Metraipolis › archive › dc_5db6d747c4fc

FFMT: Frequency-Factorized Multi-Rate Tokenization for Music-to-Dance Generation — arXiv:2611.07384

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.07384
License (as archived)
CC0-1.0
Archived
2026-11-06T14:53:50.583Z
SHA-256 of the archived text
78fee9111e3ba0aa11703e0a920f61aef74ab897111f4298656f40959132a0b2
Stable URL of this page
https://metraipolis.com/documento/dc_5db6d747c4fc

How to cite

FFMT: Frequency-Factorized Multi-Rate Tokenization for Music-to-Dance Generation — arXiv:2611.07384 — arXiv (en), https://arxiv.org/abs/2611.07384 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5db6d747c4fc (Metraipolis archive dc_5db6d747c4fc, sha256 78fee9111e3ba0aa11703e0a920f61aef74ab897111f4298656f40959132a0b2)
BibTeX
@misc{dc5db6d747c4fc,
  title = {FFMT: Frequency-Factorized Multi-Rate Tokenization for Music-to-Dance Generation — arXiv:2611.07384},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.07384},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5db6d747c4fc; sha256 78fee9111e3ba0aa11703e0a920f61aef74ab897111f4298656f40959132a0b2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07384. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Dance motion combines slowly varying body configurations with rapid local movements. We investigate how explicit temporal-frequency decomposition can guide the temporal resolution and codebook capacity of discrete motion representations. We introduce frequency-factorized motion tokenization (FFMT), which separates root-relative motion into a low-frequency choreographic scaffold and a complementary high-frequency residual. FFMT uses a lower token rate and a smaller codebook for the low-frequency component, and a higher token rate and a larger codebook for the high-frequency residual. This representation further induces a plan-then-articulate autoregressive factorization: low-frequency tokens establish coarse choreography, while high-frequency tokens realize local motion details conditioned on the generated plan and music. Experiments on AIST++ and FineDance show strong distributional fidelity while preserving realistic diversity and music--motion synchronization. Ablations consistently support frequency factorization, asymmetric temporal allocation, and low-to-high conditioning.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Sound (cs.SD) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2611.07384 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-06T14:25:42.632Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).