I have reverted the matrices I changed in 14802. I do so to remove uncertainties prior to measuring new transfer functions for upcoming tasks.

Again, I use Mark's old derivation for these matrices. See figures for the modified matrices.