数据堂—762小时非西班牙人说西班牙语手机采集语音数据

拉美西班牙语语音数据_朗读（手机），基于给定的脚本朗读并模拟录制，涵盖通用、交互、车载、家居等多类别，内容丰富。此数据集标注了文本内容等多种属性，由1630名墨西哥、哥伦比亚等非西班牙国家本土发音人参与录制，口音正宗，文本经过人工校对，准确率高，为语音识别相关研究及应用提供了丰富的资源，经多家AI公司验证：有助于模型面对真实世界的多样性时能够表现出色。我们严格遵循数据保护法规和隐私规定，确保数据采集、存储和使用的过程中维护用户的隐私和合法权益，所有数据均遵循GDPR, CCPA, PIPL

Latin American Spanish Read Speech Dataset (Mobile Recorded): Recorded by reading pre-defined scripts with simulated mobile recording conditions, this dataset covers diverse categories including general scenarios, conversational interaction, in-vehicle and smart home scenarios, with rich content. This dataset is annotated with multiple attributes such as the corresponding text content. It was recorded by 1,630 native speakers from Spanish-speaking countries other than Spain, including Mexico and Colombia, with authentic regional accents. All source texts have undergone manual proofreading, ensuring high accuracy. It serves as a valuable resource for speech recognition-related research and applications. Verified by multiple AI enterprises, this dataset enables models to perform robustly when confronted with real-world diversity. We strictly abide by data protection laws and privacy regulations, safeguarding users' privacy and legitimate rights and interests throughout the entire process of data collection, storage and utilization. All data complies with GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含762小时的西班牙语语音数据，通过移动电话采集自1,630名非西班牙语母语的使用者，录音内容涵盖对话、交互、家庭自动化和车内命令等多种类别。数据采用16kHz、16位、单声道的WAV格式，转录准确率不低于98%，专用于西班牙语语音识别模型的测试任务。