first submesh vert count is 16 bytes before vert start address
face count is 8 bytes before face indices start address

I ended up with a face count of 6114
UVs at address 0xa26ab
and a vertex count of 1619

first submesh go backward 8 bytes from each start address to see the face/UV/vertex count

Sorry man i have no clue what kind of image this is, it kind of looks like an archive of some sort.
4 bytes - BODY
4 bytes - data length
data
next sub header


since this is a PC game have you tried ripping the textures with Ninjaripper or something?

Nothing, just a little time to understand it, Hex2obj is free for every one to use :D from that sample i posted before 0x0 - length of xml data 0x11a - num submeshes(?) 0x1B1 - num triangles (multiply by 3 to get face count) 0x1B9 - vertex count ive seen 3 different vertex strides in various samples...

wheels_7oqkocc3

looks like you aren't alone http://forum.xentax.com/viewtopic.php?f=18&t=13122&p=107983&hilit=Monster+Hunter#p107983 :D I've tried neosis script it doesn't work. do you mean the script here? http://forum.xentax.com/viewtopic.php?p=103731#p103731 edit: i see you already posted again here ...

In the words of aluigi

No sample, no party

i added support for your non working diffuse texture samples. the normal map texture samples look morton swizzled but i still could not get correct image scale so i just commented out that in the script and is there if you want to play around with it to get it working right. i have no clue what form...

while you can extract the bdae file down to 3 files you gain nothing because the data is already uncompressed but you should extract it anyway so the offsets in the model file make sense. this is the first submesh of little_endian_quantized.bdae from your car_delivery.bdae sample http://www.mediafir...

i got a point cloud from your SUB1.map sample under the VRTX section


i could not figure out the face indices though

0000000000000000.bre sample mesh header little-endian 0x18D0 - offset to vertex block start 0x18D8 - length of vertex block 0x1910 - vertex stride/FVF (first occurance) 0x48 = 72 dec 0x19D0 - vert count (add 1) 0x19D4 - face count 0x19E0 - offset to face indices start 0x19E8 - length of face indices...

None of it is compiled, all the scripts in the attachment are readable python source.

Sukone Tei wrote:where do you find the fmt_phantStarOnl2.py?

there is no script file by that name in this thread, the Noesis script is in the first post and it looks like you already have it.
some nonworking samples are needed if the script is to be updated by someone.

I tried to get something out of your samples with an experimental Noesis script i made using the unswizzle part of mikulover39's gxt script posted here http://forum.xentax.com/viewtopic.php?f=16&t=14419&p=119216&hilit=gxt#p119216 i gave your samples a gxt extension, the text2_n sample lo...