Fewer-token Neural Speech Codec with Time-invariant Codes

TL;DR

This paper introduces TiCodec, a neural speech codec that uses time-invariant codes to reduce token count, improve speech reconstruction quality, and enhance zero-shot TTS performance.

Contribution

The paper proposes a novel time-invariant neural speech codec with a consistency loss, reducing token count and improving TTS quality and similarity.

Findings

Fewer tokens are needed for speech representation with TiCodec.

Speech reconstruction quality is improved with TiCodec.

Zero-shot TTS performance is enhanced, with higher naturalness and lower word error rate.

Abstract

Language model based text-to-speech (TTS) models, like VALL-E, have gained attention for their outstanding in-context learning capability in zero-shot scenarios. Neural speech codec is a critical component of these models, which can convert speech into discrete token representations. However, excessive token sequences from the codec may negatively affect prediction accuracy and restrict the progression of Language model based TTS models. To address this issue, this paper proposes a novel neural speech codec with time-invariant codes named TiCodec. By encoding and quantizing time-invariant information into a separate code, TiCodec can reduce the amount of frame-level information that needs encoding, effectively decreasing the number of tokens as codes of speech. Furthermore, this paper introduces a time-invariant encoding consistency loss to enhance the consistency of time-invariant code within an utterance and force it to capture more global information, which can benefit the zero-shot TTS task. Experimental results demonstrate that TiCodec can not only enhance the quality of reconstruction speech with fewer tokens but also increase the similarity and naturalness, as well as reduce the word error rate of the synthesized speech by the TTS model.