数据堂—35小时有声读物文本拼音标注语音数据

35小时有声读物文本拼音标注语音数据由5名发音人参与录制，其中男性 3 人，女性 2 人，对语音内容做汉字和拼音标注，拼音标注声调。35小时有声读物文本拼音标注语音数据可用于语音识别、机器翻译、声纹识别等任务

The 35-hour audiobook speech dataset has its text content annotated with both Chinese characters and pinyin, and the pinyin is marked with tone marks. It was recorded by 5 speakers, including 3 males and 2 females. This dataset can be applied to tasks such as speech recognition, machine translation, speaker verification and other relevant tasks.

搜集汇总

背景与挑战

背景概述

该数据集包含35小时的有声读物语音数据，由5位说话人（3男2女）录制，音频格式为44.1kHz、16bit、wav、单声道，内容涵盖散文、小说和儿童故事等类别，并标注了汉字和带声调的拼音。它适用于自动语音识别、机器翻译和说话人识别等任务，遵循Apache 2.0许可。