"it sounds like congested chaos, has a rough texture (especially noticeable in voices)and overall is just very brash sounding."
Two things are likely causing this:
1) too much audible jitter in your digital source - typical for rough textured vocals
2) FLAC files - typical for listening through a cave effect
#1 can be fixed by upgrading your USB converter and using a better power supply and S/PDIF coax cable for it
#2 is more difficult. You either have to be willing to live with .wav files and minimal tags or get a server like the Antipodes that sounds identical with both. Another option is to try JPLAY in conjunction with Jriver. This may or may not help.
Jitter numbers are all but useless when they get to these levels. The problem is that jitter like most phenomena is not something that occurs in one instance or repeats the same every time. There are billions of instances and each one is different. This creates a Gaussian distribution of jitter events that can be examined. Some of the jitter events may be data correlated and others may be random. When one says that a particular device has 500psec of P-P jitter, this is incomplete information because the 500Psec event may only occur every 30 seconds, or it may occur every millisecond. Big difference in these two. Also, the shape of the peak of the distribution may be very narrow in one case and very wide in the other. In other words, the distribution of the jitter events must be studied in order to determine whether one device is actually better than another. Then of course, one must listen to music as well. The human ear is the only thing that will tell you definitively whether one jitter distribution will sound better than another.
I would recommend upgrading the USB converter first, as well as an upgrade power supply and Coax cable. Most of these have money-back guarantees, so your risk is low.
Steve N.
Empirical Audio