We'll only be able to simulate/mod from the late 90 early 2000" era, as LFS doesn't support modern complex electronics (yet).

Classic cars should be possible too, but we need old type of road/slick tyres if we want to mod something like an 70' F1 or touring car accurately, unless you could reduce the tyres friction/grip of the default compound which is an approximation but should be ok for the time being.

Quote from Tazka :

Wow that's really nice! You got some talent on car design!

Thanks , I just wish I could bring this project to completion.

I forgot to mention the car dimensions :

Length : 3.90m
width : 1.81m without mirror (2.01m with it)
height : 1.35m
wheelbase : 2.40m~

Last edited by lfsrm, Sat 6 Nov 2021, 01:04.

Hi,

I thought it could be a good idea to start something to kill time before the mod support test patch, so I started a project to test some ideas and why not a practice opportunity since I am not super comfortable in making low res assets.

It's a sort of concept for a newer version or a replacement for XFG, basically it's your typical everyday 2002~2008 small city car with a relatively powerful engine (GTI).

I am still undecided about of the engine configuration or other technical aspects, let's say it will have around 170 hp max (for now).

So here are some pictures of the design, I will probably change/add some details later when I make some progress but this is mostly the final version.

Quote from Federico Testa :

is anything like this then possible to be added for instance to mods? https://racesimstudio.com/gt-ferruccio-55
Of course made from scratch...

Totally douable.

Nice trailer, please consider to add the ball as default autocross object !

Quote from Bose321 :

I believe car manufacturers have rights on the design of a car, so that won't work. Only in China this would be allowed.

But I can imagine it is possible it slips through the cracks if someone checking the upload doesn't pay attention.

That brings up a question, what if an approved mod has to be taken down, for whatever reason, what happens then? Does it stay on our PC's if it has been downloaded and just get removed from the server which won't make it useable anymore?

I think it would be a problem only if the modder is selling the mod, changing the logo and the name should probably be enough.

Quote from ACCAkut :

Not being able to import UVs would be a serious oversight. But looking closer at the videos, I doubt it would be in any way feasible to downgrade these models to work in the new Editor.

I mean, those of us that have worked with Blender/3DSMax for years now would not want to switch to something else.

Well since you can import the base mesh I am somehow more positive about this, as modeling inside the editor is all but intuitive.

Personally I can manage with the editor texturing system, but people should not expect the same result as a fully modern texturing software (eg : heavy use of tillable without any variations, stretched textures etc ).

Quote from ACCAkut :

Ok, I will have to read up on everything in this thread here. As someone very active in another sim's modding scene for close to a decade now, it certainly reads like a great approach, especially the way mods need to be approved before they can be used online. My bunch could be ready rather soon. Apart from the tris count they fit your requirements rather well already. And ofc there's also the RaceAbout GTR resting on my HD for many years now.



Also can't state enough that you should not make it too easy, or else your servers will be flooded with 63485434843 Nissan Silvias ... ehm, "Nussian Sovokias", as in models ripped straight from Forza, Need For Speed and whatever games there are out there.

You have to redo the texturing part (Uv'ing) into the LFS editor, because you can't import the UV data.

Quote from Scawen :

I am currently coding for a limit of 65535 triangles, or 65535 vertices, in the combined output model, including all subobjects.

This is massively higher than any official LFS car at the moment, so should be fine if people adopt the right approach. You need to make a 'game quality' mesh rather than a 'film quality' mesh, if you see what I mean. There are technical reasons to stay below 65536.

To change the code to overcome these technical reasons would be quite a change and, although possible, doesn't seem a good idea at this time. The fact that you can already make models hugely more detailed than any LFS model, seems more than enough for now.

I'm already concerned, with this high number, that we will get a glitch when you leave the pits and your car is generated, and also each time your vehicle is damaged, the vertices must be sent to the graphics card again. But this change is all about allowing people to be creative so I'm trying not to get in the way of that!

I think this limit is totally fine for the time being , it will force to make use of textures for fine details instead of meshes.

Speaking of textures, I just really hope that you could implement importing UV maps/textures with the base mesh.

Quote from MacedoSTI :

yes i see on video is possible export in obj xD

now just wait for the update

about a authorization to get a real names for mods? like subaru/volkswagen/toyota... how it will works?

No you can't use real brand names without an authorization ( from the manufacturer or the license holder ).

And to get that authorization, it means that you are ready to buy a license that will probably cost you +50 000 $ for one vehicle (if they accept to negotiate with you in the first place), there is some exceptions but this is how things work.

So you should always make slight modifications to your subaru or toyota model to make them not as similar as the original model, and of course make sure to use a fictional brand.

Last edited by lfsrm, Mon 1 Nov 2021, 14:58.

Quote from MacedoSTI :

I hope is possible use other 3d applications, like 3dsmax... blender is...

Any 3D app with an OBJ exporter should work for the base mesh,

and nothing wrong with blender .

I only wish that we could make the texturing part in Blender too, because from what I saw, everythings except the base mesh should be done in the editor.

I mean planar projection isn't very practical especially for complex shapes.

