ARSCNView.pointOfView.simdWorldTransform VS ARSCNView.session.session.currentFrame.camera.transform

I tought these two are the same, but in reality they are not. My observation is:


if

ARSCNView.pointOfView.simdWorldTransform = [c1, c2, c3, c4] // c is column vector

then

ARSCNView.session.session.currentFrame.camera.transform = [-c2, c1, c3, c4]

that is 1 column negation and exchange.


Why?

ARSCNView.pointOfView.simdWorldTransform VS ARSCNView.session.session.currentFrame.camera.transform
 
 
Q