C. Baume, G. Fazekas, M. Barthet, D. Marston, and M. Sandler, “Selection of Audio Features for Music Emotion Recognition Using Production Music,” in Proc. AES Conference: 53rd International Conference: Semantic Audio, Jan. 2014, Paper P1-3. [Online]. Available: https://aes.org/publications/elibrary-page/?id=17110
Baume C, Fazekas G, Barthet M, Marston D, Sandler M. Selection of Audio Features for Music Emotion Recognition Using Production Music. In: AES Conference: 53rd International Conference: Semantic Audio. Audio Engineering Society; 2014. Paper P1-3. Available from: https://aes.org/publications/elibrary-page/?id=17110
@inproceedings{Baume2014_17110,
author = {Baume, Chris and Fazekas, György and Barthet, Mathieu and Marston, David and Sandler, Mark},
title = {{Selection of Audio Features for Music Emotion Recognition Using Production Music}},
booktitle = {AES Conference: 53rd International Conference: Semantic Audio},
note = {Paper P1-3},
year = {2014},
month = jan,
publisher = {Audio Engineering Society},
url = {https://aes.org/publications/elibrary-page/?id=17110}
}
TY - CPAPER
TI - Selection of Audio Features for Music Emotion Recognition Using Production Music
AU - Baume, Chris
AU - Fazekas, György
AU - Barthet, Mathieu
AU - Marston, David
AU - Sandler, Mark
T2 - AES Conference: 53rd International Conference: Semantic Audio
M1 - Paper P1-3
PY - 2014
DA - 2014/01/06
UR - https://aes.org/publications/elibrary-page/?id=17110
PB - Audio Engineering Society
LA - en
AB - Music emotion recognition typically attempts to map audio features from music to a mood representation using machine learning techniques. In addition to having a good dataset, the key to a successful system is choosing the right inputs and outputs. Often, the inputs are based on a set of audio features extracted from a single software library, which may not be the most suitable combination. This paper describes how 47 different types of audio features were evaluated using a five-dimensional support vector regressor, trained and tested on production music, in order to find the combination which produces the best performance. The results show the minimum number of features that yield optimum performance, and which combinations are strongest for mood prediction.
ER -
