46人多风格平均音色合成库

46人多风格平均音色合成库，由专业声优采集，46种风格分别为北京话、电影解说、华妃、纪录片解说、美食解说、小说解说、中青年磁性男、怼人杜飞、孙悟空、温暖男声、TVB 女声、广西表哥、林黛玉、哪吒、如来、申公豹、太乙真人、温柔桃子、徐志胜、周杰伦、河南大叔、气泡音、容嬷嬷、天津小哥、天真幼童、正气男声、咆哮哥、四郎、海夜叉、导购女声、海绵宝宝、唐僧、鸭子嗓、阳光男声、东厂公公、新闻女声、乖巧童声、软糯童声、有声书男、有声书女、GG-bond、坚韧少御、温情大叔、温柔白月光、蛮横大妈、貂蝉。A 46-speaker multi-style average timbre synthesis dataset collected from professional voice actors. The 46 included speaking styles are: Beijing dialect, movie commentary, Imperial Concubine Hua, documentary commentary, food commentary, novel commentary, mid-young-aged magnetic male voice, sharp-tongued Du Fei, Sun Wukong, warm male voice, TVB female voice, Guangxi-accented cousin, Lin Daiyu, Nezha, Tathagata Buddha, Shen Gongbao, Taiyi Zhenren, gentle Taozi, Xu Zhisheng, Jay Chou, Henan-accented uncle, bubble voice, Rong Momo, Tianjin-accented young man, naive child voice, upright male voice, yelling guy, Emperor Yongzheng, Sea Yaksha, shopping guide female voice, SpongeBob SquarePants, Tang Sanzang, duck voice, sunny male voice, Eastern Depot Eunuch, news female voice, well-behaved child voice, soft and sweet child voice, audiobook male voice, audiobook female voice, GG-bond, tenacious young elegant female voice, warm-hearted uncle voice, gentle white moonlight, overbearing aunt, Diaochan.

搜集汇总

背景与挑战

背景概述

46人多风格平均音色合成库是一个专为语音合成设计的中文普通话数据集，由专业声优在专业录音棚环境中录制，包含46种不同风格（如北京话、电影解说等），数据格式为高质量48kHz、24bit的wav单声道音频，并附带文字标注，适用于训练多样化的语音合成模型。