FlowSynth: Instrument Generation Through Distributional Flow Matching and Test-Time Search
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908610297069568 |
|---|---|
| author | Yang, Qihui Leistikow, Randal Zang, Yongyi |
| author_facet | Yang, Qihui Leistikow, Randal Zang, Yongyi |
| contents | Virtual instrument generation requires maintaining consistent timbre across different pitches and velocities, a challenge that existing note-level models struggle to address. We present FlowSynth, which combines distributional flow matching (DFM) with test-time optimization for high-quality instrument synthesis. Unlike standard flow matching that learns deterministic mappings, DFM parameterizes the velocity field as a Gaussian distribution and optimizes via negative log-likelihood, enabling the model to express uncertainty in its predictions. This probabilistic formulation allows principled test-time search: we sample multiple trajectories weighted by model confidence and select outputs that maximize timbre consistency. FlowSynth outperforms the current state-of-the-art TokenSynth baseline in both single-note quality and cross-note consistency. Our approach demonstrates that modeling predictive uncertainty in flow matching, combined with music-specific consistency objectives, provides an effective path to professional-quality virtual instruments suitable for real-time performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_21667 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FlowSynth: Instrument Generation Through Distributional Flow Matching and Test-Time Search Yang, Qihui Leistikow, Randal Zang, Yongyi Sound Virtual instrument generation requires maintaining consistent timbre across different pitches and velocities, a challenge that existing note-level models struggle to address. We present FlowSynth, which combines distributional flow matching (DFM) with test-time optimization for high-quality instrument synthesis. Unlike standard flow matching that learns deterministic mappings, DFM parameterizes the velocity field as a Gaussian distribution and optimizes via negative log-likelihood, enabling the model to express uncertainty in its predictions. This probabilistic formulation allows principled test-time search: we sample multiple trajectories weighted by model confidence and select outputs that maximize timbre consistency. FlowSynth outperforms the current state-of-the-art TokenSynth baseline in both single-note quality and cross-note consistency. Our approach demonstrates that modeling predictive uncertainty in flow matching, combined with music-specific consistency objectives, provides an effective path to professional-quality virtual instruments suitable for real-time performance. |
| title | FlowSynth: Instrument Generation Through Distributional Flow Matching and Test-Time Search |
| topic | Sound |
| url | https://arxiv.org/abs/2510.21667 |