To wrap up the filter stuff I mentioned earlier, here are some graphs that illustrate most of them. I found them very helpful, and the credit goes entirely to
John Westlake, the designer who made them available in
this post on the PFM forum. If you're interested, you can follow the discussion following that post. One important point that John made is that the graphs below are just part of the resulting audio, and that the time domain (not represented) may be as important if not more.
Fast filter(s) (i.e. Sharp/Slow Rolloff I believe), only included for reference, not very good in any way:
Minimum Phase (apparently the most "analogue"-like, recently increasingly popular among PFMers):
Optimal Spectrum (the most technically advanced. i.e. oversampling I believe)
Optimal Transient filters (less advanced but potentially nicer sounding, i.e. non-oversampling I believe)
The difference between the 3 OTs is just different math to arrive to the same general result, so differences, if any, are minimal.
Optimal Transient XD
Optimal Transient DD
As far as I'm concerned, switching between these filters has the biggest impact on SQ (and even then it's rather subtle), more so than switching inputs, power supplies, cables etc. Only switching outputs has a similar if not bigger impact, because the XLR out signal is not just balanced, but also stronger and cleaner (S/N ratio), as per earlier post.
For headphones that have well-controlled (or even recessed) treble, like my favourite LCD3 or the Sennheiser Momentum, I like Minimum Phase or Optimal Spectrum. For those with some degree of excess treble or are prone to sibilance, like the Edition 8 or to a lesser extent the TH900, I prefer one of the Optimal Transients. Hope this helps fellow Head-Fiers & M-DAC owners.