Allowing for UVs export is something that maybe you should consider in the future.

WOW! plus the top notch integration !

Thank you Scavier !

Quote from Eclipsed :

Few messages and the speculations circus begins...
What's the point worrying so much? LFS still works very well and when something get's updated,it will be fun to enjoy it. Until then let's just have fun with racing or have other hobbies.

Let's not be harsh with how other people perceive the thing, I must admit that with the usual quality of Scavier work it's hard to contain the excitement about an incoming update .

At this point any discussion (within the rules) in this forum is a happy event in my book, it shows that there is still an active small community that cares about the game.

That's good to read ! maybe mods support with an editor for imports/exports ?

I am pretty sure that Scawen is the first to admit LFS needs better presentation (graphics), that's why they are putting much efforts into modernizing the core graphic engine.

I could add one or two pretty neat features (important one IMO) which is ingame integration with modern social/streaming platforms, an ingame hub with realtime spectating and news about the game itself, leagues or other organizational events, pretty much like what other massive multiplayer games are doing (like Dota or CS), no other simulations are doing this right.

I think this could make LFS the true and the most modern multiplayer platform for simracing

Automation and BeamNG is very similar (but more complete) to the LFS and tweak combo, I wish LFS devs could implement something equivalent with better and seamless integration to the probable modding tool, this could open endless possibilities.

Always happy to hear some news from LFS devs, the new south city is truly becoming an "open world" map. cruisers, racers, and layout makers should be happy about that!

I think releasing the patch without Fernbay is a good decision, we should have plenty of content to play with in the first version until further improvements or new track/cars.

If it means the beginning of some sort of integrated "tweak" into LFS, where we could customize every aspect of the car (tyres, engine, suspension etc..), I can only agree with this approach.

Last edited by lfsrm, Sat 9 Jan 2021, 13:12.

It will be so much more easy for Eric if LFS supported multi-layers textures, because as of now he needs to add some sort of "floating" geometry with alpha textures if he wants to add tags or other type of imperfections, which will of course make the scene heavier on the hardware.

Personally I am not that bothered with a "too clean" environment for the time being, those kind of things can be easily updated after the public release, I am pretty sure that even after the final release Scawen will continue to update the graphic engine.

Quote from nacim :

Hello !
Nice type of update, feels like dev notes, really makes me discover the complexity of the engine and tool set, love it !

On a more artistic note, Eric you are rocking it ! Your texturing work is impressive !
I feel like you are using hard edges everywhere though, any reasons you don't have some nice bevels on the edges of buildings or sidewalks for example ? I feel like it would create a better surface for the lighting to "catch on", and create a smoother, less video gamey feeling
Also like three_jump said just above, I really feel like everything is just too clean, older building could benefit having some grime and dirt applied on them

I guess Eric is more concerned about polycounts, Bevel can easily double or triple them depending on the complexity of the object.

There are some techniques to fake them without affecting the base geometry (bevel shader ?), but I never saw this kind of shaders in a video game.

Quote :

I found two small things in a screenshot, see the attachements, it looks like the occlusion is a bit too strong, maybe it isn't aware that the mesh for the fence is transparent and creates and occlusion like an opaque mesh ?
And a small shadow light leak, like you already fixed tons of already

Anyway, as usual, keep it up guys!

Same with the image I linked bellow, but I prefer a bugy vertex AO over floating objects.

Anyway @Scawen it's always good to read your progress reports, thanks for keeping us updated! and I hope 2021 is the year for us to try all those goodies !!

I think it's perfectly douable for only one man to make the core physics engine, Stefano Casillo (ex Kunos physics/graphics main dev) is the perfect example for that, I guess even SMS and Iracing have only one or two guys working on such tasks.

Audio, networking API and graphics assets are the most dependent on staff numbers, sometimes specialized developers for specific domain like AI and VFX are often needed too.

South city becoming a true city ! I see why it took them more time than usual.

Hopefully the core graphics engine is near completion.

LFS suffer from the comparison with the larger scale simulations,we should take it as it is, means as a small scale indie project.

I guess Scavier know that and they are probably transitioning from the everythings in-house mindset, as it is impossible for them to compete directly with other simulations, I think they are slowly shifting LFS to a semi-customizable platform model, the modding discussion started by Scawen can prove that.

Quote from Racer Y :

"Environment design/building is definitively one of those area where you'd better work with a big team." Uh I'm involved in these things all the time at a certain level. I'm currently involved with a "big team"project. And it's a total cluster. And you could say the same thing about coding the car - this guy figuring out the torque and that one the window tint. I dunno. I'm inclined to think track too. But the intricacies of the car?I have no real clue whatsoever as to how they make either one.

I mean it's better if you have ten artists building assets at the same time than one, unless you are not constrained by deadlines, it doesn't have the same complex management as other area like core programming, which is a nightmare to coordinate.

I do agree about cars physics, it's generally a one person work if we only take into account the simracing genra, but it's totally possible to find persons who understand your work and know how to apply numbers ( well some actual physics devs started as modders ), I mean if you want to make a 200 cars racing sim you definitively don't want to work alone applying and testing data for all those cars.