Metraipolis › archive › dc_b265a930dcad

StepAudio 3 Music Technical Report — arXiv:2609.16034

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.16034
License (as archived): CC0-1.0
Archived: 2026-10-15T22:39:59.538Z
SHA-256 of the archived text: d3d78be53cf233997080d3e6b789112c013fb8d7b344a05b3594ed22780734f9
Stable URL of this page: https://metraipolis.com/documento/dc_b265a930dcad

How to cite

StepAudio 3 Music Technical Report — arXiv:2609.16034 — arXiv (en), https://arxiv.org/abs/2609.16034 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b265a930dcad (Metraipolis archive dc_b265a930dcad, sha256 d3d78be53cf233997080d3e6b789112c013fb8d7b344a05b3594ed22780734f9)

BibTeX

@misc{dcb265a930dcad,
  title = {StepAudio 3 Music Technical Report — arXiv:2609.16034},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.16034},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b265a930dcad; sha256 d3d78be53cf233997080d3e6b789112c013fb8d7b344a05b3594ed22780734f9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.16034. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We introduce StepAudio 3 Music, a large-scale, long-form music generation model that supports explicit musical planning and open-domain text-controlled generation. The StepAudio Music Tokenizer represents audio as a 50-Hz stream from a 65536-entry single codebook, using semantically informed self-supervised and multi-task training to preserve musical structure and reconstruction-relevant information. A flow-matching diffusion Transformer (DiT) predicts continuous StepAudio VAE latents, which our VAE decoder converts into 48-kHz audio. This discrete-continuous design is guided by comparisons of single-codebook VQ, Semantic and Acoustic RVQ, and different DiT configurations. For explicit planning, a Mixture-of-Experts autoregressive model uses ABC notation to produce an intermediate arrangement plan (ABC-CoT) before predicting music tokens, making harmony, rhythm, and melodic structure part of the generation context. A progressive training curriculum and supervised fine-tuning support song and instrumental generation, accompaniment generation from dry vocals, and cover-song synthesis for up to 5 minutes and 30 seconds. With reinforcement learning via direct preference optimization (DPO), the final model achieves the highest AudioBox Content Enjoyment, Content Usefulness, and Production Quality scores and the highest MuQ-MuLan similarity among the evaluated systems, with competitive SongBench results. On the preliminary Artificial Analysis Music Arena Vocals leaderboard, it obtains a Quality Elo of 1105, behind only Suno V5.5 and Mureka and ahead of Suno V5, MiniMax models, and other systems. Audio demonstrations are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-16 in eess.AS).
Categorie arXiv: Audio and Speech Processing (eess.AS) ; Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2609.16034 — 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 2026-10-15T22:18:00.483Z dalla pagina https://arxiv.org/catchup/eess.AS/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).