I put a DDC between my SOtM bridge and my Amber 3. This is a $1500 bridge with a good onboard clock… still, the DDC still made a noticeable difference??? There was a noticeable improvement in instrument separation in particular. Nothing else really changed a lot even though it all sounded better but it was the instrument separation that really stood apart. So it seems like there’s a lot to be gained by getting off USB even at the higher end.
USB for audio suffers from two things, 1) USB noise from the power chain itself since noise generated in your PC can be introduced into your DAC via the USB cable. 2) The time domain which is muddied by USB audio signal needed to be converted into I2S to communicate with the DAC.
We take care of the first issue by separating (external) power supplies, galvanic isolation via USB decouplers, however way, it all helps, yet still they do nothing to address the second issue.
DAC’s commonly re-clock incoming USB signals, they sometimes regenerate it entirely, but most often they correct it with a PLL They still have to have a clock though they aren’t asynchronous devices. ref @Polygonhell You see the clocks in this Amanero 384 USB (which is in a lot of decent mid to higher tier DAC. Or Shiit’s very own UNISOM USB, another well regarded USB implementation.
So if via USB the signal is regenerated or cleaned up before it’s sent to the DAC why do I see improvement? DACs only speak I2S they don’t speak USB so anything coming into the DAC via USB needs to be converted. Perhaps it’s this point that seemingly is an Achilles heel?
The SOtM I referenced above doesn’t suffer from USB noise coming from power (unlike a dirty PC environment). I have a very high quality ifi Power Elite feeding it, and the SOtM cleans and re-clocks the signal it sends out to a DAC, but it only sends out that signal via USB. Yet still there’s performance improvement when you re-clock regenerate this already clean signal. So what’s in USB that is inherently noisy?
There are other streaming bridges like the Pi2AES that perform this same cleaning (re-clocking) up of the signal but they send it out via I2S or AES/EBU, BNC or SPDIF Coax. So it bypasses the USB to I2S conversion.
The difference is if you send via USB, the DAC is responsible for the clock, if you send via anything else the clock is on the wire and for everything except I2S has to be de-multiplexed out of the signal. ref @Polygonhell This may also explain why often people claim that an I2S out from a re-clocked signal to an I2S input in your DAC is “the” way to go. But it also goes to show that there’s clearly room for improvement in a lot of DACs by getting off of USB altogether even if you’re already giving it a clean re-clocked signal as in my case. Perhaps it also explains why things like the Pi2AES sounds so good to begin with for the price range. It’s simply not burdened to begin with.
Clearly the ultra high end world solves these issues but we’re talking upwards of $5k DACs and well into the price of a small 4 door hatchback. But for gear less than that, try to move your gear around. I had my DDC sitting unused in the closet and what do you know? A nice little upgrade to my 2 channel stereo for zero dollars spent.
Hopefully we can use this thread to combine a lot of the disparate conversations about DDC in the site and have a good central point for discussion so feel free to add your experiences or correct stuff that’s not entirely clear or accurate in my write up.