In a recent paper [1] a scheme for reducing error masking probability by modifying the output data before compaction has been proposed. The aim of the present comment is to direct one's attention to another method of ...