Schiit Yggdrasil Less is More (and MIL and OG) Impressions + Measurements

Discussion in 'Digital: DACs, USB converters, decrapifiers' started by purr1n, Sep 9, 2021.

?

Should SBAF get a loaner Yggdrasil Less Is More?

  1. Yes, please!

    75.4%
  2. Only if there isn't anything else more interesting

    11.5%
  3. No, I would prefer for a loaner an overpriced planar that looks like it's from House Harkonnen

    6.2%
  4. Save up for something better

    6.9%
  1. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I think most of you guys by now know that the guys at Schiit cook up all sorts of stuff, most of which never makes it to market. And two-thirds of the stuff that eventually makes it to market goes through successive iterations and actually makes to market something like 16 years later. Jason and I have a joke going between us. It usually goes like this:

    When will it come out?

    In a few months, toward the end of the year.

    Oh, you mean seventeen months from now.

    You are probably right.

    Anyway, I did have the privilege of being able to listen to a prototype DAC using the DAC8812 chip, x2 of them in hardware balanced configuration. No, this wasn't the great thing in the world defeating many other DACs ten times its price. It wasn't the most resolving. It didn't slam (at least not as much as the Bifrost 2, Yggdrasil 2 - henceforth known as "OG"). However, it did bring something new to the table: very even tonality, neutral perhaps, smooth highs as would be expected but doesn't always happen for an R2R DAC, and perhaps the nicest most correct timbre that I had ever heard.

    I asked Mike Moffat (designer of Schiit DACs) all sorts of questions about this DAC: Can I wire in a proper LPS, can I do this, can I do that? I bet this would be really good if.. Can you Yggdrasil-fy this, that is build a high-end version around this, with all the bells and whistles that the Yggdrasil offers: choke PS, crazy multiple shunt regulators, etc. It turns out that Mike had already thought of this.

    (This is why I cryptically mentioned a few times "good things happen to people who wait" in the past several weeks for folks who were asking me about a more neutral, less colored, less organic sounding DAC to Yggydrasil OG which resolved and soundstaged - not to be confused with headstage - like Yggdrasil.)

    5F4A0028.JPG
    I only have one DAC. It's left open so I can swap boards.

    Well, the Less Is More - LIM - is the extension of this prototype (let's call it development idea) to a flagship level product. The end product is very similar to what I had heard from the little prototype box. The tonal signature remains very even. The LIM doesn't have the OG(A2)'s organic lows and slightly emphasized top octave. The LIM doesn't have Gungnir A2's midbass reticence or forward upper mids. The LIM doesn't have an emphasis or recession in any one area of the spectrum. It does however have a very slight dark tilt. It's also slightly harmonic richer sounding than the prototype, but less rich than the Gungnir and Yggdrasil.

    Everything else in terms of technicalities is raised quite a bit from the development idea. Proper bass slam is now restored - as good as the rest of the Schiit DAC lineup > $700. Bass articulation is better than Yggdrasil. Clarity is better than Yggdrasil and Gungnir. The macro-dynamics are on another level to earlier Schiit DACs. I feel the 8812 chip has been unleashed by the Yggdrasil motherboard (which really is the power supply for all the components). The SE and BAL outputs sounded more or less the same (trying to take into account other unavoidables which makes a true apples-apples comparison impossible). Retrieval of low level information is behind Yggdrasil A2/OG, maybe on par with the current Gungnir (maybe even somewhere in between Gungnir A1 and A2 - still no slouch). If you are a microdetail whore, you may miss this. I consider myself a resolution whore; and TBH, I didn't miss it that much because it was a matter of the totality of the parts. This is why I have to say that my preference for the LIM must be caveated with other considerations. (BTW, most people need not worry about the microdetail / plankton thing - they don't have systems resolving enough - this goes for many https://www.ORFAS.org audiophiles I know.)

    And no, it's not matter of new toy syndrome. I've had the DAC8812 development idea in my house for what seems years now. (TBH, I don't know how long because the Xi-virus has skewed my sense of time). I've been waiting for this for years. I also didn't know the announcement was being made today. People think I know everything that goes on at Schiit. I really don't. (FYI: I'll be heading back to cybersecurity in M&E. Thanks to Amazon, Netflix, and Apple pushing up salaries in the biz.)

