01-16-2016, 02:48 AM
The bits that bleed on the alpha mask appear to only coincide with with black (which I assume are transparent) areas on the lower image, and vice versa. When extracted using the mask it should all turn out fine.
Can you tell me where you found those files though? I haven't found them yet, although I found some .tga portraits. The format is weird, Windows Photo Viewer can't open them but IrfanView and GIMP can. Maybe that's normal though, I don't work with .tga often.
![[Image: xyMF9gB.png]](http://i.imgur.com/xyMF9gB.png)
I made a GFXtract script to turn them into .png, but I noticed that they have a very slight transparency. Similar to some 3DS files I've seen, each channel only has the upper 4 bits set, the lower 4 are all 0. So the alpha is always F0 instead of FF, but that's the same with R, G, and B as well. I'm pretty sure this isn't intentional, so that means that the palettes are slightly off.
GFXtract can't do this yet but I might add a thing to allow these sort of modifications.
Can you tell me where you found those files though? I haven't found them yet, although I found some .tga portraits. The format is weird, Windows Photo Viewer can't open them but IrfanView and GIMP can. Maybe that's normal though, I don't work with .tga often.
I made a GFXtract script to turn them into .png, but I noticed that they have a very slight transparency. Similar to some 3DS files I've seen, each channel only has the upper 4 bits set, the lower 4 are all 0. So the alpha is always F0 instead of FF, but that's the same with R, G, and B as well. I'm pretty sure this isn't intentional, so that means that the palettes are slightly off.
GFXtract can't do this yet but I might add a thing to allow these sort of modifications.