The exact compensation for the bias in the classical polar-to-Cartesia
n conversion is shown to be multiplicative and to depend on the statis
tics of the cosine of the angle measurement errors. An unbiased conver
sion is presented. A comparison between this unbiased conversion and t
he previously presented debiased conversion is made. The unbiased sphe
rical-to-Cartesian conversion is also presented and evaluated.