The VG Resource

In this thread I'll be posting my model rips.
Here are first three. If everything is going to be well, there will be some more soon. Also I'm not used to .dae format so I included SMD files along with DAE for rigged models in case I messed up something.
S.T.A.L.K.E.R. Call of Pripyat (PC)
Mutants are rigged

Metro 2033 (PC)
No rigging

Sword Art Online: Infinity Moment (PSP)
Characters and creatures are rigged

Fairy Tail: Zelef Kakusei (PSP)
Characters and creatures are rigged

Fairy Tail: Portable Guild (PSP)Characters are rigged

Sword of the Berserk: Guts' Rage (Dreamcast)
Characters are rigged

History's Strongest Disciple Kenichi (PS2)
No rigging

World of Tanks (PC)
No rigging

Fate/Unlimited Codes (PSP)
No rigging

Fate/Extra CCC (PSP)
Characters are rigged

They are good, however we don't except .tga textures. So could you convert them to .png?

Sorry, didn't know about that. Updated the links in the first post, all the textures should be in png now.

OK, I'll upload these when I can.

Here are some more models (all rigged):
Sword Art Online: Infinity Moment (PSP)

Previews:
I didn't actually thought that I'll manage to rip any models from this game, but it turned out that all models were in GMO format. I think I'll rip some more models of characters/enemies/weapons and then write a tutorial on how to do this. Still have to figure out a couple of things about this game's file formats.

OK, these are all up.

So, I am finally back. Here is a couple of models from SAO:

There are a few other models pending and I'm going to put them here as soon as they get (most probably) accepted.
Also I'm probably going to start a similat thread, but in texture section. The first upload - God of War 3 Kratos textures is still pending too. As for the models from this game, it's like this:

I can get vertexes, I can get UV, I can get textures, but I can't get vertexes to conect into triangles. It's frustrating to no extent, but unfortunately I can't do much about it, at least not by myself. And my script is pretty useless as it can't get all the data automatically, you need to manually set offsets for it for every mesh of any model. I based it off script, written by somebody at ps23dformat. He did a great job trying to reverse the format, but omg, why would anyone write a model conversion script in quickBMS scripting language? As a result it is insanely hard to understand the script and while it's capable of converting any model in a bunch of .3ds files, the model looks like this, has no UV and some vertexes are missing.

I know how do u feel with those gow3 models. I spent some time to work to a method to generate an aceptable mesh by vertexes without great results. Anyway i can tell you that furryfan's script converts all the vertex points. Fortunately I didn't find any missing vertex in all the models i looked at. Just for courosity, Is the red pic UW map? how did u find it?

(01-30-2014, 07:07 AM)Orwell Wrote: [ -> ]I know how do u feel with those gow3 models. I spent some time to work to a method to generate an aceptable mesh by vertexes without great results. Anyway i can tell you that furryfan's script converts all the vertex points. Fortunately I didn't find any missing vertex in all the models i looked at. Just for courosity, Is the red pic UW map? how did u find it?

Yes, it is. When I got my hands on the script I was trying hard to understand how it works, but to no avail. And then I found out that quickBMS has a html output, where the data, read by the script is highlighted. I still couldn't grasp the way it works, but i managed to write a simple Noesis script, that imports vertex data for one specific model (that harpy). I had some troubles with getting all the vertexes imported, and by the way that led me to belive that the script doesn't convert all of them too, as I copied the values it used to calculate the number of vertexes.
While I was trying to solve the issue I stumbled upon a data block, that starts right where the vertex block ends. It contained floats (fractions), all of which had values less than 1. I thought it may be UV, did some math, and it turned out that the number of floats is twice the number of vertexes, so I assumed it to be U and V coordinates for each vertex. I was actually right, as you can see from the screenshot. Although it's quite surprising that it wasn't found before, as it was pretty obvious even for me, and I'm just a beginer at reverse engineering .

