We should probably sit back and give them the time they need to work everything out. GECK is naturally going to be the lower of the priorities, and I think that's only fair, considering the ratio of modders to players. I'm going to bet that Beth/Obsidian are busy as heck right now with patching the actual game and whoever works on the GECK is maybe not working on it this second. I'll be paying attention to it to see if I learn anything helpful. I think I had some strange things happen. it might be deeper than just debug messages.
To me this points to a problem in the binary, and I humbly request that Obsidian/Bethesda take a look at this while your in the patching-phase for the game, as I think we may need either a patch for the NV-GECK _or_ a clue on why the assets won't render in most of the interior cells (perhaps there is a.
Lastly, I decompressed all BSA archives into the Data directory to see if loading the loose assets versus loading them from BSA's would make a difference - no change, same behavior. When I select ALL references from the Cell View window (minus the navmeshes) and move them around, I get Less of the references showing up then when I draw a box in the render window around All references and move them - very odd.ĥ. This affects ALL references types, even Xmarkers and Idle markers, not just architecture pieces.Ĥ. If you hit F5 to re-render the cell, any references that you had made visible by moving things around will become invisible again.ģ. If you then move them again, or select different groups of references, you can Sometimes get more of the references to render! Sometimes references vanish when I move them - the behavior is _not_ consistent even within the same cell.Ģ. If you then control-z and undo the move, the same references that just became visible will stay visible. Moving references: If you draw a box in the render window and select many/most references in a cell and Move them slightly, Some of the invisible models will render.
I thus turned my attention away from the assets and to the GECK itself, and have tried several things in the GECK with some fascinating results:ġ. This has to be something else that changed between the time the level designers finished their final debugging and the game went to publishing - otherwise it would work. To confirm this I copies the entire data\dungeons\utility\pits folder from the Fo3 assets to the FNV assets data directory (renaming the old directory out of the way), and tried that - same exact behavior using the UtilPit* kit set from Fallout3 (these assets are shared between both games, so the test was legal). In checking flags between a dozen or more Fo3 and FNV models that are shared between both games, the flagging is identical in most cases for all of the flag fields in NifSkope - they look like mirror copies. Jerry Cox Posts: 3409 Joined: Wed 1:21 pmĪfter looking at several models and testing old collision, I cannot change the behavior in the GECK regarding the rendering of these cells. This is actually an issue worthy of a response from Bethesda or Obsidian, as we're using the NV-GECK with NV-Content and this seems very out of place. I'm going to run one more test in which I load one of the invisible models into Blender, save it out with the less-efficient/older collision layer that Blender can create and see if that makes a difference in the GECK. This is what lead me to examine the collision layer as a potential culprit. It makes me wonder if the changes made during post-processing where they applied a higher version of Havok to the models are part of the problem - as that would have "changed" the models just before launch and after the level designers were 100% done making and fixing the cells. One has to imagine that when the cells were created, the level designers could see everything just fine. I've tried changing all sorts of settings in the GECK to no avail as well. There are different flags on the collision node than other models, but of course we can't read the exact meaning of those flags (if they are even related to this). What I have learned is that we can "Preview" the item from the GECK just fine, and that NifSkope seems to load the objects without issue. No response yet of course, that sub-forum is like an active volcano of new threads and posts.
They are very likely un-related, but I took screenshots of the REPCONN Test Site Basemant cell (in that I could only see half the objects), and asked if they were related. I actually reported this as yesterday, as possibly attached to the problem I have been having (in one cell in particular) of falling off the world geometry.