Comments on Profile Post by k.e.

My current understanding is that it's a big, fat 'it depends'. Higher sampling speeds help with the analogue filtering stage (e.g. Yggdrasil). Higher bit depth, up to the realistic ~20 bits, reduces quantization error (though care should be taken if it is SOTA 20 bits ADC or using a digital filter to emulate 24 bits).

But as Marv points out, things like mastering and DR compression can dwarf all that:
http://www.superbestaudiofriends.org/index.php?threads/paul-simon-diamonds-on-the-music-analysis.4050/

Also, as Mike points out, higher sampling speeds help with Nyquist conundrums:
http://www.superbestaudiofriends.org/index.php?threads/the-mike-moffat-2-at-schiit-blog.3507/page-20#post-153352

The nutshell version after years of both theory and practice in pro audio and enthusiast playback: for consumer, higher SR makes filters easy and less audible, fuller harmonic series. For pro, captures more info for processing (noise reduction, more harmonics, bias tone of tape transfers).

Only get the HD Audio if the master tapes are used and dynamic range is "plenty" for enough headroom volume to match with your audio system. I prefer early pressings when possible for CD. For downloads 16 bit FLAC is often good enough...