USB Nervosa Thread Decrapifiers, pro interfaces, and bears oh my

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by zerodeefex, Sep 28, 2015.

  1. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    Bear in mind that even basic AOIP solutions using Ethernet are significantly more robust than even the best USB implementations can be. They have to be. On Ethernet you have to be able to withstand packet collisions, re-transmits (something USB Audio cannot do), and out-of-order packet delivery. Also, proper Ethernet is electrically isolated - without having to spend big money to get that.
     
  2. winders

    winders boomer

    Banned
    Joined:
    Feb 13, 2017
    Likes Received:
    1,596
    Trophy Points:
    113
    Location:
    San Martin, CA
  3. Kattefjaes

    Kattefjaes Mostly Harmless

    Pyrate
    Joined:
    Sep 5, 2016
    Likes Received:
    4,521
    Trophy Points:
    113
    Location:
    London, UK
    Hey @Torq - stupid question time.. How good is the "Adapticlock" stuff in the Yggdrasil/Gungnir Multibit?

    To be more specific:

    If you have a source with jitter (frequency smearing, dulled transients etc.) audible on a lesser DAC*, but which doesn't light the BBGL on those two big ones, does that mean that the VCXO is used, and the signal gets reclocked cleanly? How clean is cleanly if so, and how does that compare to some of our favourite ghetto SPDIF sources?

    I'm not sure that I trust myself to work it out by ear, and I don't have the measuring-fu right now.




    * That is to say, the Jittery source sounds jittery on a lesser DAC, but the same stream sounds a lot better from a cleaner source like a Digi+ Pro fed into the same DAC.... so the DAC is audibly benefiting from the cleaner S/PDIF.
     
  4. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    I talked about this a bit here.

    In general, at least from a clock perspective, using "better" external clocks with a re-clocking DAC isn't going to get you anything. The DAC is going to operate off it's internal clock(s) in either case, and that will be the limiting factor in terms of it's jitter performance. There's always the lovely argument that a better source clock will result in the re-clocking in the DAC itself "doing less work" to "correct" things, but this is a bit of a red-herring - as there's no guarantee that your "better" external clock won't cause MORE drift in relation to the DAC's own internal clock (and thus make the re-clocking do "more work")*.

    With Yggdrasil and Gungnir I have not heard differences in performance between sources that DON'T trigger the clock-mode-indicator (BBGL) that could be specifically attributed to the source clock/jitter. Which is not to say that there can't be other causes to differences in sound - while the differences are tiny, and have to be specifically listened for, the Aries beats the RedNet3 and RedNet3/Mutec combination when driving Yggdrasil in my testing. AND adding a Mutec to the Aries makes things worse.

    Other DACs exhibit different behavior. Bifrost, for example, benefits from a better input clock and is noticeably cleaner sounding if you compare it between a RPi/Digi+ Pro vs. the Aries or something similar. The Spring DAC, despite using it's own clock for everything other than I2S, also benefits audibly from having some of these units in play beyond a basic, isolated, S/PDIF input.

    What I've not done is test very deeply to see how much of the difference between, say, an RPi/Digi+ Pro is either clock or noise (transmitted or radiated) related. And the RPi is a noisy little f**k of a board compared to what you'll find in, say, the full-fat Aries or a microRendu.

    The short version is that, while adding S/PDIF re-clocking ahead of DACs like Gungnir and Yggdrasil might yield benefits, I don't think they're coming from jitter reduction. Which means I'd be inclined to focus on other areas to improve than the external clock.
     
  5. Kattefjaes

    Kattefjaes Mostly Harmless

    Pyrate
    Joined:
    Sep 5, 2016
    Likes Received:
    4,521
    Trophy Points:
    113
    Location:
    London, UK
    (I'm not convinced about the "more work" thing, first refuge of a scoundrel and USB "reclocker" vendors..)

    Good, that was what I surmised from reading Schiit's brief description of what's going on. That's why I was reluctant to drive myself nuts trying to be sure through listening, and just came out and asked.

    The context is that I have been messing with a Gungnir Multibit, unsurprisingly. I plugged in the jittery motherboard optical from my PC (which I usually only use for gaming), and I couldn't actually hear the jitter. I compared it to the Digi+ Pro output, both coax and optical, and couldn't be sure that I heard any difference.

    Obviously, with the Modi Multibit, it doesn't reclock- it seems to just PLL against the incoming data as best it can or something, right? Anyway, the practical upshot is that you can hear jitter if present, really clearly.

    Well, hot damn, so that "Adapticlock" stuff does actually do something.

    Oddly, I'm pretty sanguine about the noise from my Pi. It's powered from a semi-OK APC filtered mains block, and then that funny iFi wall wart, which is decent for a switcher. It's clocked down somewhat and has all the radios disabled. I even did some semi-woowoo things like making the memory clock exactly half of the CPU clock to appease the gods of magical harmonics.

