Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.23759 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914302569480192 |
|---|---|
| author | Wang, Ting-Kang Peng, Yueh-Po Su, Li Cheung, Vincent K. M. |
| author_facet | Wang, Ting-Kang Peng, Yueh-Po Su, Li Cheung, Vincent K. M. |
| contents | While automatic music transcription is well-established in music information retrieval, most models are limited to transcribing pitch and timing information from audio, and thus omit crucial expressive and instrument-specific nuances. One example is playing technique on the violin, which affords its distinct palette of timbres for maximal emotional impact. Here, we propose VioPTT (Violin Playing Technique-aware Transcription), a lightweight cascade model that directly transcribes violin playing technique in addition to pitch onset and offset. Furthermore, we release MOSA-VPT, a novel, high-quality synthetic violin playing technique dataset to circumvent the need for manually labeled annotations. Leveraging this dataset, our model demonstrated strong generalization to real-world note-level violin technique recordings in addition to achieving state-of-the-art transcription performance. To our knowledge, VioPTT is the first to jointly combine violin transcription and playing technique prediction within a unified framework. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_23759 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | VioPTT: Violin Technique-Aware Transcription from Synthetic Data Augmentation Wang, Ting-Kang Peng, Yueh-Po Su, Li Cheung, Vincent K. M. Sound Machine Learning While automatic music transcription is well-established in music information retrieval, most models are limited to transcribing pitch and timing information from audio, and thus omit crucial expressive and instrument-specific nuances. One example is playing technique on the violin, which affords its distinct palette of timbres for maximal emotional impact. Here, we propose VioPTT (Violin Playing Technique-aware Transcription), a lightweight cascade model that directly transcribes violin playing technique in addition to pitch onset and offset. Furthermore, we release MOSA-VPT, a novel, high-quality synthetic violin playing technique dataset to circumvent the need for manually labeled annotations. Leveraging this dataset, our model demonstrated strong generalization to real-world note-level violin technique recordings in addition to achieving state-of-the-art transcription performance. To our knowledge, VioPTT is the first to jointly combine violin transcription and playing technique prediction within a unified framework. |
| title | VioPTT: Violin Technique-Aware Transcription from Synthetic Data Augmentation |
| topic | Sound Machine Learning |
| url | https://arxiv.org/abs/2509.23759 |