An Attempt at a 2026 NORS Mod

Phantasmicality

Phantasmicality

Test Rank
Joined
Sep 7, 2022
Posts
23
Reaction score
17
Points
3
Hey, so I mostly come from a modeling background (i.e. roblox) and I decided to start doing this mainly because why not and give a wack at it

So far I have not gone into any importing or scripting for the mod yet, but once I get the models to a certain point I will get into that. and try and document my thoughts on that and my experiences on modeling stock car models.



1691985673120.png
1691985676242.png
1691985655935.png
1691985702494.png

I'll be honest, I'm not sure how far I'll be able to go. I've had previous instances of trying to install software for modding that wouldnt cooperate.
 
Good start. Will defintiely need to lower the poly count for NR2003 significantly. Reducing the redundant verticies for example.
 
Mystical said:
Good start. Will defintiely need to lower the poly count for NR2003 significantly. Reducing the redundant verticies for example.
Click to expand...
yeah thats something i'll definitely work on once i get somewhere with actually being able to put a model in NR, current goal with the models is to make them to a good quality and scale it back with optimization from there.
 
Second post of this and some big things on the model has been happening. I've got the Camry model to the point where I want it and have gone ahead to unwrapping and starting a test template as well. Having not made the other bits that are supposed to go on the template I've left some space so I can figure out how to configure everything, having said that, I've made some... interesting decisions.
1692486626026.png1692486604746.png

(Current tri count is 6,482)
Having not made the other bits that are supposed to go on the template I've left some space so I can figure out how to configure everything, having said that, I've made some... interesting decisions.
test uv.png

In the meantime I also have continued on working on whatever I've had on the interior. Pretty much all that has been added is a roll cage that I created by using curves.
1692487477281.png1692487516333.png


Now onto to the future of this project; I start college on Monday. Therefore, in order to accommodate time for college and such, I am going to pivot it the project towards 2024 with the some of the new mustang models being unveiled.

Next time I post it will most likely being updates on interior and a Camry model inside of nr2003.
(I still have no idea how to write these types of forum posts 😭)
 
The nose UV islands would definitely be very painful for painters to connect all their details together (if they are not using a 3d paint program). A little warping but having it all connected together would be a lot easier for painting than that many segmented islands.

As for UV space if you need more you can make non-square UV template as well. Like with ICR I did 4096x2048 for painters, then compress it down to 2048x1024 for the game. Having more width in template makes it a lot easier to fit all the UV car islands than a regular square.
 
A little warping but having it all connected together would be a lot easier for painting than that many segmented islands.
Click to expand...
I've already tried the whole thing unwrapped, and theres a bunch of warping all over the place (blaming the camry nose for being so angular). So taking into account about painters being confused about what goes where, I've grouped the two paintable islands together and grouped the manu nose bits together a bit downwards.1692499925904.png
 
Well, it's been an awful long time since I've last updated this thread, and an awful lot has changed; within recent months the entire modding process has changed dramatically, which I'm quite thankful for.
Trying to navigate an ancient version of 3DS Max 8 was quite frankly a nightmare, and probably the biggest bottleneck in terms of why NR modding has stagnated so much.


Having said all that, the top 3 series need updated mods desperately.


Introducing An Attempt at a 2026 NORS Mod.
(working with Chris Lilly for these)

1775685047693.png

Obviously the first matter will be optimization, but there are some other points of matter after receiving feedback from the roblox community. Includes hood island shaping, warping, and adjustment to fit for a driver template.
 
Welcome to the world of NR2003 (car) modding. The ever consuming project where by the end of it you feel like you were in the Nascar shop itself building a car but the payoff is extremely huge, exactly what you want to race with in NR2003.

The papyNR has truly opened the door for more people to add mods to the game and its great to see. My vision and hope was this forum would be a gateway tutorial hub to unlock more potential in the community and its great to see.
 
Mystical said:
Welcome to the world of NR2003 (car) modding. The ever consuming project where by the end of it you feel like you were in the Nascar shop itself building a car but the payoff is extremely huge, exactly what you want to race with in NR2003.

The papyNR has truly opened the door for more people to add mods to the game and its great to see. My vision and hope was this forum would be a gateway tutorial hub to unlock more potential in the community and its great to see.
Click to expand...
Considering I built pretty much the entire thing from scratch, I pretty much have lol. Am so so SO thankful that PapyNR plugin has gotten to this point.
 
