CMU ARCTIC

CMU_ARCTIC数据库是在卡内基梅隆大学语言技术研究所构建的，是语音平衡的美国英语单说话者数据库，旨在进行单元选择语音合成研究。
有关数据库的结构和内容以及记录环境等的详细报告可作为卡内基梅隆大学，语言技术研究所技术报告CMU-LTI-03-177获得，也可在此处获得。
数据库由大约1150种话语组成，这些话语是从古腾堡计划的版权外文本中精心选择的。数据库包括美国英语男性 (bdl) 和女性 (slt) 演讲者 (均具有丰富的语音才能) 以及其他重音演讲者。

The CMU_ARCTIC database, developed at the Language Technologies Institute of Carnegie Mellon University, is a phonetically balanced American English single-speaker corpus designed for unit selection speech synthesis research. Detailed reports covering the structure, content, recording environment and other relevant details of the database are available as Carnegie Mellon University Language Technologies Institute Technical Report CMU-LTI-03-177, and can also be accessed here. The corpus comprises approximately 1150 utterances carefully selected from public-domain texts sourced from Project Gutenberg. It includes American English male (bdl) and female (slt) speakers (both with outstanding speech capabilities) as well as speakers with other regional accents.

搜集汇总

背景与挑战

背景概述

CMU ARCTIC是一个由卡内基梅隆大学构建的语音合成研究数据集，包含约1150条美国英语男性和女性演讲者的录音，适用于单元选择语音合成研究。