Skip to content

v2.0.5: Modernized Skins support, full NPC support, etc.

Latest
Compare
Choose a tag to compare
@ZeroGravitasIndeed ZeroGravitasIndeed released this 05 Feb 22:15
· 6 commits to main since this release
96cb99d

THIS VERSION REQUIRES THE USE OF THE LATEST VERSION OF THE .GR2 IMPORTER ADD-ON. THE DOWNLOAD SECTION BELOW INCLUDES IT.

Character Assembler tool compatible with modernized SWTOR skin materials

What the change does is replace the paths.json's texturemaps information with that of the .mat files also referenced by it, which is always correct for the asset extraction they belong to, be it modernized or "classic". So, depending on which resources folder is selected in the add-on's Preferences (extracted before or after Game Update 7.6), we ought to get the correct classic or modernized results.

(This should also solve those mistakes TORC's Character Designer sometimes does, like those wrong Cathar PC texture filepaths. The insidious "/.dds" ultimately seems to be a symptom of another issue, discussed in the next point, that we now cater to)

The tool saves a paths_corrected.json file alongside the original, which is what it feeds to the .gr2 add-on's actual character importer.

This is a partial solution. The idea is to end up being able to use both classic and modern textures without having to change the resources folder in Preferences, and saving each set in a separate subfolder.

CAUTION: its use will replace a PC/NPC folder's already existing textures with classic or modernized ones, depending on the selected resources folder in Preferences. If wanting to avoid that, use the tool in a copy or a newly decompressed .zip's PC/NPC folder.

(It will also litter the subfolders with unused textures of the opposite type because BS didn't keep the same filenames in all cases. We are seeing about having separate Materials subfolders to solve that)

Creature-type NPC assembling works now

The .gr2 Importer Add-on's code has been revised to cater to Creature-type NPC exports from TORCommunity.com's database. It ought to solve issues such as second materials that aren't eyes (see Darth Malgus Lvl. 75's case), characters without a second material at all (see Massassi), etc. This version of the ZG SWTOR Tools Add-on has been revised in turn to be compatible with those changes.

We've found out the cause of broken "/.dds" texturepath entries: mismatches between the expected shader (Creature, which doesn't use paletteMaps) and the stated one (Eye, which does). If finding those malformed entries in the paths.json files, don't correct them. We are using their presence to detect those mismatches and handle them correctly.

Also, the skeleton.json information was incorrect for creatures (it included a "new" in their names that didn't exist). The assembling gathers and applies the correct skeletons now, and also creates "corrected" versions of the .json files for reference.

Support for the materialSkinIndex flag has been restored

This flag says which material in an armor/clothing object is meant to be skin instead of garment, if any.

Player Characters-only, though: non-Creature NPCs' paths.json files lack the data, so, for those we are adding to the tool a checkbox, pre-set to choose skin as the less inconvenient default. Correcting wrong results is a matter of applying the first Material Slot's garment material to the second Slot, replacing the skin, just a couple of clicks.

(A typical example of the use of this checkbox is the cantina dancer NPCs and companions in underwear)

Revised image deduplicator

As SWTOR 7.6' skin modernization can lead to misdeduplications of textures with matched root names but different contents (classic and modernized) if one has projects that use both, the deduplicator checks the images' sizes in pixels before acting. We are looking at using content hashing in a future version, for a more rigorous process.

Revised Physics-related Vertex Groups Merger tool

It has better console reporting, and corrects some bugs that made the use of manually entered common VG text not work at all. We've found cases where the best guessing of the common text fails: we'll try to add more smarts to it.