音乐标记方式比模型大小更影响生成质量，性能级标记显著提升音符生成效果。
Agogic: Performance-Timed Music Tokens for LLM-Native Text-to-Symbolic-Music Generation
- 采用性能级音乐标记（PMT），精确到10毫秒时间与力度，实现高保真音符表达。
- 0.8B模型用PMT时FMD达159，远低于27B模型在节拍网格下的286，差距最大达2.8倍。
- 效果稳定可复现，且不依赖特定模型或数据集，适合关注音乐生成质量的研究者。
文本转音乐的语言模型通常默认使用某种音乐标记方式，但其独立影响从未被测量。我们固定Qwen3.5（0.8B-27B）模型、数据、预算与解码策略，仅替换七种不同标记方式，以各表示的无模型上限为基准。结果清晰且出人意料：表示方式而非模型规模是分布保真度的关键变量。模型扩大34倍，弗雷切特音乐距离（FMD）几乎不变；而更换表示方式，FMD减半。我们提出的性能分辨率标记（PMT，10ms精度，每音符力度，多轨纹理；609个符号）在0.8B模型上达到FMD 159，显著优于节拍网格（FMD 272–286，降低1.7–1.8倍，局部高达2.8倍；置信区间不重叠）。0.8B PMT模型甚至超越27B节拍网格模型。该优势在2600万条从零训练的骨干网络和另一款性能标记中重现，表明属于类别特性。即便将PMT对齐至节拍网格分辨率，仍领先67–129 FMD。该效应为分布层面的，是否可听仍待人类评估验证（已预注册研究）。轻量级解码约束可使乐器识别F1从0.28升至0.60，调性正确率从0.16升至0.35，且无分布代价。我们发布评估框架、25+检查点、两个语料库（共86.6k对齐样本，含歌词/音频/ABC/MIDI；625万标注文本，为最大音乐语料），以及印刻诊断工具：现有文本转MIDI系统在不同输入下仍保持近似训练分布（72%对比71%和弦-时间匹配）。今后的表示主张可被实证衡量，而非仅宣称。
原文摘要 · Abstract (English)
Text-to-music language models begin with a choice usually made by default: how to tokenize music. Normally entangled with backbone, data, and recipe, its effect has never been measured in isolation. We fix pretrained Qwen3.5 (0.8B-27B), data, budget, and decoding, and swap only the representation across seven tokenizations, anchoring texture metrics to each representation's model-free ceiling. The ordering is clean and surprising: representation, not model size, is the binding variable for distributional fidelity. Scaling the backbone 34x barely moves Frechet Music Distance (FMD), whereas switching representation halves it. PMT, a performance-resolution stream we release (10 ms timing, per-note velocity, multi-track texture; 609 symbols), reaches FMD 159 at 0.8B against 272-286 for beat grids (1.7-1.8x lower, up to 2.8x elsewhere; non-overlapping bootstrap CIs), so a 0.8B performance-resolution model beats a 27B beat grid. It reappears on a 26M from-scratch backbone and a second performance-resolution tokenizer: a property of the class, not one lucky vocabulary. Nor is it a finer-lattice artifact: snapping PMT's onsets to the beat grids' resolution still leaves it 67-129 FMD ahead of both (n=500). The effect is distributional; whether it is audible is a separate question, left open by our probe, with a human study pre-registered. Native caption adherence is weak but separable: a lightweight decode-time constraint doubles instrument-F1 (.28 to .60) and Correct-Key (.16 to .35) at no distributional cost. We release the harness, 25+ checkpoints, two corpora (86.6k aligned across caption/MIDI/ABC/audio; 6.25M captioned, the largest for music), and an imprinting diagnostic: published text-to-MIDI systems reproduce their training distribution near-invariant to the caption (72% vs. 71% chord-time on disjoint domains). The field's next representation claim can now be measured, not asserted.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。