Well, then, here's a thought. Have an option for, instead of one single pass per strip, 2 passes per strip. If any pair of the same patch is different by more than the expected deviation, consider the read an error. If all is okay, average the two readings. Allow bi-directional reads, and you'd make one pass forward and the other backwards on each strip. Re-position the guide, lather, rinse, repeat. In practice, it should mean not that much added overhead to the tedium of reading a chart. Cheers, b&