cs.SDNov 7, 2026

Steer Unified Audio Editing Toward Reliable Source Selection

Authors: Youquan Fu, Dongnuan Cai, Yu Pan, Henghui Du, Yujia Zhang, Ping Luo, Quande Liu, Yake Wei, 

Organizations: Renmin University of China · Tencent · The University of Hong Kong · The Chinese University of Hong Kong

Abstract

Unified audio editors support diverse operations, yet often struggle to modify the intended source while preserving non-target content. We show that source selection in audio editing is shaped by early denoising states, while incorrect early selections can hardly be corrected during subsequent denoising. Motivated by these findings, we propose Source-selection Training via Early-trajectory Emphasis and Recovery (STEER), a two-stage post-training algorithm for reliable selective audio editing. STEER-Emphasis prioritizes training on task-critical early states to reduce source-selection errors, while STEER-Recovery learns corrective denoising from self-generated rollout states to recover from early deviations. To systematically assess source-selection reliability, we introduce the Selective Audio Editing Benchmark, enabling large-scale automatic evaluation of target manipulation and non-target preservation. Applying STEER to our pretrained multitask base editor gives SteerSound, a unified audio editor with leading selective editing performance among open-source audio editors. Experiments on our base editor, UNISON, and SAM Audio demonstrate that STEER consistently improves both selective editing capability and overall editing performance. Early-trajectory emphasis further brings substantial improvements in transcript-conditioned target speaker removal and separation, highlighting the broader applicability of this training principle.

Figures & tables

Explore similar work

CardsList
  1. DirectAudioEdit: Inversion-Free Text-Guided Audio Editing via Diffusion Prediction Contrast

    Jun 5, 2026Zhengkun Ge, Xiaoqian Liu, Haoran Zhang +5Audio EditingAudio Diffusion Models

  2. FreeSonic: Training-Free Temporal-Aware Decoupled Attention for Precise Audio Editing

    Jun 13, 2026Yuxuan Jiang, Mingyang Han, Yusheng Dai +12Audio EditingAudio Diffusion Models

  3. AUDEDIT: Inversion-Free Text-Guided Editing with Pretrained Audio Flow Models

    Jun 13, 2026Zhongyuan FuAudio EditingAudio Generation