This dataset contains recorded multi-speaker conversational audio in English, featuring natural dialogue among multiple participants with overlapping speech and background noise.

The recording was created as part of an undergraduate research project focused on audio forensics, multi-speaker separation, and speech enhancement.

The dataset is released under the MIT License and may be freely used, modified, and shared for research and educational purposes, provided that proper attribution is given to the creator.

You can cite this dataset using the following BibTeX entry:

  
@misc{mudasiru_2025, 
     title={{Multi-Speaker Separation and Background Speech Enhancement with Deep Learning}}, 
     author={Mudasiru, Rasheed}, 
     year={2025}, 
     howpublished={Undergraduate Project Dataset}, 
     institution={Federal University of Technology Minna}, 
     license={MIT}, 
     note={63-minute multi-speaker English conversation with background noise for speech processing research.} 
}