Diffusion-based lip synchronization models achieve strong visual quality and audio-visual alignment, but full-sequence bidirectional attention and many denoising steps make them impractical for real-time inference. We present Lip Forcing, to our knowledge the first autoregressive diffusion method for video-to-video (V2V) lip synchronization, which distills a 14B audio-conditioned bidirectional video diffusion teacher into causal students. At inference, the students generate each chunk in only two denoising steps without inference-time CFG, enabling real-time lip synchronization. A lip-sync-specific teacher-trajectory analysis reveals a CFG fidelity-sync tradeoff: no-CFG predictions favor reference fidelity, whereas CFG-guided predictions favor synchronization within a mid-trajectory band. Lip Forcing translates this finding into three analysis-derived components: Sync-Window DMD, a two-step inference schedule, and a SyncNet-based reward. We validate Lip Forcing at two student scales, both distilled from the 14B teacher. The 1.3B student crosses into real-time streaming at 31 FPS, $17.6\times$ faster than its same-scale bidirectional model. The 14B student, the largest diffusion model reported for V2V lip synchronization, runs $39.8\times$ faster than its teacher at comparable reference fidelity. Time-to-first-frame is sub-millisecond at both scales, far below every diffusion baseline.

翻译：基于扩散的唇形同步模型虽可实现优异的视觉质量与音画对齐，但全序列双向注意力机制及大量去噪步骤使其难以满足实时推理需求。本文提出Lip Forcing——据我们所知首个面向视频到视频（V2V）唇形同步的自回归扩散方法，该方法将140亿参数的音频条件双向视频扩散教师模型蒸馏为因果学生模型。推理阶段，学生模型无需推理时无分类器引导（CFG），仅需两步去噪即可生成每个分块，从而实现实时唇形同步。针对唇形同步特有的教师轨迹分析揭示了CFG保真度-同步性权衡：无CFG预测倾向于参考保真度，而CFG引导预测则在轨迹中段频带内更侧重同步性。Lip Forcing将这一发现转化为三个基于分析推导的组件：同步窗口分布匹配蒸馏（Sync-Window DMD）、两步推理调度及基于SyncNet的奖励函数。我们在两个学生模型规模上验证了Lip Forcing（均从140亿参数教师模型蒸馏得到）：13亿参数学生模型以31 FPS实现实时流式处理，比同规模双向模型快17.6倍；140亿参数学生模型（为V2V唇形同步报告的最大扩散模型）在可比参考保真度下运行速度比教师模型快39.8倍。两种规模的首帧延迟均低于亚毫秒级，远低于所有扩散基线方法。