DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916266997972992 |
|---|---|
| author | Cheuk, Kin Wai Sawata, Ryosuke Uesaka, Toshimitsu Murata, Naoki Takahashi, Naoya Takahashi, Shusuke Herremans, Dorien Mitsufuji, Yuki |
| author_facet | Cheuk, Kin Wai Sawata, Ryosuke Uesaka, Toshimitsu Murata, Naoki Takahashi, Naoya Takahashi, Shusuke Herremans, Dorien Mitsufuji, Yuki |
| contents | In this paper we propose a novel generative approach, DiffRoll, to tackle automatic music transcription (AMT). Instead of treating AMT as a discriminative task in which the model is trained to convert spectrograms into piano rolls, we think of it as a conditional generative task where we train our model to generate realistic looking piano rolls from pure Gaussian noise conditioned on spectrograms. This new AMT formulation enables DiffRoll to transcribe, generate and even inpaint music. Due to the classifier-free nature, DiffRoll is also able to be trained on unpaired datasets where only piano rolls are available. Our experiments show that DiffRoll outperforms its discriminative counterpart by 19 percentage points (ppt.) and our ablation studies also indicate that it outperforms similar existing methods by 4.8 ppt.
Source code and demonstration are available https://sony.github.io/DiffRoll/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2210_05148 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability Cheuk, Kin Wai Sawata, Ryosuke Uesaka, Toshimitsu Murata, Naoki Takahashi, Naoya Takahashi, Shusuke Herremans, Dorien Mitsufuji, Yuki Sound Artificial Intelligence Machine Learning Audio and Speech Processing In this paper we propose a novel generative approach, DiffRoll, to tackle automatic music transcription (AMT). Instead of treating AMT as a discriminative task in which the model is trained to convert spectrograms into piano rolls, we think of it as a conditional generative task where we train our model to generate realistic looking piano rolls from pure Gaussian noise conditioned on spectrograms. This new AMT formulation enables DiffRoll to transcribe, generate and even inpaint music. Due to the classifier-free nature, DiffRoll is also able to be trained on unpaired datasets where only piano rolls are available. Our experiments show that DiffRoll outperforms its discriminative counterpart by 19 percentage points (ppt.) and our ablation studies also indicate that it outperforms similar existing methods by 4.8 ppt. Source code and demonstration are available https://sony.github.io/DiffRoll/. |
| title | DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability |
| topic | Sound Artificial Intelligence Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2210.05148 |