English

Re-Bottleneck: Latent Re-Structuring for Neural Audio Autoencoders

Sound 2025-09-10 v2 Machine Learning Audio and Speech Processing

Abstract

Neural audio codecs and autoencoders have emerged as versatile models for audio compression, transmission, feature-extraction, and latent-space generation. However, a key limitation is that most are trained to maximize reconstruction fidelity, often neglecting the specific latent structure necessary for optimal performance in diverse downstream applications. We propose a simple, post-hoc framework to address this by modifying the bottleneck of a pre-trained autoencoder. Our method introduces a "Re-Bottleneck", an inner bottleneck trained exclusively through latent space losses to instill user-defined structure. We demonstrate the framework's effectiveness in three experiments. First, we enforce an ordering on latent channels without sacrificing reconstruction quality. Second, we align latents with semantic embeddings, analyzing the impact on downstream diffusion modeling. Third, we introduce equivariance, ensuring that a filtering operation on the input waveform directly corresponds to a specific transformation in the latent space. Ultimately, our Re-Bottleneck framework offers a flexible and efficient way to tailor representations of neural audio models, enabling them to seamlessly meet the varied demands of different applications with minimal additional training.

Keywords

Cite

@article{arxiv.2507.07867,
  title  = {Re-Bottleneck: Latent Re-Structuring for Neural Audio Autoencoders},
  author = {Dimitrios Bralios and Jonah Casebeer and Paris Smaragdis},
  journal= {arXiv preprint arXiv:2507.07867},
  year   = {2025}
}

Comments

Accepted at IEEE MLSP 2025

R2 v1 2026-07-01T03:55:01.584Z