Could you clarify a sentence you used?
In the article, you wrote:
The test environment is live, so readings are higher than would be obtained in an anechoic chamber readings.
Does this mean that there are sound reflections off of things like hard walls and floors that would make things seem louder than they actually are?
Yes. If the SPL readings were done in an anechoic chamber, they would all tend to be a bit lower. It should not be a huge difference -- maybe 2 dBA typically -- because the lab environment is pretty quiet & reasonably well damped.
You are right in that if you can keep the intake temp into the PSU low enough the fan ramp-up will not be a problem. In my test rig, the critical intake temp point seemed to be 29C, which is pretty low, IMO. Still, this will vary on conditions. The PSU test load in the lab is purely resistive while typical PC loads are not; they have resistance as well as inductance and capacitance, and so are more difficult to drive. On the other hand, most PCs don't routinely draw 150W DC continuously.
But your setup should work quietly. And if you want to get the noise down further, a fan swap w/ a Nexus 120 will make it very quiet indeed. (They usually start with 5.5V...)