The Intelligibility of Filtered-Clipped Speech in Noise