ClariCodec: Optimising Neural Speech Codes for 200bps Communication using Reinforcement Learning
Abstract
In bandwidth-constrained communication such as satellite and underwater channels, speech must often be transmitted at ultra-low bitrates where intelligibility is the primary objective. At such extreme compression levels, codecs trained with acoustic reconstruction losses tend to allocate bits to perceptual detail, leading to substantial degradation in word error rate (WER). This paper proposes ClariCodec, a neural speech codec operating at 200 bit per second (bps) that reformulates quantisation as a stochastic policy, enabling reinforcement learning (RL)-based optimisation of intelligibility. Specifically, the encoder is fine-tuned using WER-driven rewards while the acoustic reconstruction pipeline remains frozen. Even without RL, ClariCodec achieves 3.68% WER on the LibriSpeech test-clean set at 200 bps, already competitive with codecs operating at higher bitrates. Further RL fine-tuning reduces WER to 3.20% on test-clean and 8.93% on test-other, corresponding to a 13% relative reduction while preserving perceptual quality.
Keywords
Cite
@article{arxiv.2604.14654,
title = {ClariCodec: Optimising Neural Speech Codes for 200bps Communication using Reinforcement Learning},
author = {Junyi Wang and Chi Zhang and Jing Qian and Haifeng Luo and Hao Wang and Zengrui Jin and Chao Zhang},
journal= {arXiv preprint arXiv:2604.14654},
year = {2026}
}
Comments
Withdrawn by the authors due to incomplete bitrate accounting in the ILN-based pipeline. The side information introduced by ILN was not fully included in the effective bitrate, making the reported 200 bps results and related comparisons unreliable. The withdrawal does not concern the paper's core RL-based methodological idea. A corrected version may follow