known nothing about the results, i suspect that mid-ranges are poorer mainly because human frequency response is most sensitive towards mid-range aka vocal-pitch frequency