DesignLex
/人机交互/Speech-to-Text (STT)
论文入门

Speech-to-Text (STT)

语音转文本(Speech-to-Text, STT),将口语转换为书面文字的技术,常作为语音用户界面和对话代理的输入环节。

English Definition

"Technology that converts spoken language into written text, often used as the input stage of voice user interfaces and conversational agents."

用法说明 · 针对中文母语者

中文常简称 STT。在 WoZ 实验中常用'巫师'替代 STT,以避免识别错误对实验的污染——本文正是出于此考虑。

真实用例 · 1

"This approach minimized the risk of Speech-to-Text (STT) errors and provided a more realistic experimental environment, preventing degradation in response quality caused by input format or accuracy issues."

论文Pipeline extracted — review needed·说明用巫师替代 STT 以降低识别误差带来的实验污染
由 pipeline 自动采集,待人工 review