Import GLA taking very long time

[DANIEL120321]

traying to import gla file but import animation taking very long time. my cpu is i9 9700K RTX4090

why its taiking so long ? there is any way to make it faster ?

[mrwonko]

Adding another frame of animation takes longer the more frames there already are. This makes importing a large number of frames very slow. Humanoid has over 25K frames in Jedi Academy, importing all of it can take multiple hours.

I don't know if this bad performance is inherent in Blender, or if there's a faster way to import all animations into one long sequence. But this should be fixable somehow.

Do you really need all frames? You can import a range of frames to edit particular animations, then use glamerge to append your modified animations to the original humanoid file.

If you do need to import all of them, remember you only need to do it once, then you can save it as a regular .blend some, which loads much faster.

But I wonder if a better solution might be to import the animations individually by parsing the animations.cfg. Creating many short sequences instead of a single long one might prevent the slowdown.

[DANIEL120321]

Thanks for your replay. i have a file that contains the information about the frames range in these format:

m:/_animation/xsi/male01/fmove/any/stand/normal/67plus/000/normal/swim01.xsi { "startframe" "10990" "duration" "21" "fps" "10" "averagevec" "0.000 -5.152 0.000" }

m:/_animation/xsi/male01/inventory/helicoptor/sit/normal/67plus/aim/sithueystick01.xsi { "startframe" "12486" "duration" "2" "fps" "20" "averagevec" "0.000 0.000 0.000" }

and more...

so i need every time import the frame range manually ? i just need all of the animations whats the best way getting all of the sequences ?

[DT85]

@DANIEL120321 Those .frames files are generated by Carcass, which is unrelated to the Blender tools.

@mrwonko I did a test a few months back with converting the base game GLA to FBX binary using Noesis, then importing that FBX into Blender and it doesn't take very long to process all frames. Perhaps the FBX importer is using a different method that is faster.