Skip to content

Dataset · birgermoell

RAVDESS emotional speech and song audio dataset

birgermoell/ravdess

Classic dataset of audio recordings with various speech and vocal emotions for emotion recognition.

What's inside

Audio version of the well-known Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) multimodal corpus created by Steven Livingstone and Frank Russo. Contains 1,000 to 10,000 audio recordings of emotional speech and singing. Distributed under the Creative Commons Attribution license and accompanied by a scientific publication in PLoS ONE.

What you can do with it

  • Training speech emotion recognition (SER) models
  • Fine-tuning audio encoders for vocal and speech intonation classification
  • Comparative testing of paralinguistic speech analysis algorithms
Size
1K–10K rows
Downloads this month
22.5K
Likes
0

How to load it

from datasets import load_dataset

ds = load_dataset("birgermoell/ravdess")