XeNTaX

Taner038 wrote:Up mates. I'm waiting for localization tools.

I have textools.
Xbox360, Ps3 and PC.

I have serious issues with extracting .mul files, can someone help me with this please?

gula wrote:

Taner038 wrote:Up mates. I'm waiting for localization tools.

I have textools.
Xbox360, Ps3 and PC.

You can share with me pc tools, mate?
And does it work?

Anyone please I cannot seem to find a way how to get mul files to fsb without the file being cut in half, please I need help D:

Вижу тут проскакивают рашен комменты. Расскажите народ как вообще достать модели + текстуры ларки ?
Хотяб пошаговую инструкцию. Это распакуй, это переименуй хоть что-то, а то в заголовке тупо 2 проги и скрипт который хз куда девать, с блендером я не дружу... работаю только в 3ds max.
UPD: про Noesis и скрипт понял. Остались вопросы как добыть скелет / развес и текстуры к модели.

With the current version of the importer (fmt_TR2_mesh_1_4.py) Noesis throws an exception if it encounters a triangle index with the value 65535 (0xffff): "RuntimeError: Index buffer was not valid for drawing."

I modified the script so that it skips those triangles. It is slower but is able to import larger meshes.
However, Noesis crashes when I try to load very large meshes (like 46MB).

The modified script: https://www.sendspace.com/file/60ocwn

moxxy wrote:With the current version of the importer (fmt_TR2_mesh_1_4.py) Noesis throws an exception if it encounters a triangle index with the value 65535 (0xffff): "RuntimeError: Index buffer was not valid for drawing."

I modified the script so that it skips those triangles. It is slower but is able to import larger meshes.
However, Noesis crashes when I try to load very large meshes (like 46MB).

The modified script: https://www.sendspace.com/file/60ocwn

Thanks, can you upload the mesh file which gave you this error as well?

Cheers.

TIGER Unpacker v0.0.3c r7 - Download

* Base with names updated. Added 6890 filenames.
* Fixed unpack files with too long path more then 260 bytes

DRM Dumper v0.0.3a r7 - Download

* Updated for 1.0.610.1 game version (bigfile.update2.000.000)

Ekey wrote:TIGER Unpacker v0.0.3c r7 - Download

* Base with names updated. Added 6890 filenames.
* Fixed unpack files with too long path more then 260 bytes

DRM Dumper v0.0.3a r7 - Download

* Updated for 1.0.610.1 game version (bigfile.update2.000.000)

Ekey, how can I repack ?

And, you're any working the localization tools?

Gh0stBlade wrote: Thanks, can you upload the mesh file which gave you this error as well?

Here you are: https://www.sendspace.com/file/x0vvq6

Section 70.tr2mesh: Can be imported using my patched script.
Section 68.tr2mesh: Crashes Noesis

I checked the amount of the memory that Noesis uses while importing those meshes and it gets very high
(more than 3400MB in case of Section 68). So the crash seems to be caused by low memory. Python is not my primary programming language
and my patch is just a dirty hack. Maybe you can make the script more efficient.

moxxy wrote:

Gh0stBlade wrote: Thanks, can you upload the mesh file which gave you this error as well?

Here you are: https://www.sendspace.com/file/x0vvq6

Section 70.tr2mesh: Can be imported using my patched script.
Section 68.tr2mesh: Crashes Noesis

I checked the amount of the memory that Noesis uses while importing those meshes and it gets very high
(more than 3400MB in case of Section 68). So the crash seems to be caused by low memory. Python is not my primary programming language
and my patch is just a dirty hack. Maybe you can make the script more efficient.

Thanks, just checked both of those which seem to crash Noesis. On the bright side, when I load them into my engine I don't get a crash at all. The file should load fine, it looks like there's a bug in Noesis so there's not much we can do.

That's not the case, Noesis allows 0xFFFF for 16-bit indices by default. If you're feeding 32-bit indices, you will need to set 0xFFFF as an explicit strip ender using rpgSetStripEnder. However, the error "Index buffer was not valid for drawing" is not caused by out of range indices (this is a different error and informs you of which vertex buffer would be accessed out of range by the index), it's caused by one of the following:

1) Number of indices < 3 for triange list, strip, fan, or polygon.
2) Number of indices not divisible by 3 for triangle list.
3) Same as above 2 but 4 instead for quad lists.

Sounds to me like you're not correctly recognizing triangle strips and are trying to draw them as lists.

Hi there!

MrAdults wrote:That's not the case, Noesis allows 0xFFFF for 16-bit indices by default. If you're feeding 32-bit indices, you will need to set 0xFFFF as an explicit strip ender using rpgSetStripEnder.

The face indices are unsigned 16 bits. Triangle strips are usually reset/ended by 0xFFFF index. But this doesn't seem to be the case, 0xFFFF only occurs at the end of the face index buffer which suggests it's most likely triangle list like all previous games.

MrAdults wrote: However, the error "Index buffer was not valid for drawing" is not caused by out of range indices (this is a different error and informs you of which vertex buffer would be accessed out of range by the index), it's caused by one of the following:

1) Number of indices < 3 for triange list, strip, fan, or polygon.
2) Number of indices not divisible by 3 for triangle list.
3) Same as above 2 but 4 instead for quad lists.

Sounds to me like you're not correctly recognizing triangle strips and are trying to draw them as lists.

After debugging this I found the mesh has 65536 vertices:
Section 70.tr2mesh
Vertex Start: 2028
Vertex Count: 65536
Vertex End: 1574892

If the highest face index is ‭65535.‬ Somehow it exceeds vertex[65536] which is what I don't understand.

####

This should work fine, I'm puzzled so there's not much that can be done.

If you have a legitimate list that happens to use index 0xFFFF, it's probably just hitting the default 16-bit path which rejects 0xFFFF, thus truncating the list by one and making it not divisible by 3. You could feed the indices as 32-bit (just expand them with one line of Python prior to commit), partition the commit calls, or get the next build I put out, whenever that is, and I can give you a hook to disable the ushort rejection.

Ekey wrote:TIGER Unpacker v0.0.3c r7 - Download

* Base with names updated. Added 6890 filenames.
* Fixed unpack files with too long path more then 260 bytes

DRM Dumper v0.0.3a r7 - Download

* Updated for 1.0.610.1 game version (bigfile.update2.000.000)

@Ekey Hi
The scene model appears to be a bit of a problem.
Could you fix it
THANKS!