HiFiBerry Digi2 Pro Impressions

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by Woland, Mar 30, 2021.

  1. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I hate leading people down this rabbit hole because these kinds of tweaks very personal and may or may not work for others. This is how I did it.

    PXL_20210516_165028938.PORTRAIT.jpg

    Doing it this way instead of BNC-to-coax or coax-to-coax does sound better. To me, it's discernible enough where I am pretty sure I can pass a blind test after some practice. Using the pi2AES AES inputs results in a less grainy, less thin, less gritty, smoother, and more mellow "analog" sound. "Analog" meaning more similar to my TT source.

    Here is my chain. Too many headphones to show. (Ignore the Soncoz DAC - that's there for evaluation). YMMV.
    PXL_20210516_170308829.PORTRAIT.jpg
     
    Last edited: May 16, 2021
  2. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    Would you have any hypothesis about why that might happen? I understand that cabling might impact on jitter, but I wouldn't have expected minor jitter to manifest like that.
     
    Last edited: May 16, 2021
  3. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    No hypothesis. Same thing as cables. Why do cables of different material or same material but different geometry sound different? Some people even feel that power cable sound different. I guess you can argue that the approach I took was essentially giving the Gungnir a balanced AES input. FWIW, I also felt Yggdrasil AES input was better than its SPDIF input using the same source.
     
  4. RestoredSparda

    RestoredSparda Friend

    Pyrate
    Joined:
    Jan 7, 2021
    Likes Received:
    3,638
    Trophy Points:
    93
    Location:
    WI
    Trying not to crap up this thread, but I have to agree with you on the benefits to using AES out from Pi2AES. (In my case AES from Pi2AES -> SESCOM adapter -> BJC BNC to Coax cable -> Bifrost 2 -> Lyr 3 -> HD650 KISS w/ brand new pads).

    Compared to BNC from Pi2AES - sibilance is reduced a bit, overall more full sound (less thin), slightly blacker black-ground, a hair more plankton, and as a result, a better sense of space in recordings.

    These are all things I tend to enjoy, as I prefer a warm, while still clear and detailed, sound. Any edge or glare in the highs tends to set me off. This is the most "listenable" my terribly mastered albums have sounded, and the best my well-mastered albums have sounded.

    Thanks for the suggestion! Total spend was about $44 for a short AES XLR cable and the SESCOM adapter. I've spent more on regrettable tubes.
     
    Last edited: May 19, 2021
  5. dllmsch

    dllmsch Friend

    Pyrate
    Joined:
    Jan 3, 2016
    Likes Received:
    280
    Trophy Points:
    63
    Location:
    Molvanîa
    Holy shit it works.
    Since I'm too cheap to buy a sescom I built my own AES to Coax adapter per @gixxerwimp 's guide here
    https://www.superbestaudiofriends.o...ehringer-deq2496-digital-eq.9295/#post-306727
    [​IMG]
    [​IMG]
    Much tighter bass and a bit better sense of depth right off the bat listening to some daft punk.
    Similar impression as @RestoredSparda with BF2.
     
  6. gixxerwimp

    gixxerwimp Professional tricycle rider

    Pyrate
    Joined:
    Sep 30, 2015
    Likes Received:
    5,773
    Trophy Points:
    113
    Location:
    small island claimed by China
    Glad it's working for you. Mine looks better though :cool:
     
  7. mtavares

    mtavares Acquaintance

    Joined:
    Apr 7, 2019
    Likes Received:
    70
    Trophy Points:
    18
    Location:
    Brazil
    @purr1n
    I would like to inform you that the tweak you mentioned worked for me also. I know your concern to share such type of information, but I consider interest to inform. Thanks for sharing.

    The results were the same described by you, @RestoredSparda and @dllmsch. I would like to emphasize that it has a little more bass, a little more "warm" to the sound.

    See below the configuration used:

    [​IMG]

    Previously I was using a Black Cat Veloce BNC cable.


