(11-15-2019, 02:24 PM)Bighead Wrote: Sorry I had to take a break from this due to lack of time (and really frustration). Many of the changes I made will most likely be scrapped as the idea is apparently not working as intended. My goal was to try to get images to look identical to the original, but I apparently have a lack of understanding of how TGA and TIFF files work. On top of that, while all the programs this script supports can make both, they all have very different results so I have no idea what is "right".Don't worry, you already did tremendous work! At the least now I have beta 3 to use compressonator with tiffs and texconv with tgas! Before it was impossible to automate it.
TexConv/Compressonator can create corrupted images when working with certain types of TGA. I also get corruption when working with some TIFF images. ImageMagick appears to get all of them right, but only when converting to PNG and it is cropping alpha channels? I have no idea why TGA has so many orientations and the impact it has if it is changed. I have very little understanding of TIFF images in general. It also doesn't help that I have zero test cases (are these two image formats used in games?), so I have no idea if the images created are as intended.
Since I've never had to work with these image formats before, and since they've been around for awhile, I blindly assumed the programs would just know what to do when going between formats. I guess there is a lot I just don't understand. This is the first I've seen "interleaved rgbrgb" so I don't don't know what that is, the point of it, or what advantages it has. I have no idea what "opaque" alpha is, do you mean a value of 255 for all pixels in the alpha channel? In that case, what would be the point of an alpha channel? I'm not sure what you mean by "TIFF behaves like PNG (does not store alpha channel)" as PNG can have an alpha channel with a separate alpha value per pixel.
I'm not a graphics artist outside of hobby (which I basically lost interest in), nor am I a programmer, so I don't have a deeper understanding of the image formats and terminology used. This script is first and foremost just a frontend for the programs it supports (with a few of my own twists to manipulate images). The only format I know well is DDS, and that is only because I spent months learning and experimenting with it in my free time. And that actually has great program support, but still had many issues I had to work around. I say all this because in the end, getting this to work may be completely over my head. At least for now.
I'm too just an enthusiast and learning about all of this. I think "opaque" alpha is used to hide part of the texture without destroying it (in the example 1 first texture with "opaque" alpha has moving door in the hidden part) and transparent is used for translucency, reflections and such(and it needs separate alpha channel to achieve this maybe?). I "mean by behaves like png" that in your script tiff converts alpha channel to png transparency(example 2) thus cuts parts of the image. And in conversion by hand in PS or xnconvert alpha channel remains separated.
Quote:>nor am I a programmerIf that's the case...WOW.