Abstract
Generating realistic and expressive audio-driven talking avatars remains a challenge in digital human synthesis. Existing methods often depend on intermediate representations for natural body motion, restricting flexibility and leading to visual distortions. Moreover, many approaches rely on discrete emotion labels to regulate expression. Such categorical supervision fails to capture the continuous and fine-grained speech dynamics (rhythm, energy, intensity), resulting in limited synchronization and emotionally shallow motion. To overcome these limitations, we present SyncDreamer, a unified diffusion Transformer framework that generates identity-preserving and emotionally expressive talking avatars from only a single image , speech audio, and text prompt. We propose a visual adapter with Attention Localization Loss to maintain identity fidelity, further incorporating an audio dynamics encoder for rhythm-and emotion-aware motion, and an RL-based Cross-Modal Prompt Enhancer grounding tex-tual cues in visual context for fine-grained motion control. Extensive experiments on portrait and full-body benchmarks demonstrate state-of-the-art performance in realism , synchronization accuracy, and semantic controllabil-ity, establishing a scalable foundation for expressive digital avatars in interactive and creative applications. https: //fnazarieh.github.io/SyncDreamerWeb/