Because Rob Watts has a strong opinion regarding his design and you agree with that opinion, doesn't make our real world experiences wrong.
He's been quoted speaking against the use of Linear Power Supplies, yet a cottage industry thrives in spite of his advocating for the use of his superior SMPS. I've A/Bd the stock SMPS power supply against a LPSU… I chose the latter. He might think it sounds good, but I'd wager most of us prefer the latter.
I have no experience with re-clocker. I just post what Rob said. But, based on my experience/listening tests with power supplies, I fully believe him.
I used to be a LPS guy, but after I did the test
(you can search my previous posts), I found the stock PS sounds closer than a unregulated battery than a LPS does. The LPS adds too much low frequency harmonic, and make the sound too thick and too warm, but harsher treble.
I always believe the signal path should be as simple as possible. Adding any active stage in the signal path increase the chances of getting more noise, which affects the SQ.
If I was going to add a USB re-clocker to my system, I would ask myself two questions:
1. Does cleaning the source jitter before Mscaler help, especially Mscaler generates jitter (measured by ASR, and confirmed by Rob)? I don't think so. If it does help, would it be more helpful to put re-clocker at the DBNC side of the Mscaler (if there is a such product).
2. Will adding a USB re-clocker increase noise or reduced noise when Mscaler's USB decoder and galvanic isolation are powered from the source side of the USB? Don't know the answer. Depends on the re-clocker.
At the end, I would just use toslink as a reference to compare USB. If they sounds identical, I am done. If they sounds different, I will then start worry about it. In my current system, they actually sounds identical, so I am done.