Soncoz SGD1 ASR Measurement Discrepancies

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by GoldenOne, Jun 17, 2021.

  1. mitochondrium

    mitochondrium Friend

    Friend
    Joined:
    Sep 1, 2017
    Likes Received:
    828
    Trophy Points:
    93
    Location:
    A Cell
    If I had to chose someone to score a vital penalty in a soccer match, would I chose an amateur with nice new shiny boots or CR7 with a shabby 15 yo pair of shoes.

    And the answer is….

    Even if the amateur had an instruction with him how to execute the perfect penalty that would not change my choice

    Expertise beats equipment any time

    Sorry for the soccer analogy I am not very familiar with popular US sports
     
  2. GoldenOne

    GoldenOne Friend

    Friend
    Joined:
    Jul 17, 2020
    Likes Received:
    826
    Trophy Points:
    93
    Location:
    England
    Attached is a full meaasurement report for the LA-QXD1 and the SGD1. Fed by galvanically isolated USB (ifi iGalvanic). Balanced out. Config is shown in the report but most measurements were done at 0dbfs and/or -3dbfs

    AudioQuest Mackenzie XLR cables used

    VMV D1SE and Chord Mojo also included for comparison:

    SGD1: https://thegoldenone.co.uk/Soncoz SGD1.pdf
    LA-QXD1: https://thegoldenone.co.uk/Soncoz LA-QXD1.pdf

    SMSL VMV D1SE: https://thegoldenone.co.uk/SMSL VMV D1SE.pdf
    Chord Mojo: https://thegoldenone.co.uk/Chord Mojo.pdf

    If anyone spots any config issues please do let me know, but report should include just about everything about the config so that it is completely transparent.
     
  3. penguins

    penguins Friend, formerly known as fp627

    Friend
    Joined:
    Sep 26, 2018
    Likes Received:
    2,749
    Trophy Points:
    93
    Location:
    SouthernCA
    If we want to go over bad measurements, we could waste a lot of time making a thread that shows in simple terms, complete with a 1 line conclusion, easy to follow pictures, “scroll to here to see super best results”, and everything else “science” as an ad can show - how easy it is to obtain different measurements.

    Just on the fly, it could include the following:
    1. Build a cherry picked simple DIY project and then build, maybe multiple, “this is what you’d probably get off the shelf” version/s of the same.
    2. Buy 10 or however many pieces of inexpensive components XYZ produced over multiple batches or time and show variations in measurements. Inexpensive parts bc that’s what you usually get in $100 gear.
    2b. Have a few of these components interact with each other. Swap at random and repeat to show how individual units may vary.
    3. Show the importance of measurement device settings some more. And how this can be used to majorly alter results.
    4. Show how much measurements can deviate with the same person and gear just based on “time of day”. I.e. repeatability and maybe why we should care less at -120db.
    5. Show how single changes in hardware used can alter (or not alter) measurements. Can be something bad from a “wrong” cable , wrong probe, dirty power or just something neutral like a “different” XYZ. Again, shows why we should care less after a certain threshold.
    6. Have 3 or 5 or X different people measure the same gear and show variation in results. This would be akin to students in electronics lab getting different results. Even better, have someone do it on quick auto pilot and have someone do “a good job”. This would show how for something simple it may not matter who does it why for more complex gear people can make a good living “doing measurements”... or how much one person can bork results.
    6b. Have the people range in education and experience if possible and safe. One is a high school or middle school student (supervised only for safety), one random college student or random person off of an audio forum, one is a lab technician where they’re exposed to these things but maybe not necessarily all of the science and engineering behind what they do, one is a a master level professional who measures stuff for a living. Etc. See how many “minor oversights” (such as insane volume levels) occur at each level.
    OR
    Don’t do any of the above. Those who want to believe will believe anyways. For everyone else let people fall a few times and hopefully learn from the consequences and mistakes of their bad ASR influenced purchases... And then go do something enjoyable instead ... and maybe actually listen to music.
     
    Last edited: Jun 29, 2021
  4. GoldenOne

    GoldenOne Friend

    Friend
    Joined:
    Jul 17, 2020
    Likes Received:
    826
    Trophy Points:
    93
    Location:
    England
    Figured it'd be worth sharing.
    To preface: I am NOT accusing Amir of anything or implying that he's done anything wrong. Just showing another example of where the retail version of a product is not performing to the same standard as the unit amir measured/published.

    This time it's the SMSL SU9.
    Full report is attached.

    SINAD itself is not too different, though the actual harmonics on Amir's unit are much lower. His are all <-130dB (which he prominently labels on the graph), on this unit, even at -3dBfs there are harmonics at about -125dB.

    MzSnZn9JDm.png

    The biggest issue though is the IMD hump.
    Amir's unit showed 'the slightest hint of IMD hump':

    SMSL SU-9 Balanced USB DAC IMD Distortion versus level Audio Measurements.png

    Whereas on this unit it's very prominent:

    upload_2021-9-3_21-15-45.png


    32 tone test (using the wav file Amir posted) is also 5-10dB worse in places.

