Know your enemy

September 18, 2022
 by Paul McGowan

I get a lot of flack for my dislike of USB as a medium between the computer and DAC.

Truth is, it’s not USB I don’t like. It’s what USB connects with that bothers me.

Noisy computers.

Here’s the deal. We all know computers are vile, noisy affairs that hopefully are as far away from our pristine high-end audio systems as possible.

When we want to extract music from these noise-ridden contraptions what’s the best arm’s-length, noise-free way to do it?

Ethernet—an isolated one-way communication medium.

What’s the worst way to connect?

USB—a two-way connected open-gate flood of data and noise.

Our goal should be to get the data out of our computers with as little noise and connection as possible, which is why Ethernet (or long-ago fiber optics before Toshiba’s TOSLINK format screwed up our chances for high sample rates) or WiFi are the best choices.

I get it. It’s a heck of a lot easier just to fire up a USB cable and be done with it. Ethernet and WiFi are pains in the keester.

But better.

One possible solution is to use a USB reclocker device like the Matrix.

It helps, but it isn’t perfect.

Perfect is a new DirectStream MK2 DAC with 100% galvanic isolation.

With an MK2 in the system, you can use USB and stop worrying about it.

Subscribe to Paul's Posts

32 comments on “Know your enemy”

  1. Such a simple solution strongly works against the audiophiles’ urge to fiddle and buy fancy boxes 😉

    I guess if the noise is still inherent in the USB signal and just blocked at the DAC, this is uncritical for the DAC’s audio but maybe still less optimal than better connectivity from a “cabling radiation towards other components and environment” point of view?

    For the DSII…is there no difference anymore between USB and I2S?

    1. Same here. Still after all these years I’ve never had an interest in using a computer or laptop as a main source for music.
      Couldn’t agree more.
      A CD player or cd transport gives out way more assurances and one of them is noise floor reduction. 🙂

  2. Aren’t here the biggest enemies lousy power supplies? My DAC sees at the USB-B input only the cleanest DC voltage from a battery supply. And my active SPDIF cable also features now battery supply. The exchange of the standard Ac to DC power supplies made a huge improvement. And no need for exotic mega-buck digital cables.Galvanic isolation of LAN and USB is added of course by simple boxes – no rocket science.

  3. Neither WiFi or Ethernet are keester pains when integrated well into a piece of electronics designed for them, especially when good instructions are included. One requires wires the other doesn’t.

    Galvanic isolation has been around for quite some time. Fiber transmission of data has been around for quite some time. It seems that the audio community made a conscious decision to ignore the prospect of light data noise isolation, but now are embarrassing that as perfect.
    Or to put it another way, the source side is now perfect and it exists in the digital world.

    For years the elimination of digital jitter was paramount for high end audio playback. Now that has been “licked”, the concept of optical galvanic isolation is now proclaimed as perfect. I guess any future DAC development now just involves a prettier case. Ted just designed himself out of a job. 😀

    1. Wasn’t “clock precision” the core focus of Ted’s DAC design approach? I still wonder what happened with Ted’s Statement/Reference DAC”. Not to mention the PS Audio streamer. For my 15 years old music streamer the addition of internet-radio and streaming was just a matter of sw-upgrade. And sound quality was simply improved by a new player-sw.

      1. PS
        To answer your questions best I can,
        I assume the clock will only be improved if possible. I have the impression it already has been.
        The TS ref dac has probably just changed in name only.
        But these questions should Be answered by Paul McG

        From where I sit…
        The current FPGA can’t handle any more and probably is now obsolete? So that’s the 1st nail in the coffin.
        The current USB only handles up to 2xDSD and since O.R. Is up to 4xDSD that needs to be addressed. That sample level may be handled by the current transport, but typical file size of 4xDSD is bigger than what can be put on a data dvd. So either the amount of tracks or play time has to be reduced, or That leaves usb as a primary convenient source. Since it’s inherently noisy that needed to be addressed. (Optical galvanic isolation was PSA’s solution). The only solution to get 4x into the current DSDAC is thru the I2 buss. That makes reliance on someone else a necessity.

        Add in the fact that if you’re going to replace a current top of the line model and you don’t want to regress then new and different circuit and software topologies are needed. So why not point out the one that may have the biggest effect?

