succha wrote: ↑

Wed Apr 29, 2020 11:14 am

Hi
so how can I extract GT6 tracks?
use gttool? any command?

Unless you know how to work with binary files, pretty much no, even doing a script its kinda hard cause the meshes have too many changes, it requires manual labour one way or another

I basically did the whole script as a way of helping SRP, and yeah shinjuku might be on the works.

Also, did you do it through ninja ripping???? XD

Interesting to see such post... First of all, are the binaries encrypted? That it uses same format sufix doesnt always mean that the structure is the same, can you post an example file? My aproach would be to try and work with the pc version(even if it would be compressed or encrypted you can always...

That looks compressed, if you wanna test out for inages use texturefinder, with rawtext you can convert them to dds if that is the case, but again, that looks compressed, try maybe offzip

Well, i have been already been able to.... precisely wanted to post an update, been working on akasaka and also on valencia, first of all, though i have been able to pull it out, its not easy, specially for tracks, the geometry is in various files, sky, road and simple stuff are with such names, tha...

so funnywise 3d max complaining as always while blender doing a wonderfull job at importing.... more previews....

this is working!

a little tweak on the fi generating triangles.... and voila, normals working more or less properly, without the triangle strips and shit

Whew, kinda had me thinking this whole thing for long, i was pretty close, by doing an unpacking of the last byte and getting only the last bit i got where the jumps have to happen to solve the problems, now just trying to figure out what i am gonna do about the tripple jump readings, i know that i ...

so.... been trying some things.... by using the last byte of those ints getting interesting results.... now i was able to split the trees properly, even the mesh i showed before, used the first FF as a marker, then ignored a second one, as if those bytes would mix together to form 0xFFFF, kinda funn...

oh so basically either taking it signed or unsigned and dividing by the max value, interesting, will give it a try though here an interesting insight.... https://i.imgur.com/VfTh0EW.png this is the normals breakdown on advanced research model note that 5 and 6 should have 0 for normals... but then h...

okay, sorry to bother once again, but how do i turn those bytes into decimal values? just been trying to find more information about how can i code that myself but.... i know about how to do it with halfs, bytes/chars dont even know where to start, how do you do it in your program?

So thinking about it this might be actually an interesting aproach, sorry, i didnt quite read the offsets at the first image you sent, interestingly normals could actually work as helpers for proper backface culling.... I will do some testing, thanks for a new insight! :wink: Regarding possition and...

import mrp f = mrp.get_bfile() verts = [] uvs = [] faces = [] curfi = -1 mat = open('mat.txt', 'w') variables = [] print ("***************") meshnum = f.readInt() jump = (meshnum*16-4) f.seek(jump,1) print ("start "+f.tellHex()) for i in range(meshnum): verts = [] uvs = [] faces = [] curfi = -1 cur...

Nice! thanks for an early repply guys! the only problem for me here is that rn im kinda trying to find a fix for the fis, for instance in my screen or the one that shakotay just posted, more or less its how it shall look, but there are lots of wrong triangulation that binds parts of the meshes that ...

So, been lately trying to understand the FI format of the models from TXR Drift 2..... but at the same time the model files dont seem to make much sense, all i know is that there are vertex formats and then there are uvs, then there is a int, maybe shorts? that dont make much sense to me.... if anyo...