Page 3 of 3
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 7:10 pm
by moonpaladin
Shakotayyyyyyy! I was trying to find a way to separate the meshes and I found a file that will help alot with that step its called .xdb file it content more data about the geometry and textures!I'm attaching a file, please if you got it working, tell me some steps to make it working too! I'm trying to get some mounts that I really love u.u
https://www.mediafire.com/file/2ykan3wd ... no.7z/file
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 8:38 pm
by shakotay2
moonpaladin wrote: ↑Sun Jan 31, 2021 7:10 pm
Shakotayyyyyyy! I was trying to find a way to separate the meshes and I found a file that will help alot with that step its called .xdb file
Well, more confusion than help for me:
Code: Select all
<modelElements>
<Item>
<lods>
<Item>
<indexBufferBegin>0</indexBufferBegin>
<indexBufferEnd>8040</indexBufferEnd>
<vertexBufferBegin>0</vertexBufferBegin>
<vertexBufferEnd>2072</vertexBufferEnd>
</Item>
<Item>
<indexBufferBegin>68757</indexBufferBegin>
<indexBufferEnd>74145</indexBufferEnd>
<vertexBufferBegin>0</vertexBufferBegin>
<vertexBufferEnd>2072</vertexBufferEnd>
</Item>
<Item>
<indexBufferBegin>74145</indexBufferBegin>
<indexBufferEnd>77499</indexBufferEnd>
<vertexBufferBegin>0</vertexBufferBegin>
<vertexBufferEnd>2074</vertexBufferEnd>
</Item>
</lods>
, tell me some steps to make it working too!
file offset 0x0004 (4 bytes before start of vertices at 0x0008)
A8 23 0A 00 -> 0x0A23A8: 664488, size of vertex block
FVFsize 36 (for other models: if it doesn't fit try 28 or 32)
vertex count: 18458 (664488/36)
Enter the params (see pic below), toggle button 'noPtC' to 'PtCld',
press 'mesh' button
At the bottom of the left hand list box find next address: 0xa23b0
edit:
now update the params as follows, then toggle button 'PtCld' back to 'noPtC'
Find "size of face indices block" (DWORD) at 0xa23b4: 36 1D 04 00
0x41D36, decimal: 269622
face index count: 269622/2=
134811
start of face indices: 0x
a23b8 (enter address without the '0x')
.
MountDino.png
There's 4 submeshes, afaics.
(You need to be a coder to separate them into smaller pieces.)
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 9:58 pm
by moonpaladin
Thanks for the fast response Shakotay! I just try with the the parameter that you used loading the decompress bin model, and got this :C
dino_error.png
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 11:18 pm
by moonpaladin
I'm trying with other mount and got the same model, only vertices xD.
mount_ho.png
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 11:27 pm
by shakotay2
ehm, yeah, read my post completely!
face index count: 269622/2= 134811
start of face indices: 0xa23b8
Has to be applied, of course (thought this being obvious
.gif "Smile")
)
Then switch to 'noPtC'.
edit: well, this was the misleading part in my previous post, sorry
"Enter the params (see pic below), toggle button 'noPtC' to 'PtCld',"
Re: Allods online updated final request
Posted: Sun Jan 31, 2021 11:50 pm
by moonpaladin
hahaha yeah! was my mistake ! thanks Shakotay!!
Re: Allods online updated final request
Posted: Mon Feb 01, 2021 8:21 am
by shakotay2
You can use a whatever face index count but it must be < calculated face index count (see my post somewhere above).
But you can't set the vertex count independently! Because there's a relationship: the maximum face index (FI) equals the vertex count.
Max FI is calculated and displayed in the lower left listbox when pressing the go1 button.
(See the tutorial (tut button))
(The maximum face index is not the same as the face index count, btw.)
Re: Allods online updated final request
Posted: Mon Feb 01, 2021 3:40 pm
by moonpaladin
Yeah, I'm still practicing but got the mount!!!!!love you hahaha
Re: Allods online updated final request
Posted: Tue Feb 02, 2021 8:51 am
by moonpaladin
Sorry for so many posts :c
Re: Allods online updated final request
Posted: Tue Feb 02, 2021 9:51 am
by shakotay2
well, sometimes a further step is required: press the
go1 button when you have entered your params.
at the bottom of the left lower listbox you'll find the
max face index: 37773 (
MountBeetle). Use it as the vertex count for step 3
Is different from the calculated (overall) vertex count here.
--------------------
For the 2nd mesh of the
Hoverboard we didn't get the face indices so far:
(maybe this time the .xdb file could come in handy?)
.
Hoverboard_sails.png
Re: Allods online updated final request
Posted: Fri Nov 04, 2022 2:36 pm
by Cyn
Raising this thread from the dead, sorry
Working on exports using the BMS script and Noesis plug-in provided in earlier posts. Personally, i work a lot with wow .m2 format, so these files arent too out of my realm of understanding. I can see how the LOD detection works in each file and ive noticed that changing small bytes and chunks of data to match the expected search of "00 00 01 00 02 00" resulting in the destruction of error #48 and a nice render produced. Im getting to a spot where i can reliably extract whatever model it is that im looking for.
Apparently we used to have a texture viewer for Allods but i dont see it working anymore. Anyone have any info on how we can get these textures decompressed? Ive been able to "fix" or extract the following models:
MountCar
MountCarCabrio
MountChopper
MountChopperAtoll
MountChopperGreen
MountChopperDead
MountCopter
MountChariot
Each separate piece of the v14 exoskeleton
If more work will be done on the file format, id be happy to explore building a 010 template for the .bin (geometry) files
This is my first post here, thank you for being a community before i arrived here
EDIT: Using texture finder i can see some of the texture in DXTC3 format, got it extracted using 256 width and 8 byte offset produced the most detailed image in TF 2.0
Notes:
-The first 8 bytes of the file are a pointer to the offset of the first LOD, At that offset (by 8 bytes) is the pointer for whatever section is next (presumably the second LOD)
-At the location of this pointer seems to be an 4 byte identifier for the mesh (usually just 1) followed by a 4 byte bit-by-bit length of the indices section (usually in raw byte form) these two pieces of information should give us reliable meshes produce via the noesis script