Jump to content

ndix UR

Member Since 12 May 2016
Offline Last Active Today, 09:35 PM

Posts I've Made

In Topic: Salk's "SW:KotOR Upgrade 2.0"

18 March 2018 - 11:25 PM

Feel free to use my K1 Ebon Hawk geometry fixes.


That particular mod is free for anyone to use in any way they see fit, credited or not, because I really never want to see those geometry errors in any sort of new work anyone ever does.


Best of luck in your endeavor.

In Topic: Assistance in getting an old mod to work

08 March 2018 - 12:21 AM

It looks like that mod is intended to be distributed in a .mod file called nih.mod ... so I would assume that the files in that package need to be packed into a .mod (easy enough to do with KTool, by the way) and then maybe it would work?


Actually there are also two .nss files without compiled equivalents ... so you might need to compile those scripts also? I haven't done much with scripts so I'm not sure on that part.

In Topic: How I managed to create rgb bump maps

06 March 2018 - 09:10 PM

So, you might want to check out this thread: http://deadlystream....rch/?hl=bumpmap


Also, through some extensive research, DarthParametric and I have found that the indexed color bumpmaps are really the same as grayscale/black & white TGA bumpmaps (the way they will be parsed/handled by the game). Which is still nice, it's nice to have bumpmaps on systems that support them. You're just not getting what you might be thinking of if you have access to actual normal maps.


If you want a true full-color tangent-space normal map like the ones the game uses, you just need to convert your 24 or 32-bit TGA to a TPC file containing the appropriate 'isbumpmap 1' TXI value. The vanilla ones used by the game are uncompressed TPC files, but that is not a requirement.

In Topic: How does the game handle transparency?

26 February 2018 - 12:14 AM

The game handles transparency in a lot of ways, as JCarter426 touched on.


Most often the alpha channel in a texture isn't used for transparency, it is used as a sort of 'strength map', either for an environment cubemap or a normal map. More transparent = stronger effect.


There's an alpha controller for any mesh subtype node for specific models. This is used, but somewhat rarely.


Textures can have 'blending punchthrough' in their TXI info, which causes the alpha channel in the texture to be interpreted as transparency and properly take other geometry behind the transparent area into account when compositing.


Models have a 'transparencyhint' in the mesh subtype node header, this acts as a 'blending punchthrough' for specific nodes, and keeps you from having to specify 'blending punchthrough' in TXI information. (It doesn't let you use the alpha channel for other purposes on the same texture though).


Textures can have 'blending additive' in their TXI info, which makes them behave like a 'screen' or 'overlay' blending mode in photoshop, wherein black essentially becomes 100% transparent and white becomes 100% opaque. This might not use information from an alpha channel at all (mostly it is used in textures that don't have alpha channels in the vanilla assets), but I don't know what happens when you do have alpha channel info and are using 'blend additive'.


TPC files have an alpha blending (possibly alpha mean) value in their header, which relates to compositing behavior (you probably wouldn't need to worry about that in unity).


In your original quesion about stars/black backgrounds, just guessing, you are probably looking at a 'blending additive' situation.

In Topic: Anyone still have the .tpc to .tga executable?

23 February 2018 - 07:17 PM

My mistake, it must be a consequence of running KOTORTool through wine. I've never tried it on 'dows. I was unaware you first class citizens get blue normal maps from KOTORTool. For me, specific examples would be 'any'. I guess I'm not surprised though, you guys probably get real data in the TXI field too (I get truncated text or garbage. I pull my TXI text from the end of the TPC files w/ a text editor). I always thought it was odd too, because I had found a thread on LF where the author talked about figuring out how to process the channels properly... and yet it didn't match what my results. Thanks for clearing that up for me.