From my own experimentation, I can tell you that you get a noticeable SQ increase by outputing to the Hugo via the RCA Coaxial vs. USB. I did this via my Wadia ipod dock and other transports. In any case, I recommend you try it yourself by taking same song and outputing to HUGO via USB and via coaxial.
The article you are referring to is written in such as fashion as one can easily mix up issues he is bringing up. But in the end he is essentially recommending that you output to the Hugo via the RCA Coaxial via the USB convertors (AP2, etc.). The power issue comes in that he recommends improvements to the power being provided to the USB convertors.
I am not a big believer in the USB bandwidth related issues as I have a fairly extensive technical background and none of the arguments I have heard make much technical sense given that the examples of what could go wrong basically give outlier cases that would result in loss of USB connection more than loss of SQ. However, I often times do hear a difference from better USB cables and off loading USB to coaxial. As to why USB convertors result in better sound quality, the main theories I hear and believe is this:
USB has two SQ issues.
1) One possible feedback or interference from USB power that people often try to use better USB cables to solve. If the USB DAC is using USB power, the problem could be even worse as consistent power is vital to almost all USB operations. However the HUGO does not use USB power, but the USB power is still connected and still necessarily to make initial USB handshake. The USB convertors he mentioned all take great pains to separate power from the source device via USB to providing their own cleaner power.
2) USB has inherently more jitter. Most of the USB convertors are also taking in the digitial feed and queuing up the data. It then reclocks its queue for smoother sound. Hifi Nutter mentioned why didn't Rob Watts think of this? The key thing is that it isn't typically handled in the DAC. The Hugo is the receipt of the data and if the data is "messed" up in some way, then it really doesn't have anything to do with it. There are some desktop DACs that spend quite a bit of effort in receiving data and reclocking it so as to accomplish similar things. But it would likely involve a much bigger package than the Hugo is contained in. It wouldn't surprise me if the desktop version of the Hugo spends much of its effort around this as well as providing consistent power to the DAC chip. Although I don't know how sensitive the Chord DAC chip is to power issues and, being that battery power is so much easily to implement with less issues, it might be an expensive challenge for them to build out a better AC power solution (my guess is some linear power solution).
