Comparing Feature Sets for Acted and Spontaneous Speech in View of Automatic Emotion Recognition | Synapse