Neural Feature Predictor and Discriminative Residual Coding for Low-Bitrate Speech Coding

TL;DR

This paper presents a neural speech coding system that reduces redundancy through predictive coding and discriminative residual coding, achieving higher efficiency at very low bitrates with a scalable, lightweight design.

Contribution

It introduces a novel combination of neural feature prediction and discriminative residual coding for improved low-bitrate speech coding.

Findings

Achieves superior coding efficiency compared to LPCNet and Lyra V2 at very low bitrates.

Develops a dynamic bit allocation algorithm based on residual contribution.

Demonstrates causality and low-latency performance in the proposed system.

Abstract

Low and ultra-low-bitrate neural speech coding achieves unprecedented coding gain by generating speech signals from compact speech features. This paper introduces additional coding efficiency in neural speech coding by reducing the temporal redundancy existing in the frame-level feature sequence via a recurrent neural predictor. The prediction can achieve a low-entropy residual representation, which we discriminatively code based on their contribution to the signal reconstruction. The harmonization of feature prediction and discriminative coding results in a dynamic bit allocation algorithm that spends more bits on unpredictable but rare events. As a result, we develop a scalable, lightweight, low-latency, and low-bitrate neural speech coding system. We demonstrate the advantage of the proposed methods using the LPCNet as a neural vocoder. While the proposed method guarantees causality in its prediction, the subjective tests and feature space analysis show that our model achieves superior coding efficiency compared to LPCNet and Lyra V2 in the very low bitrates.