Supervised machine learning for audio emotion recognition

Abstract

The field of Music Emotion Recognition has become and established research sub-domain of Music Information Retrieval. Less attention has been directed towards the counterpart domain of Audio Emotion Recognition, which focuses upon detection of emotional stimuli resulting from non-musical sound. By better understanding how sounds provoke emotional responses in an audience, it may be possible to enhance the work of sound designers. The work in this paper uses the International Affective Digital Sounds set. A total of 76 features are extracted from the sounds, spanning the time and frequency domains. The features are then subjected to an initial analysis to determine what level of similarity exists between pairs of features measured using Pearson's r correlation coefficient before being used as inputs to a multiple regression model to determine their weighting and relative importance. The features are then used as the input to two machine learning approaches: regression modelling and artificial neural networks in order to determine their ability to predict the emotional dimensions of arousal and valence. It was found that a small number of strong correlations exist between the features and that a greater number of features contribute significantly to the predictive power of emotional valence, rather than arousal. Shallow neural networks perform significantly better than a range of regression models and the best performing networks were able to account for 64.4% of the variance in prediction of arousal and 65.4% in the case of valence. These findings are a major improvement over those encountered in the literature. Several extensions of this research are discussed, including work related to improving data sets as well as the modelling processes.

A Comparative Analysis of Modeling and Predicting Perceived and Induced Emotions in Sonification

“…Ntalampiras [52] compared emotion prediction using two CNNs that were designed to individually predict arousal and valence. The authors used the EmoSoundscapes data.…”

“…Part of the work performed by Ntalampiras [52] was on the EmoSoundscape dataset, and they used CNN models. The MSE values reported for arousal and valence prediction were around 0.049 and 0.11, respectively, which are equivalent to 0.22 and 0.33 for RMSE.…”

Modeling The Effect of Background Sounds on Human Focus Using Brain Decoding Technology

“…23,24 Using properties of sound as features and experience measures as labels for those features, several groups have attempted to build machine learning algorithms that can predict emotional responses based on the sound properties alone, commonly according to a valence-arousal circumplex model. 25,26,27,28,29 However results in this area remain mixed due to lack of sufficiently high dimensional measurement and modeling tools suitable for capturing the fast changes in human experience that accompany changes in sound. 30,31…”

Worship the Penguin: Adventures with sprites, chiptunes, and lasers

“…Upon connecting all three lasers, a green LED lights up, and a 'success' sound is emitted, which gives positive emotional feedback to the user through an ascending harmonic sequence of tones. This emotional feedback can be understood as an affective sound design (see Cunningham et al 2020).…”