Search papers, labs, and topics across Lattice.
This paper introduces OmniVAE, a novel approach for jointly generating synchronized audio and video by leveraging a variational autoencoder (VAE) architecture that aligns their latent representations. By employing a segment-level audio-video contrastive objective alongside distillation from pretrained semantic encoders, OmniVAE achieves fine-grained cross-modal correspondence, enhancing the quality of generated content. Experimental results demonstrate that this method significantly improves the learnability of latent spaces, leading to superior audio-video synchronization in downstream applications like text-to-audio-video generation.
Fine-grained cross-modal alignment in audio-video generation can dramatically enhance synchronization and quality, as shown by OmniVAE's innovative training approach.
Recent generative models are moving beyond silent video or standalone audio synthesis toward the joint generation of synchronized audio and video. Despite this progress, jointly generating audio and video with fine-grained cross-modal correspondence remains challenging due to their fundamental structural differences. Most existing methods use audio and video VAEs trained separately. As a result, the two latent spaces lack cross-modal alignment, leaving the downstream generative model to learn cross-modal synchronization from scratch. We present OmniVAE, a jointly trained audio-video VAE that learns fine-grained semantic alignment between audio and video latent representations. Beyond reconstruction, OmniVAE uses a segment-level audio-video contrastive objective to capture temporal-semantic correspondence and align the two latent spaces. In parallel, it distills features from pretrained modality-specific semantic encoders into each modality, improving the downstream learnability of both latent spaces. Extensive experiments show that both objectives consistently improve the learnability of the latent spaces, translating into higher generation quality and more accurate cross-modal synchronization in downstream text-to-audio-video generation. These findings underscore the importance of learning unified representations as a foundation for omnimodal modeling.1