    [​IMG]
    upload_2021-9-3_21-17-17.png


    Nothing 'major' other than the IMD hump, but still. When products can lose 10 places in the asr chart on a 2-3dB difference, this cherrypicking can make all the difference.

    Honestly when a site is as large and influential as ASR, it's to be expected that manufacturers will likely send cherrypicked units.
    With the pull Amir has, I can't imagine it'd be a problem at all to say
    "Instead of sending me a unit, I'll purchase one from a random retailer and you reimburse me".

    That way the manufacturer still gets the publicity they want (assuming the retail units are actually good), and you know that the unit being tested isn't a unicorn.
     
  5. 7seven

    7seven Acquaintance

    Joined:
    Oct 28, 2018
    Likes Received:
    33
    Trophy Points:
    18
    Location:
    Ireland
    According to him Topping offered to do this some time ago but he declined, he didn't have the time and he trusts them (????), really bizarre excuse if real (gotta love the 'quantity over quality' approach to scientific research), does not explain why this is not the procedure for every new Topping review since then.
    If he does have some connection to topping it sure is a nice anecdote to have on their flagship product review, makes topping out to be 'good guy' without them actually doing a thing.
     
  6. purr1n

    purr1n Burned out

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    81,290
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    You are far too polite (as is my experience with folks over the pond) - unless you intended this to "For Brutus is an honourable man" a la "Friends, Romans, Countrymen..."

    Anyone who positions himself as a scientist-savoir: it's their moral duty to investigate.

    I am fine by the methods of which Amir obtains gear. It's a matter of being practical and pragmatic, simplifying the workflow, the logistics. I can empathize because I have a stack of gear sitting next to my desk yet to be reviewed. One additional thing to do makes dealing with the backlog that much more difficult.

    However, we have seen several incidents now. Not only this DAC, the Soncoz, and another from SMSL. If this product were froim a western company, a company who decided they didn't want to sell to Madrona Digital, or a company whose marketing Amir disapproved, or perhaps a product in which Shenzhen Audio was NOT the middleman, Amir would be all over it.
     
  7. purr1n

    purr1n Burned out

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    81,290
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    @GoldenOne: do you still have access to this DAC - the Soncoz SGD1.

    Evidently Amir followed up on a Reddit post pointing out the measurement disparities - and found out that the SGD1 has firmware that can be upgraded via WiFi. There's some procedure to set a phone's wireless access point so that the SGD1 firmware can be updated. SONCOZ has posted an upgrade guide here: http://www.soncoz.com/uploads/1/2/7/2/127246176/sgd1_upgrade_guide.zip

    Supposedly, the specific units that @Bill-P and you received did not have the proper firmware installed, hence the poorer SINAD. I figure if we are still able, to confirm that there is a working firmware fix.

    Evidently the measurement discrepancies (from ASR measurements) both you and I discovered on separate units are being labelled as "alternate facts."
     
    Last edited: Sep 20, 2021
  8. GoldenOne

    GoldenOne Friend

    Friend
    Joined:
    Jul 17, 2020
    Likes Received:
    826
    Trophy Points:
    93
    Location:
    England
    I did try to update the firmware whilst it was here so I'm fairly certain it was on the latest (that or the installation didn't work, the update process didn't give a huge amount of info :p).
    But it's good to see that Amir did actually re-measure it.

    I don't have access to this unit anymore. The owner is in finland so getting it shipped back here and then back to him will unfortunately be fairly pricy (and I do not know if he'd be willing to do so anyway).


    In terms of the situation:

    - The units Bill-P and I measured were genuine retail units, if a firmware update was released AFTER the measurements were posted (is there a confirmed date on when the firmware was released?), or to be honest, the fact that the device performed that way out of the box anyway, means that it is completely genuine to point out the disparity between the performance of a retail unit and what Amir showed. Any attempt to spin this as me/anyone else 'attacking' amir or something is ridiculous. This is a genuine point of issue, and Amir has now addressed it, which is great. Further drama is pointless.

    - Even if the performance of the updated unit is correct (which if it is that's great), it's still lower than what Amir originally tested and the difference would still push it out of the top 20 by Amir's ranking standards. So it doesn't really change the core issue I was pointing out: Amir's measurements did not line up with retail units.
     
  9. MaxBuck

    MaxBuck Rando

    Joined:
    Sep 20, 2021
    Likes Received:
    14
    Trophy Points:
    3
    Location:
    La Quinta, CA
    I'm not sure disagreement between testing results of different equipment samples is an indictment of anyone other than the manufacturer, but I sure appreciate the information provided in this helpful thread. I'm disinclined to purchase Soncoz as a result.
     

Share This Page