Not All Frames Are Equal: Complexity-Aware Masked Motion Generation via Motion Spectral Descriptors
This work addresses a specific bottleneck in text-to-motion synthesis for applications like animation or robotics, offering an incremental improvement by incorporating complexity awareness into existing masked models.
The paper tackled the problem of masked generative models treating motion frames uniformly, which mismatches the varying local dynamic complexity in motion, and introduced the Motion Spectral Descriptor (MSD) to make generation complexity-aware, resulting in improved motion generation on dynamically complex motions and stronger overall FID on HumanML3D and KIT-ML benchmarks.
Masked generative models have become a strong paradigm for text-to-motion synthesis, but they still treat motion frames too uniformly during masking, attention, and decoding. This is a poor match for motion, where local dynamic complexity varies sharply over time. We show that current masked motion generators degrade disproportionately on dynamically complex motions, and that frame-wise generation error is strongly correlated with motion dynamics. Motivated by this mismatch, we introduce the Motion Spectral Descriptor (MSD), a simple and parameter-free measure of local dynamic complexity computed from the short-time spectrum of motion velocity. Unlike learned difficulty predictors, MSD is deterministic, interpretable, and derived directly from the motion signal itself. We use MSD to make masked motion generation complexity-aware. In particular, MSD guides content-focused masking during training, provides a spectral similarity prior for self-attention, and can additionally modulate token-level sampling during iterative decoding. Built on top of masked motion generators, our method, DynMask, improves motion generation most clearly on dynamically complex motions while also yielding stronger overall FID on HumanML3D and KIT-ML. These results suggest that respecting local motion complexity is a useful design principle for masked motion generation. Project page: https://xiangyue-zhang.github.io/DynMask