[Five]

the 48kHz>44.1kHz downsamples are not desirable, they normally are due to source DAT tapes recorded at that sample rates or Edirol R-09 recordings (this unit cannot do 44.1kHz), and it is only done to make it audio cd-comatable.

if you want higher than cd quality for your personal archives first thing to consider doing is capturing at 24bit rather than 16bit before considering upping sample rate. 24bit is the limit of the FLAC codec, so if you capture at 32bit you cannot archive as flac without truncating to 24bit along the way or keeping them as MASSIVE wav files. fwiw wavpack has support for 32bit but only when using Adobe Audition (and I have not tested this filter personally).

If you cannot set wavlab exactly to 24bit for capture use audacity. if you do decide to capture at a higher sample rate, use the highest quality setting to resample (preserving pitch). For taking it down to 16bit from 24/32bit consider this: dither is for weenies. The advantage of dither is stopping stuttering during fadeouts to silence at the cost of dulling the entire audio content. I defy anybody to post a 24bit sample plus a 16bit dithered version that sounds better than undithered 16bit version. yes, this might seem radical but try it out, if I am wrong I will eat my words.

similarly, resampling from 48kHz to 44.1kHz will produce a dulling effect, so the question is does it sound more clear recording at 44.1kHz than recording it at 48kHz then resampling to 44.1kHz. conventional wisdom dictates that the shortest path is the best, but try it out.

try it all out, take this advice as a starting point but test everything here to confirm. record at 16bit, 24bit, 32bit, 44.1kHz, 48kHz resampled to 44.1, 96kHz, consider the 24bit FLAC limitation etc etc and draw your own educated conclusions, not just following blindly some things that you read. Then post back with what you find to give us new ideas of things to try, better ways to do things.