    Sorry for the off-topic.

    Regards,

    QJA

    Marcos
     
    Last edited: May 31, 2021
  8. AukePauke

    AukePauke Acquaintance

    Joined:
    Jul 23, 2017
    Likes Received:
    33
    Trophy Points:
    18
    Location:
    Netherlands
    I can't find a sescom in Europe, so i ordered a CANARE BCJ-XJ-TRC. It gives a little more space around the different instruments, a little more dept and its just a little less dark. Not more bass. Its not earth shattering, but for now i keep it.
     
  9. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    Just to balance out the past few comments, I have continued switching between sources (Pi4AES through AES, Sescom, BNC and Pi3 HiFiBerry Digi2Pro coax) into a Gungnir, as I tweak the Volumio configurations.

    I observe no differences at all between coax, BNC and BNC via AES. That's what should be expected from any clean digital connection into a reclocking DAC.
     
    Last edited: May 31, 2021
  10. AukePauke

    AukePauke Acquaintance

    Joined:
    Jul 23, 2017
    Likes Received:
    33
    Trophy Points:
    18
    Location:
    Netherlands
    I decided to not use the Canare anymore. The symbals where to sharp, not natural on some recordings.
     
  11. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    LOL, maybe you don't observe any differences because you do not expect any? With respect to connections, a transformer solution will isolate grounds. This would be true galvanic isolation. Told you this was going to go down a rabbit hole where even different digital cables (or power cables) make a difference! I'm not going that far.

    P.S. Try comparing to Toslink optical. In theory, that should be even better since there is complete electrical isolation. Yet, it sounds like ass.
     
  12. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    I hope this will be one of the first things you test with your new toy! Jitter is one of the areas where simple measurements should be sufficient to settle the issue definitively.
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Hmm, I can already do that with the AverLAB.

    I'm going to echo @atomicbob and restate that we really need to get around the really super duper dumb notion that an APx555 is even necessary.

    People cannot hear distortion or noise as "high" as -85db on an analyzer. Heck, Amir couldn't even hear -35db THD, most of the "nasty" third-order, from the JBL plate amps from a monitor that he subjectively said was amazeballs.

    BTW, I get better JTEST results from the AverLAB digital outs than the APx525. This was based on some measurements I took of Modius and JNOG. I don't know about the APx555.

    P.S. A lot of what we hear can be seen in measurements. A lot of what we hear isn't seen in measurements, a lot of stuff can't be explained from any measurement (at least not yet), and one single measurement, i.e. jitter, hardly explains everything. I've heard lower jitter result in sharper more focused sound. I've also hear lower jitter result in more organic tubier sound. Jitter is also very 2012. It's so rarely a problem now. Well, unless -132db sidebands cause you grief.
     
    Last edited: Jun 9, 2021
  14. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    AVERLAB to Gungnir A2 (XLR output measured)

    RED = toslink out -> toslink out
    BLU = SPDIF out -> BNC in
    GRN = AES out -> SESCOM -> BNC in

    upload_2021-6-9_12-14-11.png

    In theory, we cannot hear any of this (I normalized a -6db 12kHz signal to 0dbFS - we can't stuff 110db down from that). I don't have the Unison USB board on the Gungnir, but I'm pretty use it would measure better than any of the below. Unison has been spectacular with respect to jitter / sidebands. However, I prefer the sound of AES from a good source over Unison.

