So there seems to be a lot of discussion about I2S vs usb here. I have my own experience which favors I2S but not going to extrapolate this is universal for all setups. I2S does have inherent advantages in that it is native signal path in dacs, it separates data from clock, a lone possible disadvantage is an external clock may be used vs internal dac clock, the theoretically better placement of clock is closest to signal path. So I have the choice to use the external OXCO clock in my Gaia DDC vs internal clocks in my dacs, both Femto, in other words I can sync or unsync clocks, in both cases the external OXCO clock in Gaia provides better sound quality than local clocks (this using 1/2M Tubulus Ximius I2S cable).
So for those who claim none of this is an inherent advantage of I2S, is it an inherent advantage to send signal on detour through usb boards of who knows what quality, only then to convert back to I2S? Is it an advantage to require the dac to separate clock from data?
Not saying usb can't sound good, I used it for at least a decade without complaint, only recently did I discover the I2S setup beat it. And over that decade I've used all manner of streamers with various usb implementation, tweaks and isolation of usb, very nice usb cables, dacs with some very nice usb boards including custom implementations. I2S belongs in the category of less is more which some seem to espouse, point being why do needless conversions.Not going to argue this point as I sometimes experience more is more such as optical conversion.
All in all, I'd take the battle of I2S vs usb on a case by case basis, optimize both and report back, actual experience counts for more than second hand repetition of some other person's experience.