    P.S. Pi2AES AES out used for impressions. Unison results in a warmer sound.
     
    • Like Like x 33
    • Epic Epic x 4
    • List
    Last edited: Sep 11, 2021
  2. k4rstar

    k4rstar Britney fan club president

    Pyrate BWC
    Joined:
    Jun 11, 2016
    Likes Received:
    6,941
    Trophy Points:
    113
    Location:
    Toronto, Canada
    So it sounds like a bifrost 2 all grown up? more resolution, better transients, correct tone with slight darkness?
     
  3. Merrick

    Merrick A lidless ear

    Pyrate
    Joined:
    Jan 6, 2016
    Likes Received:
    12,569
    Trophy Points:
    113
    Location:
    Portland, OR
    Coming from a Gungnir A1 fed by Pi2AES, it sounds like LiM would be an improvement across the board? Anything I’d be missing if I went from Gungnir Multibit A1 to LiM?
     
  4. EagleWings

    EagleWings Friend

    Pyrate Contributor
    Joined:
    Jul 20, 2017
    Likes Received:
    1,709
    Trophy Points:
    93
    Location:
    India
    Thanks Marv. Does the LIM need long warm up time like the AD chip Yggys?
     
  5. sphinxvc

    sphinxvc Gear Master (retired)

    Staff Member Pyrate BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    3,320
    Trophy Points:
    113
    Can't tell from what you wrote, @purr1n, do/did you have the LiM production board?
     
  6. el Chefe

    el Chefe New

    Joined:
    May 2, 2020
    Likes Received:
    15
    Trophy Points:
    3
    Location:
    NJ
    Ohh I might need to try this!
     
  7. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The LIM took less than a day to warm-up. I had always joked that the time required for a chip-based resistor ladder to warm up was Constant*2^(bits). Yes, the LIM is based on a 16-bit part, two of them. So less than a day. it sounds muddy and congested at turn on. Sound is actually good within 4-6 hours. After a day, totally clears up.

    Wait, only 16-bits? That's horrible! if you are an audio scientist or ardent MQA-advocate such as John Atkinson, you may protest: that's not state-of-the-art! Well, how much music do you got has over 16-bits? And how much of that "24/192" stream on Qobuz sounds discernably more resolving (assuming from the same master) than CD-quality. I'm not saying bit-depth isn't important. I do have true high-res recordings where if they are sampled down to CD-quality, it's noticeable. However, these recordings are far in between and not necessarily the stuff I really want to listen to on a daily basis. Besides, with 16-bit parts, it's still possible to get betterer than 16-bit resolution. We will examine this later and verify via measurements if I think I know what Schiit may have done (not too different from the sorcery they performed on the Modi MB).

    Enough with the technicalities? What about the overall picture and feel? The most attractive quality about the Yggdrasil LIM is that is does not draw attention to itself - except that is does surprise every now and then. This is not the Kate Upton gyrating in a bikini DAC. Or layered chocolate, fudge, cream, chocolate cake DAC. This is the anti-social media, anti-Kardashian of DACs. No drama, no theatrics, no nonsense. This is this person you marry, wake up with everyday, and throughout the day you notice little things catch you off guard, surprise you, and make you appreciate. Going between the Yggdrasil OG(A2) and LIM - these differences in overall sense were immediately noticeable.

    Think of the LIM as the DAC version Gaudio Nair IEM where the sum of parts were greater than the whole, except better. Crinacle rated the Nair an A for tonality and B for technicalities. If the LIM were the Nair of DACs, then that technicality score would be A-.

    I think you know that these days I am super hesitant to recommend that people "upgrade". I do think it's an upgrade across the board. Definitely more clarity, blacker backgrounds, less grey. More natural highs even. Though it has that slightly dark tilt, and a bit of richness, you may miss the outright warmth and richness of the Gungnir A1. A switch from the SPDIF to Unison could restore some of this though. Lots of options. Like the BF2, both SE and BAL outputs sound the same and sound good.

