-
Notifications
You must be signed in to change notification settings - Fork 6
Joints are all reported as the same radius #2
Comments
Doesn't LeapC expose joint radius as |
I've made the description of the issue clearer. You are correct that |
Ah, oh well. Thanks for the clarification! |
For what it's worth here, I'd love to see some reasonable default joint sizes! I used to provide my own defaults before I switched to the radius provided by OpenXR, and now I don't have a nice way to decide if I should switch back to my own. Since I use this info for generating finger tubes, this can result in some really noodle-y looking fingers :) |
Are there any improvements for more realistic finger joint radii? I've been using StereoKit on a Varjo XR-3 and there the fingers are just way too big, making hand tracking almost unusable for certain types of interaction (as you loose precision and visual comprehension of where the fingers are). SK directly uses the joint radii provided by OpenXR. |
If you are using OpenXR with a Varjo XR-3, I presume you don't have this API layer installed? The Varjo OpenXR runtime includes built-in support for the Currently, the finger-joint radius is a limitation of the underlying tracking system. I'll pass your feedback onto the team responsible, but we are also planning to look at providing more sensible defaults through this layer in the future. |
@rblenkinsopp I'm mere merely following up the original remarks from StereoKit's developer @maluoi on the hand tracking thickness. I guess StereoKit doesn't directly use this API layer. So would it be better to report this as an OpenXR issue with Varjo directly?
Any progress would be great. Currently, this is what the finger radius looks like in StereoKit: |
Closing this issue as this GitHub is being archived now the API layer is included as part of the standard Ultraleap Tracking Download. We are working on this on our layer, but if you are only working on top of Varjo, you should report it to them as well so it works with their implementation. |
The same radius is reported for all the joints in the hand. This can lead to odd-looking visualisations of the hand if these radii are used to size finger proxy objects.
This is currently a limitation of the underlying LeapC library which returns the same value for all joints. Possible solutions to this include:
The text was updated successfully, but these errors were encountered: