ReCANVo: A Dataset of Real-World Communicative and Affective Nonverbal Vocalizations

A dataset of 7077 labeled vocalizations made by non-speaking individuals. Each vocalization lasts approximately 0.5-4 seconds and is labeled with its affective or communicative meaning. Data were acquired in real-world settings (homes, schools, etc.) and were labeled in real-time by parents or caregivers who knew the non-speaking communicator well. 

dataset_file_directory.csv provides the name of each vocalization file, the corresponding participant ID, and the vocalization meaning or label (delighted, frustrated, request, etc.).

If you use this dataset, please cite Johnson & Narain et al., "ReCANVo: A Database of Real-World Communicative and Affective Nonverbal Vocalizations". The authors are Jaya Narain, Kristina T. Johnson, Thomas Quatieri, Pattie Maes, and Rosalind Picard. This paper provides more information about the dataset, including data acquisition methodology, pre-processing procedures, and participant demographics. 

**J.N. and K.T.J. are joint first authors on this project. Please include both names in attribution when possible (e.g., Johnson & Narain et al.).