    In the less magical corner, the Digi+ Pro also has a transformer isolated output, and I was able to open the jumper with a small flourish, since unlike the Modi Multibit, the Gungnir Multibit doesn't appear to have floating ground.

    It seemed pretty quiet, even plugged into the Modi Multibit over coaxial SPDIF. I shall solder a 75 ohm BNC onto it, though, since the Gungnir Multibit supports that sort of nonsense.

    Cool, thanks so much for the sanity check. It's a bit weird to have scabby motherboard optical out sounding so good, but it's a problem that I can live with. It seems that @baldr knows his stuff. Clever bastard.

    I will keep the Pi running, though- half the point is to have a player that works when the noisy PC is turned off, great for sitting back and focussing on the music late at night. Queueing up via a mobile device, with a beverage at hand is by far the nicest option.
     
    Last edited: Apr 25, 2017
  6. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    It's more radiated noise that I was referring to with the RPi, (though it is definitely noisier across the board than some of the purpose-specific devices) ... less of an issue for digital transmission (particularly with isolation at play), but it can be an absolute bugger if you, say, actually stick a DAC and output stage on a Hat sitting right on top of the SoC (etc.).
     
  7. Kattefjaes

    Kattefjaes Mostly Harmless

    Pyrate
    Joined:
    Sep 5, 2016
    Likes Received:
    4,521
    Trophy Points:
    113
    Location:
    London, UK
    Yeah, I can well imagine. That said, if you're putting a little DAC directly on your Pi, I guess you're going for convenience above all else.
     
  8. Scott Kramer

    Scott Kramer Friend

    Pyrate
    Joined:
    May 3, 2016
    Likes Received:
    1,448
    Trophy Points:
    93
    @Kattefjaes

    Just FYI In case you missed this on the bifrost (it's PLL mode only), @Torq took a pic and I checked it out a bit:

    Rings of Gold: A Tale of Eight Bifrosts
     
  9. Wfojas

    Wfojas Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    848
    Trophy Points:
    93
    Location:
    CA
    Having gone through this, I can unequivocally confirm the the SU-1 comes behind the Mutecs MC3+ (any software version), and I also can confirm that the SU-1 is probably of any true value with the Holo, due to the I2S. In the Yggdrasil AES it sucked ass, the microrendu matching it to USB. I didn't have the Gungnir Multibit then, but I think the Holo on Coax might beat the mircorendu on a Gungnir Multibit, based on sound signature memory, but that's not saying much.

    Compounding this problem are the unit sound variations with the Mutecs that I've heard (apart from firmware versions). I never kept mine around long enough to want to figure it out, but one can be mushy (the one Marv tested) and one can be fairly pratty (the one Bazelio picked up). The Mutec I used to compare the Singxer with was an MC3+ that sounded like the latter, even with the latest firmware. Not sure why people think so highly of ithe Singxer, or look for salvation in another decrapifier. I could have had a bad sample, but I compared it with one with a custom power supply, and they weren't that far apart.
    Hopefully folks figure out this route is a dead end before they throw any more money away. Get a good dac, feed it a good source. Listen. Save.
     
    Last edited: Apr 27, 2017
  10. brencho

    brencho Friend

    Pyrate
    Joined:
    Dec 18, 2015
    Likes Received:
    7,978
    Trophy Points:
    113
    Location:
    LA
    agreed. in the ideal case, abandon USB, feed good source, etc etc. done and done.
     
  11. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    One major benefit of the SU-1, and similar devices, is that, regardless of any potential effects on sound-quality is that it provides a consistent way to talk to the broadest possible array of DACs. You only need one driver to make such devices work which, when evaluating lots of units, is a major benefit. And while I'm not particularly an XMOS (or Amanero) fan, the drivers for those are generally better behaved (both in terms of performance, e.g. buffer utilization etc.) than the schlocky nonsense that most other vendors provide (had a really good chat with @atomicbob about just this sort of thing just last week).

    If you're a Windows user, being able to switch DACs in and out without worrying about drivers is a much bigger deal than it is for the OS X/Linux crowd. Which means for anyone doing any kind of analysis, or evaluation, of lots of units, such devices are invaluable - even if not used in your normal listening chain.

    While I tend to use my Aries for similar purposes, the SU-1, or similar units, are much more approachable for anyone that rotates DACs a lot or wants to do lots of protracted auditioning and testing.
     
  12. Wfojas

    Wfojas Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    848
    Trophy Points:
    93
    Location:
    CA
    True. But it may not put the device under evaluation in the best light? Your Aries is, of course, a better, and fairer option, I totally agree. In writing this, I thought people might think I was being facetious, or disingenious. I just really didn't like the Singxer, I think, and and at the same time remember what you wrote about the Aries.
     
