PSE Dataset Demo
This is a small selection from our Sound Effects Dataset so you can get a sense of how its structured, the quality of the audio samples and associated metadata, and how it could integrate with your training pipelines.
Deep Dive
Access additional resources and information about our Sound Effects Dataset
Universal Category System
Citations
Selected research that has utilized and evaluated audio data from PSE:
- Learning Control of Neural Sound Effects Synthesis from Physically Inspired Models (ICASSP, 2025)
- Audio-Language Datasets of Scenes and Events: A Survey (IEEE, 2025)
- A Machine learning method to evaluate and improve sound effects synthesis model design (AES, 2024)
- Text-Driven Separation of Arbitrary Sounds (INTERSPEECH, 2022)
- Representation Learning for the Automatic Indexing of Sound Effects Libraries (ISMIR, 2022)
- Multiple-Embedding Separation Networks: Sound Class-Specific Feature Extraction for Universal Sound Separation (IEEE, 2021)
- Improving Universal Sound Separation Using Sound Classification (ICASSP 2020)
-
Universal Sound Separation (WASPAA, 2019)