DS-Codec: Dual-Stage Training with Mirror-to-NonMirror Architecture Switching for Speech Codec
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916767682527232 |
|---|---|
| author | Chen, Peijie Guan, Wenhao Wang, Kaidi Wu, Weijie Huang, Hukai Hong, Qingyang Li, Lin |
| author_facet | Chen, Peijie Guan, Wenhao Wang, Kaidi Wu, Weijie Huang, Hukai Hong, Qingyang Li, Lin |
| contents | Neural speech codecs are essential for advancing text-to-speech (TTS) systems. With the recent success of large language models in text generation, developing high-quality speech tokenizers has become increasingly important. This paper introduces DS-Codec, a novel neural speech codec featuring a dual-stage training framework with mirror and non-mirror architectures switching, designed to achieve superior speech reconstruction. We conduct extensive experiments and ablation studies to evaluate the effectiveness of our training strategy and compare the performance of the two architectures. Our results show that the mirrored structure significantly enhances the robustness of the learned codebooks, and the training strategy balances the advantages between mirrored and non-mirrored structures, leading to improved high-fidelity speech reconstruction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_24314 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DS-Codec: Dual-Stage Training with Mirror-to-NonMirror Architecture Switching for Speech Codec Chen, Peijie Guan, Wenhao Wang, Kaidi Wu, Weijie Huang, Hukai Hong, Qingyang Li, Lin Sound Audio and Speech Processing Neural speech codecs are essential for advancing text-to-speech (TTS) systems. With the recent success of large language models in text generation, developing high-quality speech tokenizers has become increasingly important. This paper introduces DS-Codec, a novel neural speech codec featuring a dual-stage training framework with mirror and non-mirror architectures switching, designed to achieve superior speech reconstruction. We conduct extensive experiments and ablation studies to evaluate the effectiveness of our training strategy and compare the performance of the two architectures. Our results show that the mirrored structure significantly enhances the robustness of the learned codebooks, and the training strategy balances the advantages between mirrored and non-mirrored structures, leading to improved high-fidelity speech reconstruction. |
| title | DS-Codec: Dual-Stage Training with Mirror-to-NonMirror Architecture Switching for Speech Codec |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2505.24314 |