        The server, while a cool idea I think got too complicated and required software engineer resources and probably on going updates rendering it impractical given some of the other solutions currently out there.

        These are just my observations and guesses, trying to look at things pragmatically.

  4. From the DSDDAC manual V7.
    “The design uses a carefully crafted high bandwidth audio transformer at the output of the instrument for both galvanic isolation from the outside world as well as low pass filtering.”

    One of the differences with the MK II is that it now uses galvanic isolation on the input side also? Effectively making it an island unto itself… No outside world influence? (At least on the data side?)

    1. According to my engineering studies many years ago, asynchronous sequential switching circuits are a nightmare so re-clocking may be the best solution.

      Perhaps when I go to my great reward there will be beautiful music surrounding me that does not require Physics solutions which will be totally unnecessary. I will be gently bathed by the music that I love for eternity with zero distortion.

      “This is the big one Elizabeth honey… I’m coming to join you”
      Fred Sanford

      1. ??? “The asynchronous USB DAC (async USB DAC) uses its own clock to regulate the data rate from the computer, which allows for jitter-free data transfer. ” – PC Mag

  5. In my modest setup I’ve been streaming wireless from a Ubiquiti Dream Machine router at ( 700 mbps ). To my MacBook Pro using Audirvana Studio which shuts down some hardware to reduce noise and power drain. I also try to listen on the battery supply which usually last 3-4 hours. I’m using a audio quest usb cable (10 ft. ) to my Mytek DAC. I have tried really hard to hear all this noise that’s supposed to be in my system, but I haven’t heard it yet. Just great music.

    1. As a side note; there is also some joy for the Linux users, Bandshed AvLinux has special kernels to minimize noise and keep timing accurate.

      Personally I have found it distracting, using a tablet, cellphone or laptop. Think if I had a bigger room, probably would not bother me.

      A friend built a power supply out of banks of super capacitors, noticeable difference, but had to be charged everyday. Like you had mentioned usage around 4 hours.

    1. That was a great reply Richtea. I don’t think that your system is modest by any stretch of the imagination. And your last line was just beautiful. So I would say to you…
      “Rise And Rise Again Until Lambs Become Lions”

      1. Sorry I wasn’t clearer. I know the AirLens will not have a USB out to compare. I think we are asking the same thing.

        Asking it a different (and hopefully clearer) way, does the source of the input into the DS II no longer matter? Will a DS II fed by usb cable from a streamer (or a Nucleus or a fancy music server) sound the same if fed by a I2S cable from the Air Lens?

        And one more similar question. Does the cable (be it a usb, I2S or coaxil) used to stream data to the DS II no longer matter?

        1. Yes! If it does not just „less“ matter, this is quite a breakthrough. I’m wondering galvanic isolation is such a new thing and how much effort it means.

          I don’t get yet, why not everyone uses it then. There are much more complicated things in electronic engineering think, which are more widely used.

          I’d like to understand this.

  6. My discovery was forced upon me…..

    AT&T took away the DSL in my area and mandated we all turn to fiber. Out with the copper and in with the fiber.

    At first, the higher speed was apparent. Then one day, I noticed that the new type of modem they supplied had some options I could control. I use an audiophile Ethernet cable that connects between the modem and my PC. But, the WiFi was turned on with the modem, as well as a connection for a VOIP phone which I do not use. I disabled both functions. Everything became quiet in a new way. It seems that copper wire for the DSL also caused noise. That was the first noise to go.

  7. I used ethernet with Linn from 2009 to 2014. From then I was using an Auralic Aries with internal Femto reclocking by usb into a PS Audio PWD Mk2 DAC. From 2016 I used the Auralic wirelessly into Devialet. From 2019 I used Innuos with ethernet and from this year use Innuos Sense by usb. I also use Roon wireless.

    So my experience is the complete opposite of Paul, over the last 13 years having used usb, ethernet and wireless with no problems, optimised for audio, most of the time with galvanic isolation, and never with a noisy computer.

    1. It was more the headphones people using desktop computers that needed usb noise reduction products. Some just used filters, like the extremely popular AQ Jitterbug (I had one), ifi made units using filters and galvanic isolation, SOtM tX-USBultra etc . They have been selling them for years. We now have much more expensive units like the Innuos PhoenixUSB that completely regenerates and reclocks the signal.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram