Norming across many sets of affective pictures in order to compare affect ratings across perceptual and semantic features of images.