TARO: Timestep-Adaptive Representation Alignment with Onset-Aware Conditioning for Synchronized Video-to-Audio Synthesis

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ton, Tri, Hong, Ji Woo, Yoo, Chang D.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908586110615552
author Ton, Tri
Hong, Ji Woo
Yoo, Chang D.
author_facet Ton, Tri
Hong, Ji Woo
Yoo, Chang D.
contents This paper introduces Timestep-Adaptive Representation Alignment with Onset-Aware Conditioning (TARO), a novel framework for high-fidelity and temporally coherent video-to-audio synthesis. Built upon flow-based transformers, which offer stable training and continuous transformations for enhanced synchronization and audio quality, TARO introduces two key innovations: (1) Timestep-Adaptive Representation Alignment (TRA), which dynamically aligns latent representations by adjusting alignment strength based on the noise schedule, ensuring smooth evolution and improved fidelity, and (2) Onset-Aware Conditioning (OAC), which integrates onset cues that serve as sharp event-driven markers of audio-relevant visual moments to enhance synchronization with dynamic visual events. Extensive experiments on the VGGSound and Landscape datasets demonstrate that TARO outperforms prior methods, achieving relatively 53% lower Frechet Distance (FD), 29% lower Frechet Audio Distance (FAD), and a 97.19% Alignment Accuracy, highlighting its superior audio quality and synchronization precision.
format Preprint
id arxiv_https___arxiv_org_abs_2504_05684
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TARO: Timestep-Adaptive Representation Alignment with Onset-Aware Conditioning for Synchronized Video-to-Audio Synthesis
Ton, Tri
Hong, Ji Woo
Yoo, Chang D.
Sound
Artificial Intelligence
Computer Vision and Pattern Recognition
This paper introduces Timestep-Adaptive Representation Alignment with Onset-Aware Conditioning (TARO), a novel framework for high-fidelity and temporally coherent video-to-audio synthesis. Built upon flow-based transformers, which offer stable training and continuous transformations for enhanced synchronization and audio quality, TARO introduces two key innovations: (1) Timestep-Adaptive Representation Alignment (TRA), which dynamically aligns latent representations by adjusting alignment strength based on the noise schedule, ensuring smooth evolution and improved fidelity, and (2) Onset-Aware Conditioning (OAC), which integrates onset cues that serve as sharp event-driven markers of audio-relevant visual moments to enhance synchronization with dynamic visual events. Extensive experiments on the VGGSound and Landscape datasets demonstrate that TARO outperforms prior methods, achieving relatively 53% lower Frechet Distance (FD), 29% lower Frechet Audio Distance (FAD), and a 97.19% Alignment Accuracy, highlighting its superior audio quality and synchronization precision.
title TARO: Timestep-Adaptive Representation Alignment with Onset-Aware Conditioning for Synchronized Video-to-Audio Synthesis
topic Sound
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.05684