A more core suggestion: Besides the ability to import the UV map from an obj file, it should also be possible to import the materials. Not any settings of those, just a material name per vertice. No one well versed in dedicated 3D modelling tools wants to go through their model again and reassign 300000 vertices individually.

Quote from martin18 :

Have you considered maybe like a Honda J series swap instead those motors are light and without boost they can make 400 Hp and they range from a 3.0, 3.2, 3.5, 3.7L and they can rev up to 8,000 rpms, but they make good torque and great HP built or maybe the Alfa 155 v6 Ti. https://youtu.be/WHjs-pcyg8Q

https://youtu.be/wysO0nj2bMQ

yeah I want to keep the "Saab engine" just for giggles, its all fantasy number games anyway, tho going with what I'd have as basis, and what I'd want in the end, the comparison to the 155 TI is probably the most fitting. Take the base engine variant with the biggest stock bore, bore it even wider, reduce the stroke in an effort to keep mean piston speed in a range that everything keeps working at (basically the secret to screaming F1 engines - big bore, tiny stroke), make everything else from exotic expensive metals - boom, high rpm screamer.

Quote from martin18 :

Wow this mod is really getting wings and taking off (no pun intended) love that rear diffuser. Just question those engine specs is that what you want? Peak torque and peak HP all the way at the top of the rev range? Not going to be sluggish during take off?

Engine specs aren't final, but it is my goal to create a car that has different driving characteristics and feel compared to the old GTRs. Wringing ~400 hp out of a smallish V6 will create a narrow powerband, but maybe having that power instantly without turbo spool is advantages in certain situations. Lighter smaller car with more effective aero may yield better braking performance.

when I press S I can play Lego with some frame, but there's nothing about suspension

Quote from timdecnodder :

Another car to join the GT2c

hopefully the full GTRs too

did rethink the "roofscoop" question, and came up with this. Somehow more elegant than the previous one I had (this is the exact shape the sharkmouth opening has)




this Editor is so unintuitive.. or probably just 30000% different to how I did car creation for 7 years in AC Like can anyone tell me how the suspension thing is supposed to work? In AC you put coordinates for each anchor point... now here I can set lengths, but no anchor points.. I planned on using the rear suspension I developed (with help by people that know what they are actually doing) for my KRB Lotus, Group-C style rear with anti-squat and anti-dive ... but it seems like LFS does not even allow for inclined wishbones..?

Some good points in the post above. I guess so far no modder went as far as developing custom suspension for their cars, even those with "fully custom" mesh seem to use the XRT or another LFS car's physics (suspension).

Might be worth to add a check for these "rebodies", just an algorithm that compares a mod cars underpinnings with a database of yours, and add s flag to the car's page.

Regarding the "hardtop" idea, this is what I made the doors of the car.. hence no hardtop here really, rather a full structural roof

(old 250k tris version)

ofc no use of that in LFS, so I started working on a new much simpler interior. First try at shallow lights like5tag suggested

Huh, mighty impressive!

Well there's the Blender skin render kits in the 3D render section, you can work off those.

Nice work! You could try setting up sharp edges (in the Shift-E menu, plus the Edge Split modifier) to work out your desired shape more clearly without needing more tris.

Quote from ClasK_03 :

In my opinion too much time shouldnt be spent on improving the modeller. Most people will probably use other 3D software to model.

What is very wanted by everyone though is better texturing. Project from view is not the best UV approach in all situations. Even better would be the ability to import both UV maps and textures from other software.

This imo is really important. Not just for the base skin but also other stuff. Like many cars have convoluted carbon fibre parts that could easily be mapped with some small tiled texture, but using project from view would yield awfully stretched results. In Blender, if set up well with seams, making a minimum stretch mapping needs a single click.

Those sorts of acrylic lights would probably appear a bit poorly with LFSs Shader system. Maybe I'll try something, but my vision of the car is mostly set. Same with the roof scoop, in the old thread I linked you can see the scoop I had on it before. But whole engine compartment is to change, as what I have is much too high poly and a bit unrealistic anyway with its open standoff injectors (which were just an AC scripting experiment mostly). I managed to find measurements of a "Saab V6" (X25XE) and transversal seq box for the drivetrain. Probably not going to end up in the final model, but I like to fill my cars with their needed components. Front shark mouth contains the radiators and coolers.

