Does your DAC sound better through the S/PDIF or I2S inputs than through USB?


For a long time I used a Windows NUC driving my DAC through USB. (A modified Gustard X20Pro.) Then I got an Aurender N100 which provided a better USB source. I also tried a DAC on the recommendation of a friend in the New Jersey Audiophile Society. People were excited about this there: the Gustard R26.

I found that the R26 is disappointing through its USB input, but excels through the Coax S/PDIF input using a Gustard U18 Digital/Digital Convertor.

This probably points to a subpar USB implementation, to be honest, but I started wondering if this is a general pattern among DACs that can take both S/PDIF and USB.

Anyone find something similar in their DAC?

For example, I'm interested in a Schiit DAC (not sure which model yet). Now that I have the DDC I can drive it through Coax and compare to USB. With the Gungnir 2, I may not have the choice of USB because they have a USB C connection and my only good USB cable is A to B.

Note: what is wrong with the R26 through the USB input: very light bass, unintegrated treble (sibilance kind of clouds over the sound and doesn't integrate with the body of the instruments)

magon

Showing 2 responses by sns

My take is there are no universal rules in regard to interfaces, each situation unique. One should research the quality of each interface on both streamer and dac, then try to match the two. Generally we find usb is the de facto interface these days, this based on some faulty assumptions IMO. I2S is the native signal path within dacs so why would we want to detour through a usb board when an I2S interface would be the most direct path. The most often given reason for not providing I2S interface is I2S clock not designed for external connection to dac's internal I2S data path. This is true, reflections on I2S cable can cause some amount of jitter, this is reason to keep I2S/HDMI cable short as possible. So, in considering using I2S interface we much consider some trade offs, the extra amount of jitter using external clock vs dac's internal I2S architecture. Assuming we're using DDC to accomplish this, one has to compare the quality of clock and power supply to that clock in DDC vs dac. Specifically, does the clock and power supply to that clock in DDC with the possible added jitter outperform dac's internal I2S clock and power supply. I'd posit a DDC with quality OXCO clock and power supply is likely to outperform a TXCO or Femto dac clo ck virtually every time, this even with the small amount of added jitter via DDC. And when comparing to usb one should also consider the advantages of discrete runs for data and clock using I2S/HDMI.

 

So the above assumes dac even provides I2S interface.  Minus that, usb is not a bad interface, many dacs have very nice usb boards and quality streamers also have very nice usb interface. Another consideration with using I2S, interface INTO DDC is of extreme importance as well, this means usb for many, so don't scrimp on usb even if using I2S. Bottom line, best interface highly variable.

The salient point using I2S vs all the others is the ability to use a better clock via a DDC vs dac's internal clock. With superior clocking presentation can be greatly affected. Proper clocking means data packets more correctly timed which results in a much more analog like presentation. If you investigate most dac's internal clock you'll generally find Femto clocks with rather pedestrian power supplies to those clocks, one can do better in this case. Not sure I'd want to purchase any dac devoid of I2S input, minimally I'd want to compare my Denafrips Gaia with Tubulus Ximius I2S cable to the dac's internal I2S architecture and/or it's usb implementation. The only higher level dac I know of that incorporates I2S is the Mola Mola Tambaqui.