adding 282440 to the offset might be the better choice:

What should I use to extract it? you could try out the below bms script. But be warned! Use it on your own risk! It also requires the following preparations: Copy the first 0x44F44 bytes from FILE.CACHE into a new FILE.TAB. Then rename FILE.CACHE to FILE.ARC. Run quickbms, select the script then th...

Why wouldn't you just download the latest one from the official site of FMOD ? Because of this sentence of yours: Strange enough, but using their old fmodex.dll library, we can now extract all of the latest FSB archives, including these I was not sure whether the fsb I had would be handled. It woul...

easy ways are boring, aren't they? here you go (tested with Sound Archive game55.arc only!): # Just Cause 3, tested with Sound Archive game55.arc only! # script for QuickBMS http://quickbms.aluigi.org open FDDE "tab" 0 open FDDE "arc" 1 set EXT string "fsb" get ARC_SIZE...

but if there is any way to get a blender importer with bones and weights I would love that more. guess Mariusz will care for that sooner or later if the model format is not too different. There's .mdl and .mdx files in both games? Would be nice to have their structures compared. (I'll see whether I...

yeah, totally easy. You can even cut an uncompressed file from the .arc using a hexeditor and create a wav from it using daemon1's fmod_extr : 2107115430.zip HASH BYTEOFFSET SIZE DCOMP SIZE NAME ========== ========== ========== ========== ========== 0x9437BC85 0x00000000 0x0000DD10 0x0000DD10 sound/...

Thank you for your extractor - it's a bit choosy, though. I had an elder fmodex.dll version and got this error message: FMOD lib version279634 less than header version 279638 so I decided to patch your exe: fmodex_dll_4-44-52.jpg Replaced two bytes (56) by 52 each - worked fine with an FSB5 file. (I...

read here: viewtopic.php?f=10&t=13040

Looks like I have to hunt them down then, but how? how do I determine the count on Step 1? In the case of L9.PC, it's 19344. have a look at the picture in my post as of Nov. 26th, 12:38 pm to find the end address of face indices (FIs) block. With "startaddress of face indices" (0x0034) an...

huckleberrypie wrote:Should the address be constant or do I have to expect a different address for every file?

shakotay2 wrote:

huckleberrypie wrote:Now, should I expect the same offsets/values to work consistently on most if not all models?

how should that work? Each model/level having its individual vertex/face indices counts.

how do I determine the count on Step 1? In the case of L9.PC, it's 19344. have a look at the picture in my post as of Nov. 26th, 12:38 pm to find the end address of face indices (FIs) block. Calculate length of FIs' block by subtracting 0x34. 0x1F800 - 0x34 = 128972 dec. Each face consists of 3 FIs...

*.mdx: first submesh only - it's a little bit tedious to get the second one:

You guys - some bytes have changed and you are lost - aren't you? :) This is type 0x24. Cut the first 20 bytes from 11.data. Change Sigil's script (for type 0x0C) like such for type 0x24: get WIDTH long get HEIGHT long get IMGSIZE long get TYPE long goto 0x28 get FLAGS long goto 0x38 if TYPE = 0x24 ...

(thx, I have that L9.PC already from your previous upload.) I'm seriously asking myself WHY you don't use 0x0034 as the starting address for the face indices? (See, life could be very easy if you did what you're advised to do. :D ) L9_PC.jpg Also try to solve your saving problems. If an H2O file exi...