Driver fit well in AC, but there I could scale it. But iirc I did model the roof section with a helmet in mind, it's why in my roof version I continue the windscreen shape in an arc instead of extruding it flat to the rear like in the other render.

Quote from lfsrm :

oh well, that modifier does indeed completely separate the mesh geometry.. I don't know if it's ok with the actual game engine, I am pretty sure it wouldn't do well with the future real time shadows if you want to do something for the long term.

it's what other game engines do as well, at least Unity and the Assetto Corsa engine (or any that use the FBX format, as that one will separate the mesh just the same as I did here). I'm not even sure any game engine doing hard edges does it any different

I don't want to destroy the original RAC bodylines too much, as (if Scawen implements UV import), one should be able to use normal RAC skins on this car with minimal, if any, modification needed to the skin. As such I'd like to keep the headlight array, but use the lower opening for the brake cooling ducts instead. Fender venting rather at the rear than on top.
The one I have in AC has a roof scoop, but I really don't like it much and would not put it into this here. It will be a smallish V6, no need for huge air masses. Side scoops slightly widened instead.

I aim at making a competitor to the existing GTR cars, open roof would not fit.. neither do I know how to do that yet. Or anything really in the Editor

Quote from lfsrm :

Interesting, I didn't know that you could export smooth groups directly from blender!

No, apparently you can't right now, that exactly was my result of those trials above. The moment I hit "Apply" on the Edge Split modifier in Blender, the mesh is split up according to the smoothing settings I created with those sharp edges. The LFS Editor imports this raw mesh with its Normals and bundles them into Smoothing groups according to each meshs Normals' orientation.

so ofc I tried another thing right after that post and sending the email: I applied the Edge Split modifier (which will split up the mesh and duplicate vertices at the desired edges), and then in the export settings unticked all the boxes under "Geometry" - that somehow made it import fine. So I assume Blender's exporter is the culprit here



thanks for the time again

Quote from Scawen :

I'm not sure how familiar you are with LFS modeller, but a smoothing group is a property of a triangle that allows it to share normals ("be smoothed with") neighbouring triangles that have the same smoothing group.

Consider a cube. It has only 8 vertices in the editor. Each corner doesn't have 3 vertices with a different normal (it does in the 'output mesh' that is ready for drawing, but in the LFS editor model there is only one vertex at each corner). The 3 faces that meet at that single vertex must have different smoothing groups, so that the cube has sharp edges. A cube needs 3 smoothing groups in total. A cylinder would use one for the curved surface and one for the ends.

When you export from blender, you do produce multiple vertices at a single point if there is a sharp edge there. LFS will merge these vertices into a single vertex, but only after analysing the mesh and assigning smoothing groups, so that the sharp edges are retained after processing.

When it says it ran out of smoothing groups, that means that it had difficulty assigning a smoothing group to one of the surfaces. This may mean that a lot of different smoothing groups shared an edge with that surface. It might not be a real problem - it might just mean there isn't a sharp edge where there should be.

You might need your human brain to fix it, if it is a problem. The thing to do, is go into "tri" mode in the modeller, and select "groups" at the bottom right, and see how it has assigned the smoothing groups (they are colour coded). You might find a clue there, maybe something where there are loads of smoothing groups but it's completely unnecessary. It's hard to know. I had a few test meshes for Blender exports but probably haven't covered everything that could go wrong.

Running out of smoothing groups means you reached 24 smoothing groups. But normally a mesh only needs 5, 6 or 7 groups. In 'subob' mode there is a 'compress groups' button to reduce the number of groups without changing the appearance. So it's worth having a look in there to see what's up. Maybe a lot of smoothing groups meeting at a point so LFS wasn't about to assign low group numbers. Or you could send us the OBJ file if you'd like me to have a look.

thanks for the in-depth reply. There's something odd going on with the smoothing groups then, as this is what the "groups" mode shows:

the uninterrupted mesh is the part of the model from that old "Blender skin render" package, and the harlekin mesh is everything I did new.. I'm sure this is more an issue of Blender than your Editor, given this is at its core some very old mesh that moved through generations of Blender through the years.



ok, went about trying to do the "copy the old mesh into a new Blender instance and join it to the default cube" trick, which is regularly advised to clean old mesh data of superfluous leftovers.. it didn't quite work as intended:


send you a mail via contact with the files. Maybe there's some weird quirks with the OBJ exporter going on, I'm on 2.93.0 beta

I don't quite understand what's a "smoothing group" in this context.


