[2604.01929] Woosh: A Sound Effects Foundation Model

Computer Science > Sound arXiv:2604.01929 (cs) [Submitted on 2 Apr 2026] Title:Woosh: A Sound Effects Foundation Model Authors:Gaëtan Hadjeres, Marc Ferras, Khaled Koutini, Benno Weck, Alexandre Bittar, Thomas Hummel, Zineb Lahrici, Hakim Missoum, Joan Serrà, Yuki Mitsufuji View a PDF of the paper titled Woosh: A Sound Effects Foundation Model, by Ga\"etan Hadjeres and 9 other authors View PDF HTML (experimental) Abstract:The audio research community depends on open generative models as foundational tools for building novel approaches and establishing baselines. In this report, we present Woosh, Sony AI's publicly released sound effect foundation model, detailing its architecture, training process, and an evaluation against other popular open models. Being optimized for sound effects, we provide (1) a high-quality audio encoder/decoder model and (2) a text-audio alignment model for conditioning, together with (3) text-to-audio and (4) video-to-audio generative models. Distilled text-to-audio and video-to-audio models are also included in the release, allowing for low-resource operation and fast inference. Our evaluation on both public and private data shows competitive or better performance for each module when compared to existing open alternatives like StableAudio-Open and TangoFlux. Inference code and model weights are available at this https URL. Demo samples can be found at this https URL. Subjects: Sound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs....