Contrast-PLC: Contrastive Learning for Packet Loss Concealment

TL;DR

This paper introduces Contrast-PLC, a novel contrastive learning approach for packet loss concealment that enhances long burst loss prediction by learning robust semantic representations, outperforming traditional and GAN-based models.

Contribution

It proposes a contrastive learning framework combined with a hybrid neural architecture for improved long burst packet loss concealment.

Findings

Outperforms UNet-style frameworks on Interspeech2022 PLC Challenge

Shows significant improvement for burst losses between 120 and 220 ms

Demonstrates robustness in loss-robust semantic representation learning

Abstract

Packet loss concealment (PLC) is challenging in concealing missing contents both plausibly and naturally when there are only limited available context to use. Recently deep-learning based PLC algorithms have demonstrated their superiority over traditional counterparts; but their concealment ability is still mostly limited to a maximum of 120ms loss. Even with strong GAN-based generative models, it is still very challenging to predict long burst losses that could happen within/in-between phonemes. In this paper, we propose to use contrastive learning to learn a loss-robust semantic representation for PLC. A hybrid neural PLC architecture combining the semantic prediction and GAN-based generative model is designed to verify its effectiveness. Results on the blind test set of Interspeech2022 PLC Challenge show its superiority over commonly used UNet-style framework and the one without contrastive learning, especially for the longer burst loss at (120, 220] ms.