This looks very good. Away with the "Ifs" and "Buts"! Please go on, it's good to see!

(quick shout-out to Burnout3D as he laid the ground in blender; what goes around, ...)
 
Late as hell post because I've been spending the last few hours post midnight figuring out how to use make3do by myself, and I've encountered an odd problem, which I've tried a few remedies and none have quite worked.
Screenshot 2026-04-19 021556.png

The car does not translate a scheme onto it. I've tried renaming the paintjob_521 material to paintjob (breaks the make3do), I've tried joining models together. Clearing unused UV maps. I'm quite frankly stumped. For clarity, I've included a shot of my code for building the model and what my materials look like.

1776585960525.png1776586031579.png
 
for RENAME_MAT change it to:

RENAME__MAT (paintjob_512 paintjob, window windshield)

if interior and rearview sections has a RENAME_MAT to load a pas make sure it reflects the same paintjob name. Essentially we are telling make3do "paintjob_521 is actually 'paintjob' which will load my paint scheme for my mod when I use the import button in NR2003"

I've not tried an underscore in the name for a paintable texture but it does work fine with other textures I've used so I don't think that is the issue.

Also for NR2003 no single mesh can have 2 or more material (textures) assigned to it. So you try to combine as many meshes together that use the same texture.


not related to the paintjob issue but just some feedback, I notice in your script your non-damage model for the mod has many many pieces, each separate piece is a separate draw call the game will need to render your mod (aka more performance heavy) plus it will make lods a huge pain having so many meshes to deal with for each make. If you are able to combine more meshes as long as they don't go over 1k verts that will save you a lot of time and pain later down the road. 6-7k verts total for a mod on LOD0 is pretty high but doabale ( gamemay crash in replay mode on heavy track with full field in view), if you can have a mod 2-3k vertd for LOD0 that will ensure the most optimal fps friendly mod even on really heavy tracks like Daytona BBMC.
 
The quick background on the naming convention of materials is that the sim knows 2 names only (for cars):
paintjob --- sim sources the base tex from .CAR file​
windows --- allows sim to change base tex by dirt severity (dirtscreen buildup)​
Our common accomplice, make3do.exe, will not write any mat'l name unless it is explicitly instructed it to do so:
p0e: MESH_GROUP "s0e.pas" RENAME_MAT (paintjobNamedInPASfile paintjob, windowsNamedInPASfile windows)​

***
Mystical
In blender, there's no need to separate any material from a mesh just to conform with game restrictions (i.e., a mesh always & ever holds 1 mat'l).​
make3do.exe handles that---it is splitting meshes on its own.​
 
Did a couple more steps, found the original blend scene I had used for porting over the materials, that also did nothing. I had a hairbrained idea of putting the make3do for dev1b (the current model) into the original dev1 (basic NORS car with COT wing) and it worked.
My personal working theory is that whenever i copied over the NORS26 folder I had to start working on the 2nd import of it I broke something, because slapping the make3do in the original mod folder works fine.


Mystical said:

not related to the paintjob issue but just some feedback, I notice in your script your non-damage model for the mod has many many pieces, each separate piece is a separate draw call the game will need to render your mod (aka more performance heavy) plus it will make lods a huge pain having so many meshes to deal with for each make. If you are able to combine more meshes as long as they don't go over 1k verts that will save you a lot of time and pain later down the road. 6-7k verts total for a mod on LOD0 is pretty high but doabale ( gamemay crash in replay mode on heavy track with full field in view), if you can have a mod 2-3k vertd for LOD0 that will ensure the most optimal fps friendly mod even on really heavy tracks like Daytona BBMC.
Click to expand...
Yeah i understand, I'm not quite sure how nr handles making sharp edges (marksharps in blender) so i've been spliting meshes manually. If i can figure out how it handles that out then I'll be able to shorten up the part counts.
1776613888007.png
 
TShirt said:
Mystical
In blender, there's no need to separate any material from a mesh just to conform with game restrictions (i.e., a mesh always & ever holds 1 mat'l).​
make3do.exe handles that---it is splitting meshes on its own.​
Click to expand...
Interesting, I think I ran into issues when I had meshes that were sharing more than one material so I always jusy keep the meshes separated and joined together of like-minded textures. For example a wheel where I have the rim that is printable and the rest of the tire on a different texture it wouldn't work correct in-game so just split them apart. But maybe I did something wrong.