    Last edited: Apr 27, 2017
  13. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    In the testing I've done, the Aries has been the best interface I've found. That might not be true for others. But it certainly has lower noise and better clock behavior, as well as complete isolation from source, than the SU-1.

    I've not found a DAC that did better via it's USB input than it did via either COAX or AES from the SU-1. That's not to say there aren't any, and I only added the SU-1 to my setup (to get a reasonable I2S interface) late last year, so I've not heard all the DACs I've auditioned with it. It's very hard to imagine a direct computer-sourced USB output beating it though.
     
  14. Johan

    Johan New

    Banned
    Joined:
    Feb 23, 2017
    Likes Received:
    12
    Trophy Points:
    8
    Location:
    Resident troll and general Porker
    have u tried the lynx aes16e?
     
  15. Torq

    Torq MOT: Headphone.com

    Pyrate
    Joined:
    Jan 15, 2016
    Likes Received:
    8,193
    Trophy Points:
    113
    Location:
    Seattle, WA
    See here.

    For my own purposes, even if the Lynx cards resulted in what I thought was the best sound, they're more hassle than I would put up with. But, if you have/want/need a computer in your listening rig then they're a nice option - especially if you can run them directly.
     
  16. earnmyturns

    earnmyturns Smartest friend

    Pyrate
    Joined:
    Sep 25, 2016
    Likes Received:
    3,218
    Trophy Points:
    113
    Location:
    Palo Alto
    Home Page:
    My experience is different. I recently got a SU-1 intended for a (delayed, just arrived) Holo Spring, and while waiting I tried the SU-1 between my microRendu+LPS-1 and Yggdrasil. My wife, who has way better ear than me, immediately noticed the difference -- instruments better delineated against background -- even though she does not keep track of my gear machinations. microRendu+LPS-1>USB>Yggdrasil>Hegel H360>KEF Reference 1+REL T7i vs microRendu+LPS-1>USB>SU-1>AES>Yggdrasil>... .
     
  17. Wfojas

    Wfojas Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    848
    Trophy Points:
    93
    Location:
    CA
    Your reply intrigued me, as I was curious as to the difference in sound between the Yggdrasil and Hegel H360 dac, which I was contemplating as a complement to what I have now. (Holo Spring was nice, mind you, but not nice enough long term)

    The Singxer does alter sound, there is no doubt. But so does the anything in a chain. One can alter sound by changing USB cables, if your chain is revealing enough. I am curious to see what to see the long term value is of that alteration. Not casting shade, just speaking from experience, mind you.

    More importantly, I would have thought a Hegel and KEF Ref 1 would sound clear and detailed? How does it sound for body and presence?
     
  18. ogodei

    ogodei MOT: Austin AudioWorks

    Pyrate
    Joined:
    Feb 9, 2016
    Likes Received:
    2,765
    Trophy Points:
    93
    Location:
    Chicago
    A massive second to this. One problem is that mutiple DAC companies use the same third-party packager for their drivers. So when you go to install a driver for a different DAC you have to uninstall the first one, no way to compare on the same system. And using virtual machines to test audio components never works suitably.

    That and the multiple outputs on the SU-1 make it a great unit.
     
  19. earnmyturns

    earnmyturns Smartest friend

    Pyrate
    Joined:
    Sep 25, 2016
    Likes Received:
    3,218
    Trophy Points:
    113
    Location:
    Palo Alto
    Home Page:
    The H360's internal DAC is a bit too "polite" for my taste: the natural grain of voices and instruments is somewhat smoothed out, providing a sound that is pleasant but distant and unrealistic compared to the Yggdrasil. Complemented with the Yggdrasil, the overall impression is closer to live performance for well-recorded material (I listen to a lot of live music, mainly modern jazz but also classical).

    Yes, Yggdrasil>H360>Ref 1 sounds clear and detailed. With decent or better jazz recordings, double bass and drums are right there, solid and separable, you can place the different pieces of the drum set. For a good orchestral recording, you can place accurately the different instrument groups left-to-right and back-to-front, with correct balance and impact as if you were in good (not too close) seats.

    What the SU-1 did was to reduce some slight background greyness allowing instruments to pop out more clearly, but I suspect that's from avoiding the Yggdrasil's USB path, which even its designer states is less good than AES.
     
  20. Wfojas

    Wfojas Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    848
    Trophy Points:
    93
    Location:
    CA
    Oh, great post! I suspected that part of the Hegel integrated raves were the amp section, where it let through details and microdynamics leading to a life like space. if you have a chance, try the Aries and the Lynx just for comparison. These aren't cheap, but it does put the Singxer into context. Your stuff is certainly good enough to merit trying. My chain is microrendu->Berkeley Audio Alpha Usb -> AES->Yggdrasil->preamp->amp->speakers, so I don't USB direct, either.
     

Share This Page