YingMusic-SVC: 实现现实场景下鲁棒的零样本歌声转换的Flow-GRPO与歌唱特定归纳偏置
YingMusic-SVC: Real-World Robust Zero-Shot Singing Voice Conversion with Flow-GRPO and Singing-Specific Inductive Biases
AI总结：
YingMusic-SVC通过融合连续预训练、鲁棒监督微调和Flow-GRPO强化学习，实现了在现实场景下鲁棒的零样本歌声转换。
摘要：
歌声转换（SVC）旨在在保持旋律和歌词的同时渲染目标歌手的音色。然而，现有的零样本SVC系统在现实歌曲中仍然容易受到和声干扰、F0误差以及缺乏歌唱特定的归纳偏置的影响。我们提出了YingMusic-SVC，一种鲁棒的零样本框架，它统一了连续预训练、鲁棒监督微调和Flow-GRPO强化学习。我们的模型引入了一个经过歌唱训练的RVC音色转换器用于音色-内容解耦，一个F0感知的音色适配器用于动态语音表达，以及一个能量平衡的修正流匹配损失以增强高频保真度。在分级多轨基准测试中，实验表明YingMusic-SVC在音色相似性、可懂度和感知自然度方面均优于强大的开源基线，特别是在伴奏和和声污染条件下，证明了其在现实SVC部署中的有效性。
英文摘要：
Singing voice conversion (SVC) aims to render the target singer's timbre while preserving melody and lyrics. However, existing zero-shot SVC systems remain fragile in real songs due to harmony interference, F0 errors, and the lack of inductive biases for singing. We propose YingMusic-SVC, a robust zero-shot framework that unifies continuous pre-training, robust supervised fine-tuning, and Flow-GRPO reinforcement learning. Our model introduces a singing-trained RVC timbre shifter for timbre-content disentanglement, an F0-aware timbre adaptor for dynamic vocal expression, and an energy-balanced rectified flow matching loss to enhance high-frequency fidelity. Experiments on a graded multi-track benchmark show that YingMusic-SVC achieves consistent improvements over strong open-source baselines in timbre similarity, intelligibility, and perceptual naturalness, especially under accompanied and harmony-contaminated conditions, demonstrating its effectiveness for real-world SVC deployment.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
