20.3's from this game.
There's no guarantee that two models from the same version of the game engine have the same format.

The ones you uploaded are from 20.2.0.8g_as, whichever version that is.
Which would be useful also, but I want to see if my script can load the 20.6's

Maybe bugmenotty would like to share his source since it works?

I'm just collecting a list of issues. It may be useful for others as well. Maybe I'll make a wiki article for like "Noesis python FAQ" or something later since a lot of the stuff would probably come up again. I've also resolved the big endian and binding issues after noticing the "RPG...

Client DL?

These formats I had already written in my Sanae3D tool so it was easy to convert to noesis.

I think perfect world uses the same format.

I've updated the script to load all meshes. It also parses the bone section, which is just a bunch of bone names. Consider the size of each vertex, there might be indices into these bones. Or not, considering how many 0's there are. It looks like the material names are defined with the meshes as wel...

Looks like aluigi posted an unpacker that gets the filenames. Is this the output from that one? http://forum.xentax.com/viewtopic.php?p=28471#p28471 There's a lot of stuff stored in this format, like bone names, and some other stuff after the mesh section. Header has 4 integers that will be of inter...

Upload a bunch of those 20.6's
And a couple of 20.3's

http://db.tt/8WjMD9Oq

Static meshes can be loaded (type 5)
Still trying to figure out what's wrong with the way I'm parsing the type 6 meshes lol...

http://i.imgur.com/xA3FF.jpg Some things 1: all (or at least, 99%) of the filenames are in chinese (gb2132). Don't know if it will be a problem for some people. 2: some of the referenced textures do not even exist. Some models refer to psd files, which I guess these are just test models that they d...

The contents of this post was deleted because of possible forum rules violation.

Do you think you should?

I typically start by looking for the vertices. If I can verify that, then I can look around and quickly calculate values like vert count, find where they are, and then combine that with a general skim through the format, have an idea of the structure of the format. But so far I couldn't even find th...

I wouldn't be surprised if it were.
But now I'm just looking for the vertices

Wish I had a tool that calculated half-floats on the fly as well.

2. it asks me for a script... what script?

The script that you want to use.

Haven't you been reading what aluigi was telling you?