It’s really complex, the way I see it. It’s a great question though.
SPDIF has one significant advantage: Both SPDIF transmitters and receivers are real (hi-fi) audio grade components. USB transmitters and receivers aren’t, they are computer components. They don’t deal with noise they induce (both through USB transmitter and receiver), and even transport protocol commonly used (USB isochronous asynchronous, UAC 2.0) isn’t quite tailored to hi-fi (or especially hi-end) needs.
SPDIF, although it’s really tailored for hi-fi audio, has few significant disadvantages too: transmitter side is master and receiver is slave. Clock is encoded with the signal information, however it’s subject to transport jitter by itself as well, and the reconstruction of it will not be perfect - and in case receiver (DAC) does have its own precise clock, it’s not perfectly fitted (identical) to transmitter’s clock, by which the signal is encoded. And it’s really a synchronous protocol, so precise alignment of clock on the receiving side is important. OTOH USB is asynchronous and encoded different way so it doesn’t at all depend on clock on the transmitter side - it’s able to dejitter the signal effectively using the receiver’s master clock.
So while SPDIF’s bigger weakness is jitter, USB’s bigger weakness is noise, plus occasional altering of the digital signal data, which happens because there’s no data resend in UAC 2.0 and there’s no ECC either - if there’s error in data, it’s fed in DAC as is and DAC handles it the best it can
In practice, however, much depends on quality of receiver side where various SPDIF or USB receivers have various quality in reconstruction of the digital signal (and clock in SPDIF case), plus handling the noise sent from the receiver through data, ground and voltage lines.
In case of computer audio computers make noise on their own, so it’s a real challenge for the rest of the system.
I’d say the end result can’t be unanimous, the implementation matters more which will be better. And the same device does not need to have the same quality of implementation for USB and SPDIF receiving circuitry.
Practice says cables matter as well through both. In my view SPDIF is cheaper to ensure a quality digital cable. While with USB, the quality of USB receiver matters for cable demands as well. With less good USB receiver DAC (or converter) is more sensitive to USB cable quality.
Besides, computers/PC have a different level of noise they generate and in situation when they’re especially noisy it’s quite audibly apparent. Really noisy PC/comp is capable of making USB absolutely inferior as level of noise can become too much for DAC to successfully handle.
Disclaimer: all mentioned above are personal observations, experience and conclusions. Also some errors in assembled knowledge are possible.
SPDIF has one significant advantage: Both SPDIF transmitters and receivers are real (hi-fi) audio grade components. USB transmitters and receivers aren’t, they are computer components. They don’t deal with noise they induce (both through USB transmitter and receiver), and even transport protocol commonly used (USB isochronous asynchronous, UAC 2.0) isn’t quite tailored to hi-fi (or especially hi-end) needs.
SPDIF, although it’s really tailored for hi-fi audio, has few significant disadvantages too: transmitter side is master and receiver is slave. Clock is encoded with the signal information, however it’s subject to transport jitter by itself as well, and the reconstruction of it will not be perfect - and in case receiver (DAC) does have its own precise clock, it’s not perfectly fitted (identical) to transmitter’s clock, by which the signal is encoded. And it’s really a synchronous protocol, so precise alignment of clock on the receiving side is important. OTOH USB is asynchronous and encoded different way so it doesn’t at all depend on clock on the transmitter side - it’s able to dejitter the signal effectively using the receiver’s master clock.
So while SPDIF’s bigger weakness is jitter, USB’s bigger weakness is noise, plus occasional altering of the digital signal data, which happens because there’s no data resend in UAC 2.0 and there’s no ECC either - if there’s error in data, it’s fed in DAC as is and DAC handles it the best it can
In practice, however, much depends on quality of receiver side where various SPDIF or USB receivers have various quality in reconstruction of the digital signal (and clock in SPDIF case), plus handling the noise sent from the receiver through data, ground and voltage lines.
In case of computer audio computers make noise on their own, so it’s a real challenge for the rest of the system.
I’d say the end result can’t be unanimous, the implementation matters more which will be better. And the same device does not need to have the same quality of implementation for USB and SPDIF receiving circuitry.
Practice says cables matter as well through both. In my view SPDIF is cheaper to ensure a quality digital cable. While with USB, the quality of USB receiver matters for cable demands as well. With less good USB receiver DAC (or converter) is more sensitive to USB cable quality.
Besides, computers/PC have a different level of noise they generate and in situation when they’re especially noisy it’s quite audibly apparent. Really noisy PC/comp is capable of making USB absolutely inferior as level of noise can become too much for DAC to successfully handle.
Disclaimer: all mentioned above are personal observations, experience and conclusions. Also some errors in assembled knowledge are possible.