Universal Sound Separation with Self-Supervised Audio Masked Autoencoder
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929578459529216 |
|---|---|
| author | Zhao, Junqi Liu, Xubo Zhao, Jinzheng Yuan, Yi Kong, Qiuqiang Plumbley, Mark D. Wang, Wenwu |
| author_facet | Zhao, Junqi Liu, Xubo Zhao, Jinzheng Yuan, Yi Kong, Qiuqiang Plumbley, Mark D. Wang, Wenwu |
| contents | Universal sound separation (USS) is a task of separating mixtures of arbitrary sound sources. Typically, universal separation models are trained from scratch in a supervised manner, using labeled data. Self-supervised learning (SSL) is an emerging deep learning approach that leverages unlabeled data to obtain task-agnostic representations, which can benefit many downstream tasks. In this paper, we propose integrating a self-supervised pre-trained model, namely the audio masked autoencoder (A-MAE), into a universal sound separation system to enhance its separation performance. We employ two strategies to utilize SSL embeddings: freezing or updating the parameters of A-MAE during fine-tuning. The SSL embeddings are concatenated with the short-time Fourier transform (STFT) to serve as input features for the separation model. We evaluate our methods on the AudioSet dataset, and the experimental results indicate that the proposed methods successfully enhance the separation performance of a state-of-the-art ResUNet-based USS model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_11745 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Universal Sound Separation with Self-Supervised Audio Masked Autoencoder Zhao, Junqi Liu, Xubo Zhao, Jinzheng Yuan, Yi Kong, Qiuqiang Plumbley, Mark D. Wang, Wenwu Audio and Speech Processing Artificial Intelligence Sound Universal sound separation (USS) is a task of separating mixtures of arbitrary sound sources. Typically, universal separation models are trained from scratch in a supervised manner, using labeled data. Self-supervised learning (SSL) is an emerging deep learning approach that leverages unlabeled data to obtain task-agnostic representations, which can benefit many downstream tasks. In this paper, we propose integrating a self-supervised pre-trained model, namely the audio masked autoencoder (A-MAE), into a universal sound separation system to enhance its separation performance. We employ two strategies to utilize SSL embeddings: freezing or updating the parameters of A-MAE during fine-tuning. The SSL embeddings are concatenated with the short-time Fourier transform (STFT) to serve as input features for the separation model. We evaluate our methods on the AudioSet dataset, and the experimental results indicate that the proposed methods successfully enhance the separation performance of a state-of-the-art ResUNet-based USS model. |
| title | Universal Sound Separation with Self-Supervised Audio Masked Autoencoder |
| topic | Audio and Speech Processing Artificial Intelligence Sound |
| url | https://arxiv.org/abs/2407.11745 |