    I would say resolution about on-par with Bifrost 2. Less thick, less rich, better clarity. Tonally I wouldn't associate the LIM with most R2R DACs. The LIM isn't especially special with respect to transients. I wouldn't call it fast or slow. However bass articulation, clarity, and texture is better than all prior Schiit DACs. What I do really love about the LIM is the timbre. Schiit really nailed it. Usually good tonal / timbre is faked a bit on many multi-bit type DACs with warm tone, organic lows, rolled highs (not that any of these things are bad - they do work in the right system). Think good delta-sigma DAC sensibilities without those awful highs. Think a better version of the Analog Devices parts in the Convert-2 or Lavry DA11. In this sense, I do think the LIM is highly suitable for sound work or as a "reference" DAC (whatever the F "reference" means, but long time SBAF members know all about that).

    It's a production first article board with all the parts put on by robot in an oven. The PCBA houses make a few, and then ask if it's OK before they make hundreds or thousands more. I haven't had these boards for very long. I actually had to FedEx the MIL boards down to the Schiitr for the thunderdome level-matched double-blind test.
     
    • Like Like x 13
    • Epic Epic x 3
    • List
    Last edited: Sep 9, 2021
  8. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    So this hangs with the Burl Bomber and Dangerous Convert 2, both of which go for some intentional distortion?

    What about the “clean” pro converters that just don’t care that much about slight harmonic distortion if resolving power and clarity are superb, eg Apogee, Prism, and Crane Song?
     
  9. Merrick

    Merrick A lidless ear

    Pyrate
    Joined:
    Jan 6, 2016
    Likes Received:
    12,569
    Trophy Points:
    113
    Location:
    Portland, OR
    Thanks! You’re right that I shouldn’t have used the term “improvement across the board” as that’s way too broad of a brush. I was trying to triangulate if this is a good fit for me based on my knowledge of your preferences and descriptors, and my knowledge of/preferences between the Gungnir Multibit A1 and BF2. I would be willing to lose some of the warmth/richness of the GA1 to gain the blacker background/less grey. I appreciate it!
     
  10. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Burl B2 Bomber remains one of my favorites because of their discrete output stage and ability to dial in saturation distortion at hotter output levels. It's not anywhere as colored as their marketing makes it out to be. The great thing is that they are niche and nobody, well especially audiophiles, know who they are so anyway can still grab one of their B2 DACs with the AKM 4399 which I think in hindsight is AKM's overall best chip ever. Better than the VELVETY 4490 going ever brighter and brighter up to the 4449. Not that it matters now. Just a shame that AKM made their latest and greatest high SINAD DACs either mushy poo with delta-sigma nasties on the low-end or bright as F with delta-sigma nasties on the high-end. Takes good DAC designers to work around this and why the Topping implementations sound like ass. I should grab a Burl before their stockpile of ancient 4399 chips runs out.

    Dangerous Music Convert-2 was so so promising; but ultimately super hot outputs and the use of only a single Analog Devices chip did them in. The lack of nuance, lack of plankton and microdetail, made more than one early FOTM adopter eventually sell theirs. (Yes, I notice these things). It's not for any lack of warning on my part. The soundstage was weird, but headstage worked well.

    Crane Song has their custom filter with that awesome soundstage and headstage. However tuned too bright and too clean and still has heavy vestiges of AKM VELVET nonsense. The bright and clean can be addressed with downstream gear, but the VELVET cannot. If the above two (Burl and DM) were niche among audiophiles, then Crane Song is even moreso. Solaris DAC with AKM is dead. The guy at Crane Song needs to rejigger for ESS now. Curious if he will be able to work his soundstage magic with the ESS which tend to suck at soundstage.

    I should check out Apogee Groove or Symphony. If they are like anything they were like before, they should have a house sound. Hard to say. They seem to making shifting their business for at-home artists instead of the higher-end stuff that supposed to be shoved in racks of machine rooms at studios. I should get a Prism in for loaner because they seem to be lowering their prices on their consumer stuff every time I check. Tough business. (I'm not a fan of the Prism sound - sounds like someone ran a plug-in filter).

