数据堂—201小时北美英语语音数据（手机+电脑）

201小时北美英语语音数据通过采集来自北美地区的302名发音人，录音内容包含短语和句子，场景丰富，贴近生活，有效数据时长201小时。录音环境来自安静的室内，录音设备均衡，包含PC、安卓手机、苹果手机。本套北美英语语音数据适用于北美英语语音识别的相关研发

This North American English speech dataset has an effective duration of 201 hours. It was collected from 302 speakers in North America. The recorded content includes phrases and sentences, with diverse real-life scenarios. All recordings were conducted in quiet indoor environments, using a balanced set of recording devices including personal computers (PCs), Android smartphones and iPhones. This dataset is suitable for research and development related to North American English speech recognition.

搜集汇总

背景与挑战

背景概述

该数据集包含201小时的北美英语语音数据，由302名说话者使用PC、安卓手机和iPhone在安静室内环境录制，覆盖多种真实场景的英语单词和句子。数据格式为16kHz/44.1kHz、16位未压缩WAV单声道音频，适用于北美英语语音识别模型的研发与测试，且为商用数据集。