In this work is proposed a speech emotion recognition model based on the extraction of four different features got from RAVDESS sound files and stacking the resulting matrices in a one-dimensional array by taking the mean values along the time axis. Then this array is fed into a 1-D CNN model as input.
