Analysis of Optimal High Resolution and Fixed Rate Scalar Quantization

In 2001, Hui and Neuhoff proposed a
uniform

quantizer

with

overload
for the quantization of scalar signals and derived the asymptotically optimal size of the quantization bins in the high-bitrate limit. The purpose of the present paper is to prove a quantitatively more precise version of this result which, at the same time, is valid for a more general, quite natural class of probability distributions that requires only little regularity and includes, for instance, positive Lipschitz-continuous functions of unit integral.