Investigating salient representations and label Variance in Dimensional Speech Emotion Analysis