Generating realistic conversational gestures are essential for achieving natural, socially engaging interactions with digital humans. However, existing methods typically map a single audio stream to a single speaker's motion, without considering social context or modeling the mutual dynamics between two people engaging in conversation. We present DyaDiT, a multi-modal diffusion transformer that generates contextually appropriate human motion from dyadic audio signals. Trained on Seamless Interaction Dataset, DyaDiT takes dyadic audio with optional social-context tokens to produce context-appropriate motion. It fuses information from both speakers to capture interaction dynamics, uses a motion dictionary to encode motion priors, and can optionally utilize the conversational partner's gestures to produce more responsive motion. We evaluate DyaDiT on standard motion generation metrics and conduct quantitative user studies, demonstrating that it not only surpasses existing methods on objective metrics but is also strongly preferred by users, highlighting its robustness and socially favorable motion generation. Code and models will be released upon acceptance.
Figures & tables
Figure 2: Overview of DyaDiT. DyaDiT conditions on multiple input modalities, including audio, partner motion, relationship type, and personality scores. It employs an Audio Orthogonalization Cross Attention (ORCA) module to obtain cleaner audio representations and a motion dictionary to guide style aware gesture generation.
Figure 3: ORCA reduces ambiguity between the two audio streams, allowing DyaDiT to generate realistic motion even when one person interrupts the other during the conversation. The example demonstrates the generated motions adjusts naturally as the conversation shifts.
Method
BC
FD
Diversity
( ×10−1 )
Sta. ↓
Kine. ↓
Sta. ↑
Kine. ↑
GT
-
-
-
28.42
1.97
GT Random
-
14.94
3.74
33.85
2.05
ConvoFusion
-
9.22
1.74
18.33
1.10
Audio2PhotoReal
-
8.77
1.84
19.35
1.05
DyaDiT (w/o ORCA)
7.25
7.32
1.79
23.57
1.24
Table 1: Quantitative comparison of DyaDiT and baselines in terms of Fréchet Distance (FD) and Diversity. Lower FD indicates higher realism, and higher diversity values indicate more varied motion generation.
Figure 4: Qualitative Results. Comparison of visualization results between DyaDiT , ConvoFusion [ 31 ] , and Audio2PhotoReal [ 3 ] . The gestures generated by DyaDiT exhibit higher diversity and greater realism compared to the other methods.
Figure 5: Visualization results under different personality score conditionings. All samples are generated using classifier-free guidance with CFG = 2.5.
Figure 6: Example video pairs used in the user study for evaluating participant preference in conversational gesture generation.
Figure 7: A/B subjective evaluation percentages comparing our method with ConvoFusion [ 31 ] and with ground truth. Participants preferred our generated motion due to its more natural and socially aware conversational behavior.
Dyadic conversational motion generation is essential for realistic interactive digital humans. Existing approaches typically model conversational behaviors within unified dyadic generators. However, such holistic formulations tend to couple self-speech-driven motion with partner-responsive social feedback, leaving the interaction-specific component implicit and underutilizing the speech-motion correspondence already learned by pretrained monologic motion models. We propose Learn2Chat, a unified framework that models dyadic motion as interaction modulation over pretrained monologic motion priors. This design separates intrinsic speech-driven motion from social interaction effects and enables more structured interaction modeling. Specifically, we introduce a Monologic-Anchored Motion Factorization scheme that leverages the semantic motion manifold learned from monologic data to disentangle audio-driven motion dynamics from interaction-induced modulation, yielding clean interaction representations from dyadic sequences. On top of this representation space, a Cross-Attentive Interaction Latent Prediction module maps paired speech signals to interaction latents through cross-branch attention and interaction alignment. During inference, the predicted interaction latents modulate canonical monologic motion to generate coherent and synchronized dyadic behaviors in a data-efficient manner. Extensive experiments on the DualTalk benchmark demonstrate that Learn2Chat achieves state-of-the-art performance across both quantitative metrics and perceptual evaluations. Moreover, the framework is model-agnostic and seamlessly integrates with diverse pretrained monologic motion backbones, highlighting the effectiveness of prior reuse and interaction adaptation for scalable conversational motion generation. More visual results are available on the project page.
Zikai Huang, Siyue Chen, Xuemiao Xu +4
School of Computer Science and Engineering, South China University of Technology, Guangdong, China · School of Design, South China University of Technology, Guangdong, China · Guangdong Engineering Center for Large Model and GenAI Technology +2
Co-speech gesture generation in dyadic conversations for creating immersive virtual avatars remains highly challenging. Existing models often struggle to balance kinematic stability with rhythmic responsiveness and suffer from semantic entanglement, leading to either generic mean-pose regression or unnatural micro-jitters. To address these challenges, we propose SynerGest, a high-fidelity dyadic gesture generation framework grounded in Contrastive Flow Matching (CFM). Specifically, our CFM objective explicitly penalizes mismatched semantic pairs during training, discouraging spatial shortcuts and over-smoothed predictions while improving conditional semantic alignment. Furthermore, to overcome architectural bottlenecks, we introduce the Impulse-Trend Synergistic Attention (ITSA) module. Grounded in the key insight of decoupling "what moves" from "when it moves," ITSA models sharp, speech-driven rhythmic impulses separately from continuous kinematic trends, supporting precise audio-motion synchronization and global structural stability. Extensive qualitative and quantitative evaluations on two public datasets demonstrate that SynerGest achieves state-of-the-art performance, delivering contextually responsive and kinematically fluid dyadic interactions.
Chengxiang Guo, Jian Wang, Lanmiao Liu +3
Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming, China · Yunnan Key Laboratory of Artificial Intelligence · Utrecht University Utrecht, Netherlands +2
We present DyaPlex, a streaming, full-duplex speech-and-motion model designed for dyadic interaction. To capture the continuous and reciprocal nature of human communication, this full-duplex capability empowers the agent to simultaneously perceive and generate both speech and physical motion in a streaming fashion. At its core, our method leverages the strong priors of a foundational full-duplex speech model and integrates a novel motion pathway, thereby achieving fully synchronized multi-modal interaction. Specifically, we design a dual-tower Transformer architecture that preserves the zero-shot conversational reasoning of a frozen base speech model while constructing a deeply coupled, streaming motion pathway. By introducing a unified dyadic token interleaving mechanism and guiding cross-attention via a time-aligned speech-motion RoPE, our model effectively aligns autoregressive motions with rich latent speech features. Trained on the 4,000-hour Seamless Interaction dataset, our model effectively captures cross-speaker dependencies and establishes new state-of-the-art performance across both monadic and dyadic human interaction benchmarks.