    Bottom line is that all DACs, even pro DACs have their own sound. Heck, we all have our own individual forms of hearing damage or hearing loss. It's what we do with it and how we work around it. I'd still trust deaf rock drummers turned sound engineers than younglings with their Toppings on their desks when it comes to gear.

    Even high SINAD DACs have their sound. High SINAD doesn't mean not colored. A high SINAD ESS sounds different from a high SINAD AKM. A high SINAD AKM from Geshelli sounds different from a high SINAD AKM from Topping sounds different from a high SINAD AKM from Shanling.
     
    • Like Like x 10
    • Epic Epic x 3
    • List
    Last edited: Sep 10, 2021
  11. GoodEnoughGear

    GoodEnoughGear Evil Dr. Shultz‎

    Pyrate
    Joined:
    Oct 25, 2015
    Likes Received:
    3,070
    Trophy Points:
    113
    Location:
    Cape Town, South Africa
    Looking forward to seeing comparisons roll in. Sounds like LIM may be in the DAC2541 kind of space, but a tad darker/mid focused and a step up in tonality and timbre?
     
  12. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So what about the More Is Less? Depending who you talk to, it sounds like ass or it sounds pretty good. I don't have the boards in house and TBH I'm not too worried. I'd be much more depressed if either the OG or LIM boards were not around. My impressions will be short: the MIL isn't bad. There's no such thing as a bad R2R ladder DAC (within reason). I didn't have as much time with the MIL so take these impressions with a grain of salt:

    The MIL would be a good DAC in the absence of the other two DACs. In a way, it's "referencey" than either the OG or LIM because its sound a bit more coarse and delta-sigma. This for lack of better words because it still has R2R sensibilities. There's is a bit of warmth too, so it's not lean sounding either. When it comes down to it, I didn't think the MIL was as musical as the other DACs. By musical, it wasn't a matter of it not being as harmonically rich, or as liquid, as organic, as the other two. By not as musical, what I mean is dull. Dead. Unengaging. If would call this the Schiit Yggdrasil Heresy edition because it reminds me of the Heresy amp. A nice warm precise sound with extended highs - but dull - something I would rather not listen to if I had something better.

    I'll listen a bit more when I get the boards back. Maybe I will change my mind. Maybe with a one week warmup, it will sound better. In the meantime, I can see that those folks used to run-of-the-mill DACs using AKM and ESS chips may actually prefer the MIL because it's more similar in presentation. Also, another reason to get the MIL is because of it's shockingly good measurements SINAD. (Going forward, I think we need to be smart about measurements. SINAD is really one thing to itself).

    [​IMG]
     
  13. rhythmdevils

    rhythmdevils MOT: rhythmdevils audio

    Pyrate
    Joined:
    Apr 15, 2020
    Likes Received:
    12,239
    Trophy Points:
    113
    Location:
    Bay Area, CA
    Home Page:
    Hmmmmm resolution or perfect timbre and reference. Can I stack the two on top of each other?
     
  14. Cellist88

    Cellist88 Friend

    Pyrate
    Joined:
    Nov 11, 2015
    Likes Received:
    1,607
    Trophy Points:
    93
    Location:
    NJ
    Yggdrasil A3...
     
  15. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Part but not all of it could be the slightly emphasized top octave of the Yggdrasil. Not too different from how the HD800 is super resolving. Whereas the dark tilt of the LIM could hurt it in this regard of perceptual plankton. Schiit stacked 4 DAC8812s, x2 per differential signal. I guess they could stack 8 or 16!

    --

    Many moons ago I recall speaking to Mike Moffat about DACs and glitch. Glitch is an issue with R2R DACs. Take the initial iteration of Modi MB where we saw a small spike upon the zero-crossing with certain kinds of signals. A few years ago, I wrote a some code to output and analyze a wave file bit crawl of the 5 LSB (basically triangle wave) and discovered a zero crossing glitch: https://www.superbestaudiofriends.o...easuring-dac-accuracy.4184/page-3#post-131301

    It's like when we turn on the vacuum cleaner and hairdryer in an old house causing a current spike the makes the lights dim momentarily. This happens with resistor ladders because each bit flip causes a latch on or off the power line. The more bits that are flipped for each sample, the more severe the glitch behavior. For example, 1111 (15 decimal) to 0000 (0 decimal) would exhibit much more glitch than 1000 (8 decimal) to 0000. Fortunately, from what we can gather from the examples, the really bad cases only happen with specific changes in signal. It's really not a zero crossing deal with real musical signals; however, it's a different story for triangle, sawtooth, and sine waves where we have -1 to 0 if the DAC uses all bits on for the negative number just below zero.

