Self-Supervised Singing Voice Pre-Training towards Speech-to-Singing Conversion
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929373434609664 |
|---|---|
| author | Li, Ruiqi Huang, Rongjie Wang, Yongqi Hong, Zhiqing Zhao, Zhou |
| author_facet | Li, Ruiqi Huang, Rongjie Wang, Yongqi Hong, Zhiqing Zhao, Zhou |
| contents | Speech-to-singing voice conversion (STS) task always suffers from data scarcity, because it requires paired speech and singing data. Compounding this issue are the challenges of content-pitch alignment and the suboptimal quality of generated outputs, presenting significant hurdles in STS research. This paper presents SVPT, an STS approach boosted by a self-supervised singing voice pre-training model. We leverage spoken language model techniques to tackle the rhythm alignment problem and the in-context learning capability to achieve zero-shot conversion. We adopt discrete-unit random resampling and pitch corruption strategies, enabling training with unpaired singing data and thus mitigating the issue of data scarcity. SVPT also serves as an effective backbone for singing voice synthesis (SVS), offering insights into scaling up SVS models. Experimental results indicate that SVPT delivers notable improvements in both STS and SVS endeavors. Audio samples are available at https://speech2sing.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_02429 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Self-Supervised Singing Voice Pre-Training towards Speech-to-Singing Conversion Li, Ruiqi Huang, Rongjie Wang, Yongqi Hong, Zhiqing Zhao, Zhou Audio and Speech Processing Sound Speech-to-singing voice conversion (STS) task always suffers from data scarcity, because it requires paired speech and singing data. Compounding this issue are the challenges of content-pitch alignment and the suboptimal quality of generated outputs, presenting significant hurdles in STS research. This paper presents SVPT, an STS approach boosted by a self-supervised singing voice pre-training model. We leverage spoken language model techniques to tackle the rhythm alignment problem and the in-context learning capability to achieve zero-shot conversion. We adopt discrete-unit random resampling and pitch corruption strategies, enabling training with unpaired singing data and thus mitigating the issue of data scarcity. SVPT also serves as an effective backbone for singing voice synthesis (SVS), offering insights into scaling up SVS models. Experimental results indicate that SVPT delivers notable improvements in both STS and SVS endeavors. Audio samples are available at https://speech2sing.github.io. |
| title | Self-Supervised Singing Voice Pre-Training towards Speech-to-Singing Conversion |
| topic | Audio and Speech Processing Sound |
| url | https://arxiv.org/abs/2406.02429 |