Phantasmicality said:
Yeah i understand, I'm not quite sure how nr handles making sharp edges (marksharps in blender) so i've been spliting meshes manually. If i can figure out how it handles that out then I'll be able to shorten up the part counts.
Click to expand...

Yeah in blender simply marking sharp your edges and smooth shading will work.

Phantasmicality said:
The fact the game crashed when i changed the camera on the pace car to face the infield definitely proves that the LODS will need to be quite extensive lol, but got cars on track.
Click to expand...

Whats your vert count at LOD0? (wheels in scene enabled as well)? Generally 6-7k is pushing it, recommended is 2-3k verts is best if you want the most versatile mod that can have any aggressive replay camera view on it even on highly straining performance tracks with extreme detail (like Daytona BBMC). You can start with LOD0 being 6-7k but your lod levels below have to be much more aggressive and sooner as a zoomed in camera forcing more than half the field to still stay in LOD0 will nuke the game for sure. Also take into account your damaged state of the mod adds an extra 500-1000 verts as more pieces have to be split and you have to model interior sections that aren't needed to be shown in the damaged view. On the flip side this also means if you have interior sections you never see in undamaged state you can delete those to save a lot of vert count.

My ultimate stress test is running my mods in development on Daytona BBMC with all graphics settings enabled (except shadows from cars on other cars as it always looks buggy) on a 32;9 resolution ultrawide monitor. If I can't crash the game in that scenario then it'll work for 99% of all players.
 
Last edited:
Mystical said:
Whats your vert count at LOD0? (wheels in scene enabled as well)?
Click to expand...
If i did the import of the wheel.3do properly, it looks like it'll be roughly 8,757 triangles.
1776635397835.png

I did most of the chassis detail in the event of the car coming on its lid, but most of the detail would be a waste.
Mystical said:
You can start with LOD0 being 6-7k but your lod levels below have to be much more aggressive and sooner as a zoomed in camera forcing more than half the field to still stay in LOD0 will nuke the game for sure.
Click to expand...
This was the intended goal, I'm curious as to how the LOD parameters work, and how LODs work with damage in general, that's the part i'm the most nervous about how to tackle.
Mystical said:
My ultimate stress test is running my mods in development on Daytona BBMC with all graphics settings enabled (except shadows from cars on other cars as it always looks buggy) on a 32;9 resolution ultrawide monitor. If I can't crash the game in that scenario then it'll work for 99% of all players.
Click to expand...
I'll probably try and replicate this later on to gauge the stress levels of the mod. Thanks for the feedback
 
Pulled together a LOD check mod, threw in some values that were very tight

LODBOX: STATIC_LOD (
0 BOXL1,
1 BOXL2,
3 BOXL3,
5 BOXL4,
7 BOXL5,
9 BOXL6)

It'd probably be quite a bit to draw but I'd imagine for my mod atleast I'm going to have at least 7 LOD states

(friend recorded this while I was messing around with freecam)
 
Something we discovered while pulling the lod script together was that instead of overwriting when you declare a mesh(?) variable twice, instead of overwriting, MAKE3DO blows up and yells something about "multiplying" stuff together, something along those lines
 
yeah multiplier mean you used the same alias twice in the psg script.

For example

L1_Body_main MESH maxcars mesh_body_L1

then later in your script you use the same exact alias again even if its a different mesh from your pas:

L1_Body_main MESH maxcars mesh_chassis_L1

you'd need to rename the alias to something different
 
Fiddling around with figuring out wheel positions, 1776811887337.png

Couple realization I figured out:
Wheel placement in PSG doesn't affect physics (maybe?) didn't notice anything
When configuring make_a_ui.3do, should probably just cut out the other LODs so only LOD0 shows up, or you could jank up the LOD0 distance

I don't recall exactly, but I believe the smiffsden indycar mods had wheel damage, if that's possible, theoretically how hard would it be to have damage that could associate to power loss, say, if the entire front clip comes off?

Current status of the LOD0 model, I swear I've spent more time trying to optimize this than putting it into NR.
1776812765194.png
 
Last edited:
You must log in or register to reply here.
Who read this thread (Total readers: 0)
    Back
    Top