aiqbench analysis
SCAPES introduces a lightweight, resource-efficient generative model that synthesizes environmental audio using continuous latent manifolds and Flow Matching, significantly reducing the computational burden typically associated with high-fidelity sound generation. By enabling training on consumer-grade hardware with uncurated datasets, this approach lowers the barrier for open research in semantic audio control and creative sound design.
SCAPES: Semantically Conditioned Autoregressive Prior for Environmental Sounds
arXiv:2609.04634v1 Announce Type: cross Abstract: As generative audio models grow in complexity, the computational and ecological costs of synthesizing everyday sounds have become increasingly prohibitive, often requiring industrial-scale resources and massive datasets. In this paper, we present SCAPES: a Semantically Conditioned Autoregressive Prior for Environmental Sounds. SCAPES is a lightweight, resource-efficient generative model designed to synthesize high-fidelity environmental textures through high-level semantic control. By operating on the continuous latent manifold of a neural audio codec, our approach bypasses the rigid structural constraints inherent to discrete tokenization. We propose a segmentation strategy that decomposes audio into overlapping segments, enabling a Continuous Normalizing Flow (CNF) to model the evolution of latent trajectories using Flow Matching. Our experiments demonstrate that a 36-million parameter instance of SCAPES can be trained on limited, uncurated datasets using a single consumer-grade GPU. Notably, convergence is achieved after training for approximately twice the source audio duration, yielding high-fidelity outputs with robust long-term stability and semantic consistency. Furthermore, we showcase the model's capacity for smooth semantic interpolation, providing a flexible and accessible tool for open research and creative sound design. Code, pretrained weights, audio examples, and an interactive demo are publicly available on our project page https://cordutie.github.io/projects/scapes.html
Abstract:As generative audio models grow in complexity, the computational and ecological costs of synthesizing everyday sounds have become increasingly prohibitive, often requiring industrial-scale resources and massive datasets. In this paper, we present SCAPES: a Semantically Conditioned Autoregressive Prior for Environmental Sounds. SCAPES is a lightweight, resource-efficient generative model designed to synthesize high-fidelity environmental textures through high-level semantic control. By operating on the continuous latent manifold of a neural audio codec, our approach bypasses the rigid structural constraints inherent to discrete tokenization. We propose a segmentation strategy that decomposes audio into overlapping segments, enabling a Continuous Normalizing Flow (CNF) to model the evolution of latent trajectories using Flow Matching. Our experiments demonstrate that a 36-million parameter instance of SCAPES can be trained on limited, uncurated datasets using a single consumer-grade GPU. Notably, convergence is achieved after training for approximately twice the source audio duration, yielding high-fidelity outputs with robust long-term stability and semantic consistency. Furthermore, we showcase the model's capacity for smooth semantic interpolation, providing a flexible and accessible tool for open research and creative sound design. Code, pretrained weights, audio examples, and an interactive demo are publicly available on our project page this https URL
Accepted to the Digital Audio Fx (DAFx) Conference 2026 to be held in Cambridge, USA. 8 pages, 3 figures and 2 tables
Sound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2609.04634 [cs.SD]
(or arXiv:2609.04634v1 [cs.SD] for this version)
https://doi.org/10.48550/arXiv.2609.04634 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 316 words
Related coverage
AlignAudio: A Unified Dual-Alignment Framework for Noise-Robust Speech-Conditioned Generation
arXiv · Oct 15, 2026
Memory Efficient Audio Synthesis with Decoupled Temporal Depth Diffusion Transformers
arXiv · Jul 28, 2026
FireRedAudio: A General-Purpose Audio Language Model with Decoupled Continuous Representations for Understanding and Generation
arXiv · Aug 26, 2026
AuEmoChat: Authentic Emotion Understanding and Rendering for Conversational Speech Synthesis
arXiv · Jul 20, 2026