.mesh format similar to diablo immortal models

[jenny1367]

high-poly mesh that looks similar to diablo immortal III formats
in the header instead of

Code: Select all

P3F�Q4H_T2H_T2H

,

Code: Select all

P3F_N4B_T2F�T4H_B4H

is used.
File samples provided(diablo.mesh as a reference, but the queried mesh is foralltimes.mesh).

The game is foralltimes, messiah engine

https://www.mediafire.com/folder/ph8xr0 ... esh_sample

[roswell]

It is the same format, with one minor difference. Triangle indices are stored as DWORDs instead of WORDs.

[jenny1367]

It is the same format, with one minor difference. Triangle indices are stored as DWORDs instead of WORDs.

Is there possibly any script to parse it based on diablo tools or any way to edit it to make it work?

[shakotay2]

I have no idea.
If it weren't for the point clouds:
.

foralltimes-mesh.png

[jenny1367]

I have no idea.
If it weren't for the point clouds:
.

oo bc this is what I get with viewtopic.php?p=185106#p185106

ss.png

[shakotay2]

For me it doesn't make much sense to use such complex meshes as a starting point. Don't you have simpler ones?

[jenny1367]

For me it doesn't make much sense to use such complex meshes as a starting point. Don't you have simpler ones?

Thing is the game just started using 3d models for collectables and most of them are quite detailed, I’ll try find something smaller though

[roswell]

This script should already be able to do it:
https://github.com/TaylorMouse/MaxScrip ... rtModel.ms

See, what it's doing. It's checking to see if there's more than 2^15-1 indices, and if there are, it reads in INTs instead of SHORTs.
But, I think this is buggy, as nobody stores negative indices here. I think it should be checking for unsigned shorts, which are 2^16-1 (65535).
Unless someone can find a mesh file with vertices between 32767 and 65535 that clarifies this issue.

local nTotalVerts = readLong stream
local readshorts = true
if ( nTotalVerts > 32767 ) then readShorts = false

for f=1 to nTotalIndices / 3 do (
if readShorts then
append faces (_help.ReadTris stream)
else
append faces (_help.ReadIntTris stream)
)


foralltimes.mesh has vertices over 100k, so the script would read in INTs instead of SHORTs.

[jenny1367]

For me it doesn't make much sense to use such complex meshes as a starting point. Don't you have simpler ones?

maybe try this><

[jenny1367]

This script should already be able to do it:
https://github.com/TaylorMouse/MaxScrip ... rtModel.ms

See, what it's doing. It's checking to see if there's more than 2^15-1 indices, and if there are, it reads in INTs instead of SHORTs.
But, I think this is buggy, as nobody stores negative indices here. I think it should be checking for unsigned shorts, which are 2^16-1 (65535).
Unless someone can find a mesh file with vertices between 32767 and 65535 that clarifies this issue.

Haven't worked with 3ds max in a while but will take a look when I install it back! Thank you so much><

[shakotay2]

Thanks for that smaller sample, looks very "organic".
In the midst seems to be the uvs:
.

100k_mesh.png

[shakotay2]

100k-mesh-uvs.png

H2O file:

0x76 12468
Vb1
24 16
0x61DE 2603
020000
0x0 255

[shakotay2]

Complex one is also simple:
.

foralltimes.png

H2O file:

0x76 455160
Vb1
24 16
0x1BC856 107797
040000
0x0 255

[Durik256]

oo bc this is what I get with viewtopic.php?p=185106#p185106

try my new MeshToOBJ.exe, (click up arrow)

MeshToOBJ.exe

foralltimes2.mesh.png

*(the Noesis plugin also works, but it will be slow for your files. (you need to change the loading method, parsing attributes, etc, but I'm too lazy))

[jenny1367]

try my new MeshToOBJ.exe, (click up arrow)
*(the Noesis plugin also works, but it will be slow for your files. (you need to change the loading method, parsing attributes, etc, but I'm too lazy))

It worked wonderfully, thank you so much!
Also, not a big problem but the uv seems to be flipped? Anyways thanks again><
UV maps:
https://imgur.com/a/YBYpfWo