XeNTaX

I'm not sure what "divisors" you're referring to

This is what I meant, although I may have misunderstood:

MrAdults wrote:I just looked back at a TRB, I was mistaken, you can pretty easily use those COMP chunks to correctly scale+bias without fucking around with bounds at all.

Yes, I'm using the COMP for the bias and the MESH AABBs for the scale.
Seems to work OK so far.......

Anyway, I felt like learning the Maya and Softimage APIs for some reason (no idea why).
Obviously I wanted to work with something I was halfway interested in, hence all this mayhem and foolishness.

I meant, it looks like COMP can be used to scale too. It works out just right cause they scale the box it provides with the mesh. I'll fix up the old trb loader to use it after work tonight.

Hmm, I hadn't tried that.
I figured it was working fine so I didn't bother investigating any further.
I'm fighting with Maya's skinning code right now. It was a lot simpler in Max.

It worked fine with the COMP chunks, the environments are all loaded correctly in Noesis now. Also, chrrox told me you had to assign all those materials manually, so I decided to try to get the Noesis material mapping to work a bit better. Barring actual shader emulation, the best thing would be to chart out the sampler mappings for all of the base shader types, and that will reliably give you your appropriate diffuse/normal/spec/hair offset/etc. maps. But I can't be assed to do that, so I just try to guess based on the texture and shader names. New default material settings in the Noesis preview:

It applies appropriate material settings for skin and hair now as well. Hopefully, this will keep you away from my balls for now.

Sweet Baby Jesus!!!
You the man Mr. Adults!
Maybe in another 10 or 20 years I'll know as much as you.........."maybe"

Now if I could just get the energy to finish mine......

EDIT: Here's a Noesis plugin for the PS3 xgr files:
View - http://www.mediafire.com/view/1czn3klwz ... XGR_XFV.py
Download - http://www.mediafire.com/download/1czn3 ... XGR_XFV.py


EDIT 2: I wasn't paying attention and gave those 2 lists global scope causing the textures to add up if you look
at multiple files without closing Noesis. Just move them under XGRLoadTextures to fix that.

EDIT 3: Fixed stuff n junk, updated link. Type 4 format is morton-order tiled a8r8g8b8

EDIT 4: Added support for xfv files

Nice work, I was able to look at some of the battle GUI elements and its quite fascinating how they have done them. There seem to be a lot of .xfv files accompanied by .imgb files in the vfx folder for example, have you had a chance to look into those yet?

flarespire wrote:Nice work, I was able to look at some of the battle GUI elements and its quite fascinating how they have done them. There seem to be a lot of .xfv files accompanied by .imgb files in the vfx folder for example, have you had a chance to look into those yet?

OK, I updated the link above.
I only tested out a few files though so I'm not sure it works for everything.

TheDude wrote:

flarespire wrote:Nice work, I was able to look at some of the battle GUI elements and its quite fascinating how they have done them. There seem to be a lot of .xfv files accompanied by .imgb files in the vfx folder for example, have you had a chance to look into those yet?

OK, I updated the link above.
I only tested out a few files though so I'm not sure it works for everything.

Im happy to report that most of the XFV's work right off the bat, but there are a few, especially any in the common folder thats inside the VFX folder, where they do not show any data in the preview.

So what format next? XD (jk)

Do those imgb files contain any data?

TheDude wrote:Do those imgb files contain any data?

some do and some are 0kb in size with large xfv files.

flarespire wrote:

TheDude wrote:Do those imgb files contain any data?

some do and some are 0kb in size with large xfv files.

All the imgb files I see under the "common" directory are 0 byte files, i.e., no textures.
The xfv files look like they're for visual effects and not all of them necessarily use textures.

well then I guess we have gotten all we can from the .xfv files...

On a side note, does anyone know why noesis has so much trouble properly reading the UV data for .ps3.trb's/.ps3.imgb's but has no issue with .x360.trb's/.x360.imgb's?

hi, newbie here (both 3D models and 3dsmax)..

i just following guides/hint from this forums and finally i can import Lightning model to 3dsmax (export to .dae from Noesis and import using OpenCOLLADA)

i read from this forum how to fix alpha/tranparency for the models by applying each map to their "Opacity" and most of them fixed beautifully. however i can't seems to fix Lightning's eyelash transparency, and it's driving me crazy



anyone can help?
Thank You

TheDude: You might wanna update your script to use the new Noesis rapi.imageFromMortonOrder function, it's faster and will handle badly-sized images too. Usage is rgbaPix = rapi.imageFromMortonOrder(rgbaPix, width, height). You can specify an optional argument too after height, if you want to force a byte stride, it will otherwise determine it based on the size of the data and the image dimensions. You can shift the width/height and force a byte stride to do stuff like untile DXT blocks, which isn't necessary for PS3 but is done on some other platforms.

flarespire: There are no PS3-specific UV issues that I'm aware of. In the case of FF13-2 the UV's are stored in the image bundle, which means if you have the wrong imgb your UV's will be wrong and probably complete garbage.

that's the issue though, i used the file list file to extract the white_imgu.ps3.bin file, and it came out exactly the same as the xbox360's would as well as the imgb's having valid image data that's identical to the xbox 360's its just that the UV mapping for the ps3 files (in both XIII and XIII-2) is completely F.U.B.A.R. compared to the xbox 360's files.

on the subject of file lists, has anyone cracked XIII-2's yet?