Not All Frames Are Equal: Complexity-Aware Masked Motion Generation via Motion Spectral Descriptors

TL;DR

This paper introduces the Motion Spectral Descriptor (MSD), a measure of local motion complexity, to improve masked motion generation by making it complexity-aware, leading to better synthesis of dynamic motions.

Contribution

The paper proposes MSD, a deterministic spectral measure of motion complexity, and integrates it into masked motion generation to enhance performance on complex motions.

Findings

MSD correlates strongly with motion dynamics.

DynMask improves generation on complex motions.

Overall FID scores are strengthened on benchmarks.

Abstract

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