AI 中文总结
该研究针对基于样例的复频谱分离方法，发现其选择角色失效的问题，提出用更简单的选择类修复，在 MUSDB18 上降低了准则距离和延迟。
AI 中文摘要
样例方法通过为每个源选择一个学习到的频谱并对其进行变形以解释观测结果，从而分离混合信号，这使得一个变形类同时充当重构器和选择器。我们证明，一旦该类能够进行插值，第二个角色（选择器）就是空的：此时该规则会基于其正则化项对候选进行排序，而这一选择是在数据之前做出的，无论哪个候选胜出，估计值都会重新组合成混合信号。该条件是参数数量，因此诊断在任何实验之前就可以进行。对于复频谱的逐频带自由变形，这立刻解释了观察到的病态现象：一个根据响度对候选进行排序的准则，以及一半输出是混合信号上的掩码而非样例。同一定理规定了修复方法：选择类比重构类更简单，用一个复增益和一个纯延迟对候选排序，通过联合闭式拟合的最佳对齐原子的局部组合来重构它们。在 MUSDB18 上，与掩码类的精确上限相比，准则在其自身候选池中的距离从刚性选择器的 6.2-7.7 dB 降至 0.5-2.8 dB，尽管只有 0.9-1.2 dB 到达输出，且部署规则的逐帧延迟降低了 47 至 806 倍。仍存在一个可量化的瓶颈：原子会针对混合信号评分，因此评分带有其他源的项，吸收了重构类获得的容量，使输出比上限低 10.0 dB。对足够自由以进行插值的拟合残差进行假设排序，仅基于正则化项排序。
英文摘要
Exemplar methods separate a mixture by picking one learned spectrum per source and deforming it until it explains the observation, making one deformation class both reconstructor and selector. We show that the second role is empty as soon as the class can interpolate: the rule then ranks candidates on its regulariser, a choice made before the data, and the estimates sum back to the mixture whichever candidate wins. The condition is a parameter count, so the diagnosis runs before any experiment. On free per-bin deformation of complex spectra it explains the observed pathologies at once: a criterion that ranks candidates by their loudness, and half an output that is a mask on the mixture rather than an exemplar. The same theorem prescribes the repair, a selection class poorer than the reconstruction class: one complex gain and one pure delay rank the candidates, and a local combination of the best-aligned atoms, fitted jointly in closed form, rebuilds them. On MUSDB18 against the exact ceiling of the masking class, the distance between the criterion and an oracle inside its own candidate pool falls under the rigid selector from 6.2-7.7 to 0.5-2.8 dB, though only 0.9-1.2 dB of that reaches the output, and the per-frame latency of the deployed rule by a factor of 47 to 806. One lock remains, quantified: atoms are scored against the mixture, so the score carries a term for the other source that absorbs the capacity the reconstruction class gains, leaving the output 10.0 dB under the ceiling. Ranking hypotheses by the residual of a fit free enough to interpolate ranks them on the regulariser alone.
CommentsSubmitted to IEEE/ACM Transactions on Audio, Speech and Language Processing, manuscript T-ASL-13397-2026. 11 pages, 4 figures. Companion paper: "Geometric Ceilings on Time-Frequency Masking" (arXiv:2609.03481)
