SAO-Instruct: Free-form Audio Editing using Natural Language Instructions

TL;DR

SAO-Instruct is a novel model that enables flexible, natural language-based editing of audio clips, trained on a new dataset, and demonstrates strong performance both objectively and subjectively.

Contribution

The paper introduces SAO-Instruct, the first model capable of free-form natural language audio editing, along with a new dataset and training pipeline for this task.

Findings

SAO-Instruct outperforms existing methods in subjective listening tests.

The model generalizes well to real-world audio and unseen instructions.

It achieves competitive objective metrics on audio editing benchmarks.

Abstract

Generative models have made significant progress in synthesizing high-fidelity audio from short textual descriptions. However, editing existing audio using natural language has remained largely underexplored. Current approaches either require the complete description of the edited audio or are constrained to predefined edit instructions that lack flexibility. In this work, we introduce SAO-Instruct, a model based on Stable Audio Open capable of editing audio clips using any free-form natural language instruction. To train our model, we create a dataset of audio editing triplets (input audio, edit instruction, output audio) using Prompt-to-Prompt, DDPM inversion, and a manual editing pipeline. Although partially trained on synthetic data, our model generalizes well to real in-the-wild audio clips and unseen edit instructions. We demonstrate that SAO-Instruct achieves competitive performance on objective metrics and outperforms other audio editing approaches in a subjective listening study. To encourage future research, we release our code and model weights.