{"paper":{"title":"Shao: Scaling Acoustic Token Language Models Toward High-Fidelity Music Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"High-fidelity music can be generated by progressively modeling structure and detail together inside one 64-layer acoustic token hierarchy.","cross_cats":["cs.AI"],"primary_cat":"cs.SD","authors_text":"Feng Yu, Hongjia Liu, Huijing Liang, Jiafeng Liu, Maosong Sun, Wenbo Zhan, Xiaobing Li, Yuanliang Dong, Yuming Sun, Yuqing Cheng, Zhancheng Guo","submitted_at":"2026-05-03T09:13:20Z","abstract_excerpt":"A common design pattern in high-quality music generation is to handle structure and fidelity in different representation spaces: a generator first models high-level structure, followed by diffusion-based or neural decoding stages that reconstruct fine details. In this work, we explore an alternative view: both may be progressively modeled within a single deep acoustic-token hierarchy. To study this, we build a 64-layer residual vector quantization (RVQ) acoustic representation and propose a two-stage coarse-to-fine generation framework. A backbone model first generates coarse acoustic tokens f"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Taken together, our results suggest that high-quality music generation can be effectively pursued without separating structure and fidelity into heterogeneous representation spaces. Instead, both can be progressively modeled within a unified acoustic-token hierarchy, pointing toward a simpler and more unified path to high-quality music generation.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a single 64-layer RVQ acoustic token space can capture both high-level musical structure and fine-grained fidelity sufficiently to allow emergent alignment and high quality without information loss that would require separate semantic modeling.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A two-stage coarse-to-fine acoustic token language model with hybrid attention generates high-fidelity music and achieves emergent lyric alignment without separate semantic tokens.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"High-fidelity music can be generated by progressively modeling structure and detail together inside one 64-layer acoustic token hierarchy.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"79cdbd1fc42166f36886992f6fcb8b90d3139c0d4bcb53ae57e9441fc34b3b89"},"source":{"id":"2605.01790","kind":"arxiv","version":2},"verdict":{"id":"6ff8e938-02e5-40ed-b318-49fb28c8d039","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T16:32:11.293019Z","strongest_claim":"Taken together, our results suggest that high-quality music generation can be effectively pursued without separating structure and fidelity into heterogeneous representation spaces. Instead, both can be progressively modeled within a unified acoustic-token hierarchy, pointing toward a simpler and more unified path to high-quality music generation.","one_line_summary":"A two-stage coarse-to-fine acoustic token language model with hybrid attention generates high-fidelity music and achieves emergent lyric alignment without separate semantic tokens.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a single 64-layer RVQ acoustic token space can capture both high-level musical structure and fine-grained fidelity sufficiently to allow emergent alignment and high quality without information loss that would require separate semantic modeling.","pith_extraction_headline":"High-fidelity music can be generated by progressively modeling structure and detail together inside one 64-layer acoustic token hierarchy."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01790/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T17:36:02.204905Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-20T05:01:22.725690Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T16:58:40.484608Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"4a14a298611371a5ff6ca02ea171bf38ffacad0428e6e515ccb07605be932254"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}