CMU ARCTIC

该数据集名为CMU ARCTIC，包含了18位发言人的录音，每位发言人贡献了500至1000个语音样本。因此，语音样本的总规模在9000至18000个之间，具体数量取决于每位发言人的贡献。该数据集适用于自动说话人验证（ASV）任务。The dataset named CMU ARCTIC consists of audio recordings from 18 speakers, where each speaker contributes between 500 and 1000 speech samples. As a result, the total scale of speech samples falls within the range of 9,000 to 18,000, with the specific quantity depending on the contributions of each individual speaker. This dataset is applicable to automatic speaker verification (ASV) tasks.

搜集汇总

背景与挑战

背景概述

CMU_ARCTIC是一个用于单元选择语音合成研究的语音平衡英语单说话人数据库，包含约1150个语句，支持多种口音和性别的说话人，提供16kHz波形和EGG信号，并集成Festival语音合成系统。