JAM: A Tiny Flow-based Song Generator with Fine-grained Controllability and Aesthetic Alignment
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915414633611264 |
|---|---|
| author | Liu, Renhang Hung, Chia-Yu Majumder, Navonil Gautreaux, Taylor Bagherzadeh, Amir Ali Li, Chuan Herremans, Dorien Poria, Soujanya |
| author_facet | Liu, Renhang Hung, Chia-Yu Majumder, Navonil Gautreaux, Taylor Bagherzadeh, Amir Ali Li, Chuan Herremans, Dorien Poria, Soujanya |
| contents | Diffusion and flow-matching models have revolutionized automatic text-to-audio generation in recent times. These models are increasingly capable of generating high quality and faithful audio outputs capturing to speech and acoustic events. However, there is still much room for improvement in creative audio generation that primarily involves music and songs. Recent open lyrics-to-song models, such as, DiffRhythm, ACE-Step, and LeVo, have set an acceptable standard in automatic song generation for recreational use. However, these models lack fine-grained word-level controllability often desired by musicians in their workflows. To the best of our knowledge, our flow-matching-based JAM is the first effort toward endowing word-level timing and duration control in song generation, allowing fine-grained vocal control. To enhance the quality of generated songs to better align with human preferences, we implement aesthetic alignment through Direct Preference Optimization, which iteratively refines the model using a synthetic dataset, eliminating the need or manual data annotations. Furthermore, we aim to standardize the evaluation of such lyrics-to-song models through our public evaluation dataset JAME. We show that JAM outperforms the existing models in terms of the music-specific attributes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_20880 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | JAM: A Tiny Flow-based Song Generator with Fine-grained Controllability and Aesthetic Alignment Liu, Renhang Hung, Chia-Yu Majumder, Navonil Gautreaux, Taylor Bagherzadeh, Amir Ali Li, Chuan Herremans, Dorien Poria, Soujanya Sound Artificial Intelligence Diffusion and flow-matching models have revolutionized automatic text-to-audio generation in recent times. These models are increasingly capable of generating high quality and faithful audio outputs capturing to speech and acoustic events. However, there is still much room for improvement in creative audio generation that primarily involves music and songs. Recent open lyrics-to-song models, such as, DiffRhythm, ACE-Step, and LeVo, have set an acceptable standard in automatic song generation for recreational use. However, these models lack fine-grained word-level controllability often desired by musicians in their workflows. To the best of our knowledge, our flow-matching-based JAM is the first effort toward endowing word-level timing and duration control in song generation, allowing fine-grained vocal control. To enhance the quality of generated songs to better align with human preferences, we implement aesthetic alignment through Direct Preference Optimization, which iteratively refines the model using a synthetic dataset, eliminating the need or manual data annotations. Furthermore, we aim to standardize the evaluation of such lyrics-to-song models through our public evaluation dataset JAME. We show that JAM outperforms the existing models in terms of the music-specific attributes. |
| title | JAM: A Tiny Flow-based Song Generator with Fine-grained Controllability and Aesthetic Alignment |
| topic | Sound Artificial Intelligence |
| url | https://arxiv.org/abs/2507.20880 |