, complex data structures, spaghetti code and big data files with lots of duplicate information.

<snip>

And some of that data is encrypted as well.
I suppose at least multi-threading all that loading so it takes advantage of multiple cores would really help.
But I vaguely sense it is not in the priority list...
 
RD: (question from @Will Mazeo) - I'd like to know about more pit stop work… Mandatory pits with pit window (u got 2 official GT3 packs, this is a must have) / Min pit timer stop (like ADAC GT)?

MO: So that’s a few questions. I think our plugin system is powerful enough for leagues that want to follow the rules of specific series to implement those (mandatory pit window, minimum time in the pitlane).

This is good news for leagues I suppose, but what about Single Player? :cautious:
For one, mandatory pitstops could be a workaround for the issue of AI pitting unnecessarily when Fuel and Tire usage are enabled.
 

Latest News

Are you buying car setups?

  • Yes

  • No


Results are only viewable after voting.
Back
Top