    Now if I wanted to pull something outta my ass, on the hypothesis that these primitive (and they are very primitive) measurements may allude to other behaviors (observed subjective phenomena) not yet measurable (i.e., someone hasn't invented the measurement yet).
    • Toslink sucks ass. There are two big sidebands - too high and too far away, near the edges of the graph. There are also two smaller sideband pairs. The result is this: sounds like ass.
    • BNC is OK. Lots of little sidebands all over the place, but not anywhere nearly as high or far away as the big Toslink sidebands. As a result, the sound is more grainy, more disjointed compared to the below.
    • AES-adapter-BNC is arguably "better" because its sidebands are close to the fundamental and low in level, on par with the BNC results. Maybe there's some smearing to the sound that makes it smoother or warmer. However, one could prefer the BNC on the account that the sound may be more focused.
    Would I even stake $10 on my hypothesis. Nope. There's just not enough information. However, could it possible that what the folks have observed about an AES-> BNC adapter having a different sound over the other inputs is not imaginary? Yep.
     
    Last edited by a moderator: Jun 9, 2021
  15. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    Thanks for taking a look @Marv !

    I don't know who it is that cares about APx555 etc that you/Bob are addressing. But it sounds like you're going to be spending more time focussing on measurements once it arrives, so figuring out a panel of standardised analytics for comparing sources (or source+DAC combinations) will be appreciated.
     
    Last edited: Jun 9, 2021
  16. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    Not problem at at!

    I'm too simply too curious. It's difficult for me to let possible answers to good questions lie buried under a rock, so I pulled the Gungnir out of the rack to do these. (A lot of the time, we get no good answer, so we got maybe sorta luck this time.) This is the essence of science. It's time to reclaim the word science and objectivity from those who have misused and twisted the meaning behind those words into hard dogma and religion.

    Now here's the scary stuff: USB, even "mediocre" USB solutions, tend to measure much better than SPDIF/BNC or AES. It's weird. There's a disconnect here on SBAF because many people tend to prefer Coax/BNC/AES over USB. Maybe there's stuff with USB that isn't seen in the measurements. Maybe the APx555 could tell us more? Maybe not?
     
  17. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    We need more of this kind of curiosity about music!

    A related thing I'd love to see tested: in-DAC filters vs external resampling. This is not about buying another piece of gear, it's about potentially getting better SQ from an existing system.

    For example, if you feed your Gungnir 44kHz, it uses the Combo Burrito filter. If you resample to 384k using Roon, Volumio, Audirvana, HQPlayer etc the Combo Burrito filter is not used. Are any of these improvements? There are a few hardware options on offer too.

    @GoldenOne has teased a video about HQPlayer.. I'm really looking forward to seeing that and the discussion that generates on filters/resampling/upsampling etc
     
  18. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    You mean like the standard test suites found in the technical measurements

    here:
    https://www.superbestaudiofriends.org/index.php?forums/headphone-amplifier-measurements.13/

    and here:
    https://www.superbestaudiofriends.org/index.php?forums/source-measurements.14/
     
  19. Woland

    Woland Friend

    Pyrate
    Joined:
    Jan 13, 2021
    Likes Received:
    1,311
    Trophy Points:
    93
    Location:
    a friendly land
    Yes! I was thinking of something comparable to the standard test suite on headphone amplifiers and DACs which I've seen many times and appreciate! The only difference is the interest is more on the digital stream and in varying the transport and processing prior to the DAC chip.
     
    Last edited: Jun 10, 2021
  20. Marvey

    Marvey Super Friend

    Staff Member Pyrate BWC
    Joined:
    Jun 6, 2021
    Likes Received:
    1,712
    Trophy Points:
    93
    Analysis like this can always been requested! It's not that I will be focusing more on measurements going forward. (Sure, the AP will help me a lot in this regards.)

    It's more that people have become complacent, less curious, and want to be spoon fed. Back in the Changstar days, I would get challenged a lot more, or get requests of can you explore this measurement or that measurement.

    This doesn't seem to happen as much these days. It most certainly doesn't happen on ASR. I do get cranky, but I do in the end appreciate being pushed to explore more

    There is lots of explore, so if I don't, it's a time issue.

    These days, I feel I could just wing it, plot a random graph, and write a few lines of recommended or not recommended with a pose from my skullpyrate, and people would actually buy this crap.
     

Share This Page