This is a technical question that came up over on the Steve Hoffman Music Forums (SHMF). I hope that Ted Smith and some of the other technical gurus at PS Audio can help clarify this. In a thread the statement was made that the sound of a well designed DAC will not depend on quality of the transport or the cable connecting the two together. The reasoning behind this is that a well designed DAC well take the digital data that is sent to it and re-clock it and do all of the things needed ( such as noise shaping, dithering, etc. ) to get the best analog signal out of the digital data. Seems reasonable.
For the last four or so years I have used a PS Audio DMP and DS DAC ( which now runs Windom firmware ). I was told from day one of my purchase to use the HDMI I2S connection between the two to get the lowest jitter performance from the combination of the two. That seems to say that the best quality sound from the DAC depends on the connection between it and the transport.
Recently I traded-in my DMP for a PS Audio PST. And, to my pleasant surprise, I hear even better sound quality from the combination of the PST and my same DS DAC. I am still using the same I2S HDMI connection and the only change has been the PST for the DMP. The music sounds noticeably clearer with the PST and other PST users have said the same thing. This has been attributed to the galvanic isolation used in the PST, but again this seems to say that the sound of the DAC depends on the transport.
Now I believe that even though the PST does not say so in its name it is also a memory player just like the DMP. It takes the data from the disc and buffers it and re-clocks it before sending via the I2S HDMI interface to the DAC. Does the DS DAC then re-clock the data again when it gets to it. So what I am asking is can someone explain to me in high level technical terms who does what to the data in the PST / DS DAC combination?
Yep, it’s all about the noise. The digital bits always get there correctly unless your system is “broken”. So it’s not a question of what happens to the data, it gets there uncorrupted in most all systems. (And you can run a bit perfect test to verify that.)
The timing of the bits can make a difference in many systems since reclocking is simply a low pass jitter filter. Retiming doesn’t get rid of jitter, but it reduces the higher frequency jitter spectrum. The PS Audio transports are pretty good at providing low jitter outputs and they are getting better over time. The DS is pretty good at rejecting input jitter. It’s not a question of the quality of the reclocking, they are very good so jitter isn’t really the issue.
The real problem is noise, groundloops cause noise along any loop of wires in the system (e.g. DAC to wall, preamp to wall, DAC to preamp.) The noise is directly proportional to the area of the groundloop.
Balanced interconnects can help (common mode noise rejection), great grounding can help (lower resistance alternative paths for noise.)
The quality of the cables matters, How they radiate HF noise into the system matters. How well they are shielded (or not) matters, etc.
Super useful, thanks Ted. Honestly it makes perfect sense.
I ordered an EtherRegen a while back and plugged the clean side into my MacMini then to the DSDJr via USB. I heard no difference and sent it back. Later, when I used only the Bridge II in the DSDJr I ordered the EtherRegen again and heard a huge difference. USB didn’t even compare. I was confused for a while but the extra noise caused by the MacMini then going over USB negated any possible benefit from the EtherRegen isolation the first time around.
I recently purchased a Sonore Signature Rendu SE so music is coming over optical internet and out via ultra clean USB to the Matrix, then to USB to DSD Jr. Its the best I’ve ever heard. Nothing magical about all this like many people seem to think. Its all about noise.
Can I ask a followup question? Ive heard so much about measurements of cables etc. People constantly see no differences in their measurements. Is their equipment not sensitive enough? Or are they measuring the wrong things? I know we can measure jitter and overall THD and stuff like that (excuse my ignorant non-technical speak)
Cables are a part of a system. As an example, in systems with properly terminated digital cables, where jitter isn’t much of an issue (which implies that noise isn’t high) then the length of the cables don’t matter very much. But if things aren’t well terminated or there are too many cable adapters or there’s not a lot of jitter rejection or … then the length of digital cables matter a lot more. With bad enough termination they may be completely flaky at some lengths and work fine at others.
The noise that a cable radiates to the rest of the system depends on the cable, sure, but it also depends on, well, the rest of the system. My neighbor had a stereo with a loose connection, over the speakers you could hear your digital phone doing things if you were too close to it… This isn’t a problem with almost any other system.
For digital cables having the right characteristic impedance matters a lot (e.g. 50 Ohms, 75 Ohms, 120 Ohms?) But for analog things are a lot more complicated. If you want to read about some measurements that matter check out the background papers at https://iconoclastcable.com
or search this forum for posts by Galen Gareis.
Ted, Thank you so much, as always your knowledge and insight into digital audio is so valuable. I find many people are skeptical about cables making a difference especially aftermarket power cords. I try to get them to understand that the standard “home depot” power cord makes a great antenna and that the three most important things about an aftermarket power card are shielding, shielding and shielding.