Celerra deduplication statistics

I read in a Celerra dedupe whitepaper that on average most of the space reclaimed by dedupe is in fact not deduped files, but compressed ones, which seems logical enough to me. Is there are way to determine the actual amount of data reclaimed by dedupe, and the amount by compression? The dedupe stats show the total amount, which is a little misleading, as most is meant to be reclaimed by compression.

The reason I'm concerned about this level of detail is due to having to perform file level NDMP backups of 2 x 16TB file systems that are scheduled to have dedupe enabled. I'm worried are the impact of having to redupe the deduped files during the backup process (actual deduped files rather than compressed as compressed are backed up in their "reduced space format" in 5.6.47.11) will have on the backup time. I need to prove to that the % of compressed files is much greater than the deduped files, and thus should reduce backup times, but am finding it hard to do so.

This leads me to another question about backing up compressed files being backed up on drives with hardware compression enabled, and whether there is a clever way to disabled the hardware compression by using a data mover param or something. Maybe i'll start another thread on that one.

scott