A Comparative Study on Physical and Perceptual Features for Deepfake Audio Detection