Off topic but I'm throwing in my own data point

My response is meant to be supplemental to your comments....
I tried replacing the 5V regulator in the SU-2 DDI with a Belleson [also tried the Sparkos] and heard no difference. Also tried several oversize transformers both RCore and toroid with Mundorf Electrolytics and discrete ON diodes for the bridge - no difference.
John Swenson of Uptone Audio/Sonore product design made comments about DC umbilical cables and advantages of Star Quad. He commented that digital circuits need power instantaneously where the power supply's ability to deliver these transients [my words from memory not his] is critical.
Then Antipodes Audio redesigned their LPS to be hybrid LPS/SMPS with a "white paper" on their website about the power supply needing to be "fast" [again my oversimplification]. I did this upgrade for my Antipodes server and player and experienced a 10-15% SQ improvement in terms of soundstage expansion and transient delivery.
So I took a shot and bought a Ferrum Hyposos, programed it for 5V and connected it straight to the SU-2 digital board. There was a clear 10% improvement.
Just another data point FWIW.
P.S. - I'm not implying that an OCXO needs a "fast" power supply and these references were specific to the DDI at hand.
And, as I don't have the measurement tools to make direct correlations between "fast" and "low-noise", my comments from tweaking experiences are meant to be interpreted purely as "data points"