SMS-WSJ (Spatialized Multi-Speaker Wall Street Journal)

Spatialized Multi-Speaker Wall Street Journal (SMS-WSJ) 由从 WSJ 数据库中提取的人工混合语音组成，但与早期的数据库不同，该数据库考虑了所有 WSJ0+1 话语，并严格区分训练、验证中存在的说话者集和测试集。

Spatialized Multi-Speaker Wall Street Journal (SMS-WSJ) consists of artificially mixed speech extracted from the WSJ database. Unlike earlier databases, this dataset covers all utterances in WSJ0+1, and strictly distinguishes the speaker sets present in the training and validation splits from those in the test set.

搜集汇总

背景与挑战

背景概述

SMS-WSJ是一个公开的语音数据集，专注于空间化多说话人场景，大小为941.0B，包含901个文件。它适用于预训练、AIGC、语音识别和音源分离任务，并采用MIT许可证，便于开源使用。