XeNTaX

mrstyx wrote: ↑Wed Jul 17, 2019 11:25 pm Can anybody help on finding the positions float ? I'm trying the one on 0x55BC (for one mesh obviously) but It don't seem to be the right one.

(Sorry for the late reply.) I don't get it. It's no floats, your starting address is wrong but you get the correct mesh?

(Assumed vertex count is 5846? (max Fi= 3439), but still struggeling with the face indices (as always)...)
. (needs a scaling of course, minor prob)

Not fitting FIs might be a problem of double vertices? Though they seem to belong to different faces here.
...
v 135.234375 127.957031 253.925781
v 135.234375 147.058594 252.375000
v 135.234375 145.660156 243.265625
v 135.234375 127.957031 244.703125 <<<
v 141.136719 127.957031 243.398438 ###
v 176.808594 145.460938 241.976563
v 176.808594 127.957031 243.398438
v 141.136719 127.957031 243.398438 ###
v 135.234375 127.957031 244.703125 <<<
v 135.234375 145.660156 243.265625
...

edit: what the heck did I do here?
It's like so:

shakotay2 wrote: ↑Fri May 14, 2021 3:22 pm

shakotay2 wrote: ↑Sat Apr 25, 2020 10:36 am

mrstyx wrote: ↑Wed Jul 17, 2019 11:25 pm Can anybody help on finding the positions float ? I'm trying the one on 0x55BC (for one mesh obviously) but It don't seem to be the right one.

(Sorry for the late reply.) I don't get it. It's no floats, your starting address is wrong but you get the correct mesh?

(Assumed vertex count is 5846? (max Fi= 3439), but still struggeling with the face indices (as always)...)
.
kindaWheel.png(needs a scaling of course, minor prob)

Not fitting FIs might be a problem of double vertices? Though they seem to belong to different faces here.
...
v 135.234375 127.957031 253.925781
v 135.234375 147.058594 252.375000
v 135.234375 145.660156 243.265625
v 135.234375 127.957031 244.703125 <<<
v 141.136719 127.957031 243.398438 ###
v 176.808594 145.460938 241.976563
v 176.808594 127.957031 243.398438
v 141.136719 127.957031 243.398438 ###
v 135.234375 127.957031 244.703125 <<<
v 135.234375 145.660156 243.265625
...

How with cars? Models have many visual parts.

I only have that wheel file mrstyx extracted. (And before the wheel isn't correctly converted to obj it doesn't make sense to me to care for cars.)

shakotay2 wrote: ↑Sat Apr 25, 2020 7:32 pm I only have that wheel file mrstyx extracted. (And before the wheel isn't correctly converted to obj it doesn't make sense to me to care for cars.)

Did you still working of converter for No Limits? Textures we can get using Ninja Ripper+emulator for Windows but models with correct UV isn't possible.

No. I don't have the time for fiddling any more - just working on models when there's sufficient infos.

Here "I only have that wheel file mrstyx extracted" where he said it's modified then vanished (for ever?).

Bad base to keep my interest, sorry.

Hello,

There is really not possble to extract 3d models from .sb3d files, even they are uncompressed?

I mean, thanks to "DKDave" on Discord (thank you!), that made a script to uncompress the .sb3d files, once extracted it, there are like ~700 files from 1KB to 277 KB heavy, where I think there are stored the 3d models.

BUT I tried to open this file in 3d Model Researcher, but being me so ignorant in read hex code, I can't find the start address for vertices

SO there is once again a way to open NFS: No Limits 3d models?

P.s. here a folder with the chunk files: https://www.dropbox.com/s/4lmr5haze130d ... s.rar?dl=0

The files extracted are bunch of vertex and face buffer (except for the largest one who contains the model information), for example, the file 359 is a vertex buffer that goes with the file 27, which is a face buffer.



(But you can see it need some scaling, which is exactly the same problem for the wheel that was posted)

yeah, I see, giving it a 2nd try seems it's a matter of "relative to absolute face indices conversion" for the wheel I previously dealt with
(or of collecting sub meshes one by one):
. (lower right wheel: I did a manual collection of sub meshes 'till the point where it goes bogus...)

And here for the BMW M3 from chunks (combining _e46_382 and _50):
. Scales (wild guess): y x 2.5, z x 0.8

Just a bit more info ... the very last block (688 in this case) contains the data on how to put all of the meshes together. If I can work it out, it should be scriptable (is that a word?).

@shakotay - This file contains the x/z/y scale at the start as 0.96, 0.80, 2.25 (rounding down for simplicity), so your wild guess is about right

DKDave wrote: ↑Fri May 14, 2021 7:59 pm Just a bit more info ... the very last block (688 in this case) contains the data on how to put all of the meshes together. If I can work it out, it should be scriptable (is that a word?).

@shakotay - This file contains the x/z/y scale at the start as 0.96, 0.80, 2.25 (rounding down for simplicity), so your wild guess is about right

Where did you find the scale ? Because It seems that the bounding box is tied to the scale and the position of the mesh, I think ? (I'm a bit a noob at this stuff, so I can be completely wrong)




(Testing the one at 0x1EE70)

The scale looks to be the 3 floats at offset 0x20 - they seem to match with Shakotay's info, but that could just be a coincidence!

So I wrote a (unfinished) blender script. However, the problem is, it's my first steps in blender scripting so you will need to install the zstandard module with this tutorial (since I don't know how to do a blender addon with a module included), and then go the scripting tab and run the script here.
(The script I made is for the version 2.8, so I don't know If it's going to work for the later versions.)



While it look fine here, I kinda have doubt on how the meshes are placed (Not seen here, but some car model have the chassis just under them). And the way I placed the mesh according to the bounding box is kind of dumb. While the normals are read, some mesh still doesn't look smooth, and some UV are broken (or need scaling ?)

(it's funny that this topic is trending since I was working on this format last week, what a coincidence)

Chipicao's tool worked greatly. It just doesn't support the latest models because of the newer encryption, but I'm sure that the 3D format is the same. Wouldn't it be better to just continue his work instead of doing the research for this completely from scratch?

AMG wrote: ↑Sat May 15, 2021 5:54 pm Chipicao's tool worked greatly. It just doesn't support the latest models because of the newer encryption, but I'm sure that the 3D format is the same. Wouldn't it be better to just continue his work instead of doing the research for this completely from scratch?

the SB3D format is really different than the M3G format from what I've seen, they did more than just add a compression in the new format. If you're talking about the vertex attributes, then, yeah, it would be lucky if it didn't change.

Oddly enough, the textures (.sba) doesn't seem to have changed, but for some reasons, the program (sbabrute) doesn't work.

By the way, It is possible to program a bounding box function from Advanced Mesh Reaper in blender ? It seems that the matrices at the start are important for the placement of the body kit.