If you're doing DAT > WAV transfers, first step before any processing (resample, normalize etc...) should be to remove the DC offset. First Scan - If it finds one, remove it before the processing. As I understand it, since processing is always performing mathematical algorithms on your sound files (on a computer) having a DC Offset is like having a centerline that moves. Like this:
--------____________---------------------_____________
If you perform processing on a centerline that looks like that, you might get some more audible anomalies. Note, it isn't as simple as the above example, but that analogy will hopefully help you understand why you want to get rid of it.
-Brad