Berkeley Audio Design excludes USB connections from their DACs to protect the delicate digital-to-analog conversion circuits from computer-generated electrical noise and clock phase noise. They believe that connecting a DAC directly to a "dirty" USB source inevitably degrades audio performance.
@jmrrobbie1 - I have read and heard about the detriment of USB connections from others, in addition to Berkeley. Interestingly, Aries Cerat has just released a $12K interface (Hronos-Project X) (link) that accepts a USB signal and feeds their DACs a customized discrete signal on a board that replaces the USB input board. They imply it makes the source/streamer quality mostly irrevalant and that it achieves a level of sonic improvement that is comparible to upgrading to the next higher DAC in their line-up.
Because I use two different DACs in my main system, it has been easier for me to route a high quality USB output into a DDC, which can then route two separate (non-USB) signals to the two DACs.
After reading conflicting reports about which inputs sound best for my DACs, I tried comparing the AES or S/PDIF rca inputs on my DACs (output from the Singxer SU-6 DDC) to a direct USB input (output from my Sonore Sig Rendu SE Deluxe Opt. streamer). The results I heard were close and I concluded the USB directly from the Sonore sounded very slightly more fluid.
After researching the SU-6 as well as reading more about the pros/cons of USB vs. AES or S/PDIF, I have decided to try again with the supposedly higher quality Berkeley Alpha USB Reference Series interface, which I now have on order.
There is so much to this digital source issue, including the network capabilities and isolation, streamer quality, interface choice, and of course the DAC itself and how well it is optimized for a particular interface. It seems there is clearly no one-size fits all approach.

