Dataset Open Access
The Variably Intense Vocalizations of Affect and Emotion Corpus (VIVAE) consists of a set of human non-speech emotion vocalizations. The full set, comprising 1085 audio files, features eleven speakers expressing three positive (achievement/ triumph, sexual pleasure, and surprise) and three negative (anger, fear, physical pain) affective states, each parametrically varied from low to peak emotion intensity. The smaller core set of 480 files represents a fully crossed subsample of the full set (6 emotions x 4 intensities x 10 speakers x 2 items) selected based on judged authenticity.
Construction and validation
Recordings took place at the Berklee College of Music in Boston MA, USA. The VIVAE full set contains 1085 files (i.e., all recordings that passed an acoustic quality check and were free of linguistic content). The core set was selected based on authenticity judgements collected throughout the technical validation of the dataset. The technical validation comprises data from 30 individuals providing emotion, intensity, and authenticity judgements. Details on the construction and perceptual evaluation will be shared in an academic paper. Please contact the authors for any questions at firstname.lastname@example.org.
The recordings are digitized at a 44.1-kHz sampling rate and 16-bit resolution.
Each of the 1085 VIVAE files has a unique filename following the file naming convention: [Speaker_Emotion_Intensity_Item-ID.wav]. The filename consists of a 4-part identifier (e.g., S04_surprise_peak_10.wav) defining the stimulus characteristics:
Please cite the VIVAE if it is used in your work in any form.
The VIVAE is released under a Creative Commons Attribution-NonCommercial 4.0 International License, CC BY-NC 4.0.
For VIVAE usage inquiries not specified or accredited under the CC BY-NC 4.0 license, please contact us at email@example.com to request permission.