In Blender the model has all double vertices deleted, "sharp edges" set (those in cyan), but not applied on export (the original "skin render" models distributed ages ago had this same sharp edge setup)

Is the X-shape of the front grill intentional? Really nice believable design, hats off for that

Yeah you can basically ignore that lower half of that screenshot as I will not use it, everything is much too high poly for the LFS guidelines (full car is 250k tris)
I will splice in my new aero onto an "original export" of the RAC model and make a new simpler cockpit.

Quote from Scawen :

The trouble is we have received one or more:

- mods that are just LFS cars with slight tweaks
- mods whose license we need to question
- mods that we aren't certain are 'original work'
- mods which really need more work - why not improve it a bit before submission
- mods that look like a good adjustment but could be one of many of the exact same idea

Our current thoughts are that the mods should go through a process of community evaluation and discussion before being submitted.

This was first mentioned by user luchian https://www.lfs.net/forum/post/1969335#post1969335

Right now we're considering trying to make this a workable system and a way to sort out these issues above, and remove the problem of an 'authority' (group of reviewers) making the decisions.

Current plan is, using the mods work in progress forum, a user could post their ongoing work there, in early or late stage. They might receive any kind of comments, help, suggestions, approval - depending on how advanced a modder they are. Anyway in the end, maybe aided by a voting system, it should become clear if the mod is ready for submission.

On the Submit a vehicle mod page there could be a box where you must enter the forum thread where the mod has been discussed and generally approved by the community.

How does it sound?

I've been in the AC modding community since ~2014, and this same question came up a lot, especially recently in the Race Department community, which as its problems with people ignoring the rules and clueless moderators, as anyone can upload new mods without prior review.

I'd guess that "have a valid WIP thread and link to it" would already filter out a lot of rubbish, and would give you, or whatever group of reviewers you establish, a quick information source. You just have to hope that modding culture on this forum develops in this direction, into one that values scratch work.
The Modding subforum is only open to S3 users, right?

Quote from nikopdr :

The real thing has an aluminium chassis with carbon body shell though, so there is not much weight to take out of anymore, i would expect every kilogram lost to come back when installing hardtop and a rollcage


Please keep the saab engine though. IIRC real one had a B204, you can shed almost 10 kilograms off that block if you remove the balancer shafts and it's chain pulleys

you probably got a good point there with the car's weight

With the engine, the combination of that rear axle with turbo lag made the RAC so hard to master, so a high revving N/A would make it more predictable. Also, we got so many turbo-4s already, spice it up a little Saab did use Opel/GM X25XE V6 is some of its models... I could brew up a hypothetical formula version of that... 10k Spiess-build V6 anyone?

Quote from Van Sterberkt :

Whaw cool! Super curious.. I saw the RAC GTR screenshot/impressions a few years ago, was that your work as well?

this was mine: https://www.lfs.net/forum/thread/92078-Enhanced

I remember this render from like ten years earlier, I didn't make this, had no clue about modelling then:

Ok, so some may know that I did work on this some years ago and made a thread in the render section. I did not make this in the few days since the announcement lol. More a heads-up that this is out there, I can absolutely not tell if I can get this to work and if I release it.



The fictional specs I cooked up for this were 300 hp, 725 kg weight, aero somewhere between GT4 and GT3 of comparable size. Engine is a high revving V6, similar to older Formula Renault VQ-based V6s. Suspension wise I kept it with the odd no-antirollbar at the rear just for giggles.

Some words of wisdom:

You're much better off learning the craft yourself than waiting on someone fulfilling your request. Those that can create this stuff don't sit around, not knowing what to do. And collaborating on these things does not mean one person writes down an idea, and another sits down for half a year actually making it Literally anyone can come up with some fantasy specs in 5 minutes.

I don't want to be a party pooper, but I know all this from the Assetto Corsa community.