I think that's just the sound of both drives loading their heads. Unlike older models where the heads parked on the platters in a special landing zone, current models retract their heads off the platters onto a loading ramp.
Here is a Hitachi white paper that illustrates what is happening:
http://www.hitachigst.com/tech/techlib.nsf/techdocs/9076679E3EE4003E86256FAB005825FB/$file/LoadUnload_white_paper_FINAL.pdf
BTW, I take issue with Hitachi's claim that "load/unload technology was discovered in the mid-1990s
as a viable alternative to Contact Start-Stop (CSS)". I worked on Control Data hard drives in the early 1980s that had load/unload ramps.
As for your SMART reports, the only attributes that appear odd are High_Fly_Writes for both drives, and UDMA_CRC_Error_Count for one of them. It looks like both drives will hit the threshold for High_Fly_Writes within a matter of hours, although I suspect the value will remain at 1 and therefore not trigger a SMART failure. I have no idea how this will affect data integrity.
AFAIK the UDMA_CRC_Error_Count reflects a problem with data transfer along the external interface, so I'd be checking the cable. You may want to swap cables between drives, or swap positions on the motherboard.
There also appears to be a hiccup with Spin_Retry_Count on both drives. Maybe that's insignificant, but you may want to check the +12V supply rail from your power supply.
BTW, the high raw numbers for Raw_Read_Error_Rate, Seek_Error_Rate, and Hardware_ECC_Recovered are normal. For example, a seek error rate of 1764826 means that the drive has experienced 0 seek errors in 1,764,826 seeks.
