Hi, I don't draw lines. hex2obj just sends the test_uv.obj to the mesh_viewer.exe which displays the uv map. uv data in test_uv.obj as faked vertices: v tu1 tv1 0 v tu2 tv2 0 v tu3 tv3 0 f 1 2 3 The important thing is the face indices which you don't take care of when just connecting one uv point to...

found a small mesh but faces are unsure, tried toggling backface culling in Noesis, but the result is not really good: but this may serve as a starting point.

There's some convex hull data sections (maybe collision mesh?)

Jamal al dambali wrote:And how did you find the point clouds?

it's a combination of trial and error and experience

And one more question, does this file have a submesh?

yes; I usually search for one submesh only and leave the rest of the work to the customers

(upps, wrong thread, sry)

oldman wrote:I'm looking for the GTTool

I searched this forum and found only

Hello,
viewtopic.php?f=16&t=11783&p=97169&hilit=gttool#p97169
The exe is in the bin\Release folder.
(If it doesn't run on your system you're required to compile the source.)

dunno which face indices you used in the end but seems you're on a good way: (as a deja vue only: https://forum.xentax.com/viewtopic.php? ... ds#p103400)

edit: upps, well, see I'm growing too old, you've posted that picture already...

I don't have the time for making video tuts, sry. Try to repack un changed files. It should result in an archive that's identical to the original. (for checking use fc /b newlyPacked.ipk original.ipk in windows console or any hexeditor which allows file comparison) If there's a difference the unpack...

You didn't tell where the "proper co-ordinates" of the Unique objects to be found. Anyways, since you seem to know how to place them, after a glance at t_ginza.hood : there's t_ginza_rd_48_x , unique_component 0, it's the owner of instance_component 1253 for example (and 4 other instances) I assume,...

you could easily find out this for yourself.
look for point clouds, then look for face indices, it's DWORD here:

Guess he just changed the parameters, something like this:
viewtopic.php?f=33&t=15854&p=127514&hil ... ip#p127514

If the size of the textures will remain unchanged you can try using the reimport feature of quickbms.
It uses the unpacking script for repacking.
Read the
3) Reimporting the extracted files
section in quickbms.txt.

hi eri619

For the Noesis script here: there's no simple solution. You'll need to dig into the format to find the submesh borders.
You could manually add grouping lines to the obj file such like this one:
g SM_x
Maybe some sorting of face indices could improve the result.

I'd guess for quads but autocreation of face indices might be the wrong idea: In fact there's a good chance that they use byte indices (0..255) since the model might be split up into 105 submeshes (but not sure).

not to sound sophisticated but for me it looks as if the script in the opening post only needed some swapping such like this: if DUMMY1 == 2 get DUMMY long get DUMMY long endif get NAMESZ long getdstring NAME NAMESZ get PATHSZ long getdstring PATH PATHSZ because the filename ("NAME") appears first i...

.ipk sample required