Joint audio-video generation models have shown that unified generation yields stronger cross-modal coherence than cascaded approaches. However, existing models couple modalities throughout denoising via pervasive attention, treating high-level semantics and low-level details in a fully entangled manner. This is suboptimal for talking head synthesis: while audio and facial motion are semantically correlated, their low-level realizations (acoustic signals and visual textures) follow distinct rendering processes. Enforcing joint modeling across all levels causes unnecessary entanglement and reduces efficiency. We propose Talker-T2AV, an autoregressive diffusion framework where high-level cross-modal modeling occurs in a shared backbone, while low-level refinement uses modality-specific decoders. A shared autoregressive language model jointly reasons over audio and video in a unified patch-level token space. Two lightweight diffusion transformer heads decode the hidden states into frame-level audio and video latents. Experiments on talking portrait benchmarks show Talker-T2AV outperforms dual-branch baselines in lip-sync accuracy, video quality, and audio quality, achieving stronger cross-modal consistency than cascaded pipelines.
@article{arxiv.2604.23586,
title = {Talker-T2AV: Joint Talking Audio-Video Generation with Autoregressive Diffusion Modeling},
author = {Zhen Ye and Xu Tan and Aoxiong Yin and Hongzhan Lin and Guangyan Zhang and Peiwen Sun and Yiming Li and Chi-Min Chan and Wei Ye and Shikun Zhang and Wei Xue},
journal= {arXiv preprint arXiv:2604.23586},
year = {2026}
}