VMAS: Video-to-Music Generation via Semantic Alignment in Web Music Videos

TL;DR

This paper introduces VMAS, a novel framework that generates background music from videos by leveraging large-scale web data, semantic alignment, and a new video encoder, resulting in more realistic and diverse music.

Contribution

The paper presents a new video-music Transformer with semantic and beat alignment, and a large dataset for training, advancing the quality and diversity of video-to-music generation.

Findings

Outperforms existing methods on DISCO-MV and MusicCaps datasets.

Uses a large-scale dataset of 2.2 million video-music samples.

Achieves higher human evaluation scores for music realism and relevance.

Abstract

We present a framework for learning to generate background music from video inputs. Unlike existing works that rely on symbolic musical annotations, which are limited in quantity and diversity, our method leverages large-scale web videos accompanied by background music. This enables our model to learn to generate realistic and diverse music. To accomplish this goal, we develop a generative video-music Transformer with a novel semantic video-music alignment scheme. Our model uses a joint autoregressive and contrastive learning objective, which encourages the generation of music aligned with high-level video content. We also introduce a novel video-beat alignment scheme to match the generated music beats with the low-level motions in the video. Lastly, to capture fine-grained visual cues in a video needed for realistic background music generation, we introduce a new temporal video encoder architecture, allowing us to efficiently process videos consisting of many densely sampled frames. We train our framework on our newly curated DISCO-MV dataset, consisting of 2.2M video-music samples, which is orders of magnitude larger than any prior datasets used for video music generation. Our method outperforms existing approaches on the DISCO-MV and MusicCaps datasets according to various music generation evaluation metrics, including human evaluation. Results are available at https://genjib.github.io/project_page/VMAs/index.html