You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Oct 31, 2023. It is now read-only.
It's more probably a Blender issue rather than a FrankMocap issue, but here there might be someone that already was capable of accomplish this task among this community, so this is why I am writing here.
My goal is to apply hand rotations coming from FrankMocap to a Blender character. Let's first state some concepts I think I understood while reading the paper and the documentation.
FrankMocap estimates finger rotation based on the parent-bone-axes. This means for instance that each finger's first bone will be defined based on the wrist reference system. I guess that the wrist reference system is consistent with the forearm direction. But how exactly? In blender each bone expresses the +Y axis of its local reference system, while X and Z can be set with an arbitrary roll. E.g. in the human T-pose I have arms with X point upwards. What about FrankMocap? Is there a clear representation of it? Something like this? If FrankMocap orientation is quite standard (or if it copies someone else's definition) can you point me to some documentation of it?
I assume angles measure orientations and not rotations, since FrankMocap is frame-based does not consider movements, right? I guess that rotation can be computed as a delta compared to previous frame.
[This is the real issue] In Blender you can express rotation in Euler angles, Axis-angles or Quaternions (see here). Anyway I understood that all of them are referred to the axis system located at the head (i.e. base) of the bone. Are you aware of an approach to transfer parent-based orientation to local axis coordinates? I fear this requires some robotics skill I cannot master.
Any comment on these topics might certainly be helpful. Thanks!
The text was updated successfully, but these errors were encountered:
@penincillin I can the hand movement can calculated along with your body hand joint.
For example, you have your left hand joint rotation, you can get your left hand joint position from pelvis, then just keep going get position of your hands by predicitons of hands.
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
It's more probably a Blender issue rather than a FrankMocap issue, but here there might be someone that already was capable of accomplish this task among this community, so this is why I am writing here.
My goal is to apply hand rotations coming from FrankMocap to a Blender character. Let's first state some concepts I think I understood while reading the paper and the documentation.
FrankMocap estimates finger rotation based on the parent-bone-axes. This means for instance that each finger's first bone will be defined based on the wrist reference system. I guess that the wrist reference system is consistent with the forearm direction. But how exactly? In blender each bone expresses the +Y axis of its local reference system, while X and Z can be set with an arbitrary roll. E.g. in the human T-pose I have arms with X point upwards. What about FrankMocap? Is there a clear representation of it? Something like this?
If FrankMocap orientation is quite standard (or if it copies someone else's definition) can you point me to some documentation of it?
I assume angles measure orientations and not rotations, since FrankMocap is frame-based does not consider movements, right? I guess that rotation can be computed as a delta compared to previous frame.
[This is the real issue] In Blender you can express rotation in Euler angles, Axis-angles or Quaternions (see here). Anyway I understood that all of them are referred to the axis system located at the head (i.e. base) of the bone. Are you aware of an approach to transfer parent-based orientation to local axis coordinates? I fear this requires some robotics skill I cannot master.
Any comment on these topics might certainly be helpful. Thanks!
The text was updated successfully, but these errors were encountered: