Text this: Performance of the vocal source related features from the linear prediction residual signal in speech emotion recognition