Since the addition of Denafrips Gaia, Musetec MH-DA006 dac, Tubulus Ximius I2S and new power cables have been recent additions I've been waiting for system to stabilize (over 300 hours burn in) and for me to acclimate to particular sound qualities of all this new equipment. So, in evaluations of syncing to external clock (Gaia) vs using internal dac clock it has become patently clear a limitation of this dac, and perhaps most if not all dacs is the I2S internal clock. When you look at the OXCO clock in Gaia and Hermes you'll find a rather intensive power supply for that OXCO clock, even if a dac implemented an OXCO clock doubtful they'd find enough real estate for it. On top of that, most dacs devote a fair amount of space to usb boards, space I'd think better used for I2S clock power supply.
The implications for sound quality are easily hear using Gaia clock vs Musetec clock in that using Gaia clock sound staging greatly enlarged and images far more dimensional, greater density, Sense of reach out and touch performers vs the relatively flat presentation of internal clock.
Beyond this, in my evaluations of optimized usb vs optimized I2S, I2S easily surpassed usb, and maintains this superiority even when bypassing Gaia clock. From this I surmise I2S alone (without using ddc clock) is superior to usb, this points to inherent I2S superiority via it's dedicated data and clock runs, and the elimination of detour through a usb board (I2S is native signal path in dacs). the better clock in Gaia is added major benefit. Based on this reasoning I'll extrapolate a quality implemented OXCO clock in a ddc using I2S interface should outperform any dac's usb implementation and/or likely inferior I2S internal clock.