Any successful experience when using Blender in order to animate characters for games?
polycounter lvl 7
I use Blender as my main software of choice but there's a huge problem that I noticed. Blender has terrible animation baking so whenever you export an fbx with animation it would be inaccurate which can be very bothersome in some cases. I even had an experience where it would deform part of the body so bad that I got forced to redo the animation. This has actually been a long running limitation of Blender that's not discussed well and will probably not fixed soon.
From what I'm experiencing I would suggest that you avoid animating in Blender if you want really good animation for gamedev. I've been trying to look for workarounds to make the animation accurate because I've already made a rig with Rigify in Blender which is really frustrating. Has anyone had any good experience when animating in Blender for games?
From what I'm experiencing I would suggest that you avoid animating in Blender if you want really good animation for gamedev. I've been trying to look for workarounds to make the animation accurate because I've already made a rig with Rigify in Blender which is really frustrating. Has anyone had any good experience when animating in Blender for games?
Replies
"... inaccurate"
These are very strong, but also undocumented claims. If you are running into such problematic bugs you need to document them, write precise repro steps, and bring them up to the attention of the devs.
If these are indeed bugs they will eventually get fixed. And if they are merely user error, you will get feedback/help on how to solve them.
Animation can be very delicate and the inaccuracy of baking can limit the artistic approach of animators. Despite an obvious problem and is almost never discussed, I'm asking if there is anyone who made animations in Blender to use in games without too much of a hassle.
My personal experience is limited to static poses within the limited context of existing skeletons (either as raw armature, or through the MrMannequin system), also with some occasional exporting out of animations not requiring any baking so I can't comment on much of this topic.
But judging from your observations, I would say ... why even bother with this workflow if it is proven to be broken and/or non-intuitive ? If the issue is identified but isn't actively being worked on by any dev, I think it is pointless to hope/wait for a fix. It would be more logical to use a rock solid industry proven solution instead (Max or Maya) to get things done in the meantime. In this case the license costs would be quite justified.