Temporally Aligned Audio for Video with Autoregression
This addresses the challenge of precise audio-visual synchronization in video generation, which is important for applications like multimedia and virtual reality, but it appears incremental as it builds on existing video-to-audio generation methods.
The paper tackles the problem of generating temporally aligned audio for video by introducing V-AURA, an autoregressive model that achieves high temporal alignment and relevance, outperforming state-of-the-art models in these aspects while maintaining comparable audio quality.
We introduce V-AURA, the first autoregressive model to achieve high temporal alignment and relevance in video-to-audio generation. V-AURA uses a high-framerate visual feature extractor and a cross-modal audio-visual feature fusion strategy to capture fine-grained visual motion events and ensure precise temporal alignment. Additionally, we propose VisualSound, a benchmark dataset with high audio-visual relevance. VisualSound is based on VGGSound, a video dataset consisting of in-the-wild samples extracted from YouTube. During the curation, we remove samples where auditory events are not aligned with the visual ones. V-AURA outperforms current state-of-the-art models in temporal alignment and semantic relevance while maintaining comparable audio quality. Code, samples, VisualSound and models are available at https://v-aura.notion.site