Foley Control: Aligning a Frozen Latent Text-to-Audio Model to Video

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Rowles, Ciara, Jampani, Varun, Donné, Simon, Vainer, Shimon, Parker, Julian, Evans, Zach
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918169966280704
author Rowles, Ciara
Jampani, Varun
Donné, Simon
Vainer, Shimon
Parker, Julian
Evans, Zach
author_facet Rowles, Ciara
Jampani, Varun
Donné, Simon
Vainer, Shimon
Parker, Julian
Evans, Zach
contents Foley Control is a lightweight approach to video-guided Foley that keeps pretrained single-modality models frozen and learns only a small cross-attention bridge between them. We connect V-JEPA2 video embeddings to a frozen Stable Audio Open DiT text-to-audio (T2A) model by inserting compact video cross-attention after the model's existing text cross-attention, so prompts set global semantics while video refines timing and local dynamics. The frozen backbones retain strong marginals (video; audio given text) and the bridge learns the audio-video dependency needed for synchronization -- without retraining the audio prior. To cut memory and stabilize training, we pool video tokens before conditioning. On curated video-audio benchmarks, Foley Control delivers competitive temporal and semantic alignment with far fewer trainable parameters than recent multi-modal systems, while preserving prompt-driven controllability and production-friendly modularity (swap/upgrade encoders or the T2A backbone without end-to-end retraining). Although we focus on Video-to-Foley, the same bridge design can potentially extend to other audio modalities (e.g., speech).
format Preprint
id arxiv_https___arxiv_org_abs_2510_21581
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Foley Control: Aligning a Frozen Latent Text-to-Audio Model to Video
Rowles, Ciara
Jampani, Varun
Donné, Simon
Vainer, Shimon
Parker, Julian
Evans, Zach
Computer Vision and Pattern Recognition
Sound
Foley Control is a lightweight approach to video-guided Foley that keeps pretrained single-modality models frozen and learns only a small cross-attention bridge between them. We connect V-JEPA2 video embeddings to a frozen Stable Audio Open DiT text-to-audio (T2A) model by inserting compact video cross-attention after the model's existing text cross-attention, so prompts set global semantics while video refines timing and local dynamics. The frozen backbones retain strong marginals (video; audio given text) and the bridge learns the audio-video dependency needed for synchronization -- without retraining the audio prior. To cut memory and stabilize training, we pool video tokens before conditioning. On curated video-audio benchmarks, Foley Control delivers competitive temporal and semantic alignment with far fewer trainable parameters than recent multi-modal systems, while preserving prompt-driven controllability and production-friendly modularity (swap/upgrade encoders or the T2A backbone without end-to-end retraining). Although we focus on Video-to-Foley, the same bridge design can potentially extend to other audio modalities (e.g., speech).
title Foley Control: Aligning a Frozen Latent Text-to-Audio Model to Video
topic Computer Vision and Pattern Recognition
Sound
url https://arxiv.org/abs/2510.21581