It doesn't involve any inter-frame coding or prediction, so evaluating stills is fair. It compresses a single line at a time.[1]

[1]: http://www.vesa.org/wp-content/uploads/2014/04/VESA_DSC-ETP2...