摘要:
在content encoder中使用instance normalization,可以起到去除说话者信息的作用。首先来看一下instance normalization的原理,一般会对输入语音做conv1d得到feature map,有几个conv1d filter就会得到几个feature ma 阅读全文
摘要:
原文:FREEVC: TOWARDS HIGH-QUALITY TEXT-FREE ONE-SHOT VOICE CONVERSION 原文地址:https://ieeexplore.ieee.org/abstract/document/10095191 个人总结: 1.提出mel谱缩放增强方法。 阅读全文
摘要:
原文:PARALLEL-DATA-FREE VOICE CONVERSION USING CYCLE-CONSISTENT ADVERSARIAL NETWORKS 地址:https://arxiv.org/pdf/1711.11293v2.pdf 摘要: 我们提出了一种无需平行数据的语音转换(VC 阅读全文
摘要:
论文原文:USING JOINT TRAINING SPEAKER ENCODER WITH CONSISTENCY LOSS TO ACHIEVE CROSS-LINGUAL VOICE CONVERSION AND EXPRESSIVE VOICE CONVERSION 代码和论文地址:http 阅读全文