TAVGBench: Benchmarking Text to Audible-Video Generation
Abstract: The Text to Audible-Video Generation (TAVG) task involves generating videos with accompanying audio based on text descriptions. Achieving this requires skillful alignment of both audio and video elements. To support research in this field, we have developed a comprehensive Text to Audible-Video Generation Benchmark (TAVGBench), which contains over 1.7 million clips with a total duration of 11.8 thousand hours.
We propose an automatic annotation pipeline to ensure each audible video has detailed descriptions for both its audio and video contents.
We also introduce the Audio-Visual Harmoni score (AVHScore) to provide a quantitative measure of the alignment between the generated audio and video modalities.
Additionally, we present a baseline model for TAVG called TAVDiffusion, which uses a two-stream latent diffusion model to provide a fundamental starting point for further research in this area.
We achieve the alignment of audio and video by employing cross-attention and contrastive learning.
Through extensive experiments and evaluations on TAVGBench, we demonstrate the effectiveness of our proposed model under both conventional metrics and our proposed metrics.
Primary Subject Area: [Generation] Generative Multimedia
Secondary Subject Area: [Experience] Multimedia Applications
Relevance To Conference: Audio and video are pivotal forms of multimedia expression. This article concentrates on the generation of multimedia and multimodal content, which holds significant relevance to ACM Multimedia (ACM MM) and contributes directly to the field of multimedia and multimodal processing. Specifically, we address the challenge of multimodal generation by introducing an extensive dataset, an evaluation metric, and a robust baseline. The effectiveness of these contributions has been comprehensively validated through experimental results.
Supplementary Material: zip
Submission Number: 1240
Loading