The SpeechEmotionRecognition.xyz project aims at the formation of an open-sourced, anonymized multilingual dataset of emotional speech. The recordings that are submitted to the site, will be validated, and be provided publically for research non-commercial purposes.

The resulting dataset will be an extension of the Acted Emotional Speech Dynamic Database (AESDD).

The current instance of the AESDD, and the respective documentation can be accessed here