A2SB: Audio-to-Audio Schrodinger Bridges
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_ | 1866913986888335360 |
|---|---|
| author | Kong, Zhifeng Shih, Kevin J Nie, Weili Vahdat, Arash Lee, Sang-gil Santos, Joao Felipe Jukic, Ante Valle, Rafael Catanzaro, Bryan |
| author_facet | Kong, Zhifeng Shih, Kevin J Nie, Weili Vahdat, Arash Lee, Sang-gil Santos, Joao Felipe Jukic, Ante Valle, Rafael Catanzaro, Bryan |
| contents | Real-world audio is often degraded by numerous factors. This work presents an audio restoration model tailored for high-res music at 44.1kHz. Our model, Audio-to-Audio Schrödinger Bridges (A2SB), is capable of both bandwidth extension (predicting high-frequency components) and inpainting (re-generating missing segments). Critically, A2SB is end-to-end requiring no vocoder to predict waveform outputs, able to restore hour-long audio inputs, and trained on permissively licensed music data. A2SB is capable of achieving state-of-the-art band-width extension and inpainting quality on several out-of-distribution music test sets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_11311 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A2SB: Audio-to-Audio Schrodinger Bridges Kong, Zhifeng Shih, Kevin J Nie, Weili Vahdat, Arash Lee, Sang-gil Santos, Joao Felipe Jukic, Ante Valle, Rafael Catanzaro, Bryan Sound Machine Learning Audio and Speech Processing Real-world audio is often degraded by numerous factors. This work presents an audio restoration model tailored for high-res music at 44.1kHz. Our model, Audio-to-Audio Schrödinger Bridges (A2SB), is capable of both bandwidth extension (predicting high-frequency components) and inpainting (re-generating missing segments). Critically, A2SB is end-to-end requiring no vocoder to predict waveform outputs, able to restore hour-long audio inputs, and trained on permissively licensed music data. A2SB is capable of achieving state-of-the-art band-width extension and inpainting quality on several out-of-distribution music test sets. |
| title | A2SB: Audio-to-Audio Schrodinger Bridges |
| topic | Sound Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2501.11311 |