Skip to content

The compiled instrumental music dataset used in my Master's thesis 2022

Notifications You must be signed in to change notification settings

jonasrodningen/instrumentalMER-data-thesis

Repository files navigation

Instrumental music emotion recognition (MER) dataset

The compiled instrumental music emotion recognition (MER) dataset used in my Master's thesis (2022):

Jonas Klepper Rødningen, Emotion Recognition from Speech and
Instrumental Music: Proof of Shared Emotional
Code Through Transfer Learning, 2022, NTNU

See the tracklist files for the compiled instrumental music emotion recognition dataset (static dimensional valence and arousal values, normalized to [-1,1]): Soundtracks and DEAM. For information about the compilation process, I refer to the thesis.

The chosen segments from EMOPIA, and their Youtube-titles (++) are also included.

Please see the respective original datasets' links above for credits to the authors.

It is recommended to use StratifiedGroupKFold from Sklearn to create train/val/test splits that respects that some clips originate from the same song.

About

The compiled instrumental music dataset used in my Master's thesis 2022

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published