It is desired to encode a random variable

using an

-level quantizer

to minimize the expected distortion

, where the error weighting function

is convex, strictly increasing and continuously differentiable. It is shown that if

has a log-concave density, then there exists a unique locally optimal quantizer

and Lloyd\´s Method I may be used to find

. Trushkin had earlier shown this result for the error weighting functions

and

.