I think I’ve just had the most sluggish two weeks out of the whole time developing this. What pray tell have I been doing? I’ve been fidgeting with Felix’s design again. I even went as far as to update the old game with a more accurate design so I’d have something better to work off. 3D isn’t very easy. Skinning is hard, animating is hard, and I still have very little idea what the heck I’m doing with the inverse bind pose. I’m almost certain it’ll jump up to bite me in the neck later.
But ever since I’ve updated my sprites, I’ve been able to pick up the pace and it’s only a matter of before those problems are nailed down. Right now I’m attempting to finalize the skinning and animation. It’s hard to move ahead in-game when the model isn’t exactly moving right.
I haven’t filled in the eyes yet so Felix has the look I use in concept art.
Those eyes make me think of early Disney art. It’s a cool style. And yeah, you definitely have a perfectionist side ;P i know how it is.
“Inverse bind pose” reminds me of loading MS3D models, where the vertices (stored in model-space) have to be inverse-transformed into the joint-space – that each group of vertices belong/binds to. And the only reason for that is so each joint-matrix can transform its vertices out to model space after being rotatated/moved/”animated” – Since joints could not be hierarchical (e.g. spine-shoulder-elbow) if their matrices were already in model space (each joint needs to be in its root’s space, and therefore its animated vertices need to be there, too) – That’s just MS3D though! I’m sure your tools are slightly different.
Do you use Blender somewhere now?
I just realized how utterly messed up my sentences of the joint elaboration was – whoa I barely get what I meant. Here, have some paragraphs instead : D
In MS3D model files, vertices are stored in model-space, and have to be inverse-transformed into the joint-space that each group of vertices belong/binds to.
The only reason for that is so each joint-matrix can directly transform its vertices out to model space after being rotatated/moved/animated.
The joints are hierarchical, e.g. the spine moves shoulder which correspondingly moves elbow – That would not be possible if joint-matrices would be in model space. So each joint needs to be in its root’s space, and the same goes for its vertices, which is why they had to be inverse-transformed once after the MS3D file was loaded.
Maybe it was something similar you did with inverses, I’m not sure!
I don’t think you messed up your sentences, I just think it’s weird to explain : j
I think I understand the point of it, but setting it up inside the program is was a pain and I think I’m a tad paranoid it’s going to break at some point.
As it turns out, I’ve eliminated any need to turn to other programs to do anything. After having used Blender, I found the controls to be toxic and the layout to be very very strange. I guess you can’t argue with a free program but the learning curve is way too step just to do a little outsourcing.
Also Disney is the first thing I thought of with the eyes! Feels easier to work with for some reason.