This dataset focuses on processing Japanese conversational speech in real-world settings. Designed in a conversation-based style, it captures the interactive and complex nature of everyday communication, thereby enhancing model performance in authentic conversational environments. Recordings were made using mobile devices, a choice that closely mirrors actual usage scenarios and highlights the dataset’s practical relevance. With a total duration of 10 hours, the dataset offers a diverse and realistic collection of conversational speech samples.
This dataset is part of the MAGICDATA-DUPLEX series, a collection of multilingual full-duplex conversational speech datasets designed to support research and development in natural voice AI interaction. It is also known as MAGICDATA-DUPLEX-JAPANESE, following the unified naming convention of the series. The series focuses on authentic human conversations, preserving natural turn-taking, interruptions, overlapping speech, and other interactive features essential for full-duplex speech model training and evaluation.
Sample：
Two-speaker conversation with separate tracks:
The dataset is not for commercial use. The open-source dataset may be used for academic research and must be properly cited with the source.
Citation Format：Japanese Duplex Conversation Training Dataset. 2025. https://magichub.com/datasets/japanese-duplex-conversation-training-dataset/. Beijing Magic Data Technology Co., Ltd.
For more commercial datasets, please contact business@magicdatatech.com.
