A simple Sign-Bit Probabilistic Shaping Scheme

TL;DR

This paper introduces a low-complexity sign-bit probabilistic shaping scheme for Gaussian channels that achieves most shaping gains with simplified binary sources, reducing complexity by half compared to traditional methods.

Contribution

The paper presents a novel sign-bit shaping scheme using only two binary sources, significantly reducing complexity while maintaining shaping performance.

Findings

Achieves most shaping gain with simplified binary sources.

Reduces complexity to approximately half of traditional binary distribution matchers.

Uses a Maxwell-Boltzmann-like distribution for efficient shaping.

Abstract

We propose a new shaping scheme for the Gaussian channel whose complexity is approximately half the one of a binary distribution matcher (DM). The result is obtained as follows: We first show that most of the shaping gain can be obtained via a simplified version of sign-bit shaping, which uses only two non-uniform binary sources. This is achieved by considering a stepwise Maxwell-Boltzmann-like distribution of the symbols. One of the two binary sources has a parameter $p$ close to 0. Hence, we then describe a binary DM which explicitly takes advantage of this aspect and has a negligible complexity. Since the two binary sources are used alternately with equal probability, the complexity of the proposed shaping scheme is half the one of the second binary DM.