Edit: Sometimes not all floats are less than 1, but it still seems to work. I did it about half a year ago so I don't remember all that stuff in detail.

(01-30-2014, 07:54 AM)maver1k_XVII Wrote: [ -> ]Yes, it is. When I got my hands on the script I was trying hard to understand how it works, but to no avail. And then I found out that quickBMS has a html output, where the data, read by the script is highlighted. I still couldn't grasp the way it works, but i managed to write a simple Noesis script, that imports vertex data for one specific model (that harpy). I had some troubles with getting all the vertexes imported, and by the way that led me to belive that the script doesn't convert all of them too, as I copied the values it used to calculate the number of vertexes.
While I was trying to solve the issue I stumbled upon a data block, that starts right where the vertex block ends. It contained floats (fractions), all of which had values less than 1. I thought it may be UV, did some math, and it turned out that the number of floats is twice the number of vertexes, so I assumed it to be U and V coordinates for each vertex. I was actually right, as you can see from the screenshot. Although it's quite surprising that it wasn't found before, as it was pretty obvious even for me, and I'm just a beginer at reverse engineering .

Edit: Sometimes not all floats are less than 1, but it still seems to work. I did it about half a year ago so I don't remember all that stuff in detail.

Thank you so much for the info! (even if to be honest i don't like so much write script lines, and i have too much stuffs to learn) Expecially in this case whit this weird format i preferred to work on the bms 3ds output after i made a surrender to this file format.
So then i tried to create a decent mesh from a vertex point cloud using various formula to aproximate the mesh, but with not great results.
I finally opted for the manual recreation of the mesh, I know it's not the clever method to obtain the model, but i like modelling more then reversing a 3d format XD. Actually I'm working on Aphrodite model and the main iusse is the uv map. I may recreate the seams manually but then need to be perfectly adapted to the texture...and this is a really tedious thing. Do you think u will able to extract the uv map "vertexes" in an automate way? Theoretically with a max script it possible to associate a uv map vertex points to a mesh, anyway having even an idea of the uv map of a 3d model of kratos or Aphrodite it may be quite useful for me.

This is a wip of Aphrodite model ( i hope to not offend anybody with this nude model, but i i didn't to work on clothes yet. ):

Unfortunately I can't. The problem is that I don't know how to read the UV data automatically for every mesh and character models consist of about 100 meshes, so finding the offsets and calculating the vertex number manually isn't really an option. And I don't have any idea whether I'll be able to do anything about that in the future or not.
By the way the model looks good.

Also here is a "little" update (omg that was a lot of copy-pasting):

And here are some more models:
Sword Art Online: Infinity Moment (PSP)

History's Strongest Disciple Kenichi (PS2)

There are more models (mostly from Fairy Tail), but they are all still pending. Just wanted to put those here a little earlier.

Update: here are the rest

Wow thanks for the Fairy Tail and Sword Art models! Are you using JPCSP to rip them?

(02-10-2014, 03:39 AM)Zerox Wrote: [ -> ]Wow thanks for the Fairy Tail and Sword Art models! Are you using JPCSP to rip them?

You're welcome.
And no, both games use common .gmo model format so I used Noesis. Although the models are somewhat hard to reach. I actually had to use JPCSP to decrypt the main archive of SAO:IM, which turned out to be a .cpk and those can be extracted with Quickbms script. The models were kept in some sort of containers, so I used Gitmo on them to get the models and textures.
As for Fairy Tail, I found a way to unpack the archives somewhere on Xentax. The way is imperfect though because it doesn't keep the file names and directories. Characters are kept as a number of separate models (hands, body, legs, head) and while most of the main characters have at least one model, that's already assembled, finding all the necessary pieces for some characters can be troublesome.

Oh that's a lot easier then, I have tried a script I found on xentax to unpack One Piece Romance Dawn also a PSP game. But even though the models appear to be gmo they don't open in Noesis.

If you're interested in those models maybe your method can work better.