A 4x4 matrix would be enough. Removing sheer is not an option if you want to maintain the same functionality as Blender as now.
@sybren True, there is Shearing happening. I didn’t realize it at first. And based on quick testing. It actually seems to be quite easy to get.
Showing the 4x4 matrix is possible but is not very Blender like UI. And it won’t be much of help ease the confusion. I would expect in the point when the user can understand the transformation matrix they can handle it with the current Python API.
Maybe having also Share as parameters could be option.
At least it is way more user friendly UI than 4x4 matrix.
1 Like
Maybe having also Share as parameters could be option.
The issue is: a homogenous matrix already includes sheer, it’s how 3D translation can be stored in a matrix, and it’s the reason an extra dimension is required (hence the 4D matrix). If you can find a scientific article or math book that explains how to decompose a homogeneous 4x4 matrix uniquely and unambiguously into 3D loc/rot/scale/sheer, this could be implemented.