    Going back to the MIL which uses the DAC11001, I noted in the datasheet features a "code independent low glitch: 1 nV-s", that is a deglitching circuit. I'm am guessing this has been left on (by default it's on) to arrive at those spectacular measurements. This is described as a "track-and-hold" circuit before output of the chip. It's basically a technique also known as "sample and hold". So getting back to the old conversation:

    What do you think about adding a sample and hold circuit after the chip out to deal with the glitch?

    It sounds like ass.​

    Well there you go. Most of what goes on in audio can't be explained by a steady state 1000 cycle per second sine wave. Sometimes you have to learn from the wizards. And I've said this before, a lot audio design is like cooking. There's really nothing new in audio. Nothing in audio is state of the art. These aren't 5nm GPU parts or CPUs with dozens and dozens of crazy long prediction branches. It comes down to people with experience putting the ingredients together. Those without experience who copy recipes from datasheets and confirm their cooking is good via a sine wave - they have much to learn and will probably never get better. BTW, ASRC chips also sound like ass, at least those from ten years ago.
     
    • Like Like x 7
    • Epic Epic x 5
    • List
    Last edited: Sep 10, 2021
  16. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
    I'm really surprised that more manufacturers don't use the sign magnitude approach to R-2R DACs, most notably seen in the Soekris DACs. One ladder for the + signal, one ladder for the - signal, minimizes zero-crossing glitch.

    Or maybe other manufacturers do use sign magnitude, and they just don't advertise that they do....?
     
  17. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    PCM1704 was sine magnitude. And again, it's "random" and not necessarily a zero-crossing. (Keep in mind that Schiit fixed this in firmware, probably via offset). Find me worst case instances of 1111111111111111 being flipped to 000000000000 in a 16-bit wave file. Of course there are another bad numbers too, 0000111100001111 to 111100001111000 just as bad too. Same with 1100110011001100 to 0011001100110011. Heck, even 0000001111111111 to 0000000000000000 is kinda bad based on the 12 bits being flipped.

    I'll leave you guys to convert binary to decimal.
     
  18. Merrick

    Merrick A lidless ear

    Pyrate
    Joined:
    Jan 6, 2016
    Likes Received:
    12,569
    Trophy Points:
    113
    Location:
    Portland, OR
    [​IMG]
     
  19. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    You called it. It's basically like the Soekris DAC2541 on steroids. Tonal balance very similar, except with a 1 degree tilted more downward slope. Main differences are as follows:
    1. Megacomboburritofilter soundstage - significant
    2. Better extension in the lows - moderate
    3. Better macrodynamics overall - moderate
    4. Bass slam and heft - significant
    5. More resolving - minor
    6. Harmonically richer - minor (different, not better)
    I loved the Soekris personally; however a few I have spoken to personally felt the Soekris was not for them on the account of it sounding too lean, I am attributing to #2, #4, and #6 above. The question is whether the LIM would be better suited. My answer is generally yes. So long as you understand the LIM isn't as thick as Bifrost or organic as Yggdrasil A2 (now OG) - and that if you like the Bifrost 2 or Yggdrasil A2/OG already and have no complaints, then stick with it. (Remember, I am MR. DO NOT SIDEGRADE).
     
    • Like Like x 21
    • Epic Epic x 2
    • List
    Last edited: Sep 10, 2021
  20. Inoculator

    Inoculator Friend

    Pyrate
    Joined:
    Dec 5, 2018
    Likes Received:
    3,522
    Trophy Points:
    93
    Location:
    Tacoma, WA
    Now I am even more intrigued. I enjoyed the 2541, but found the bass too lean and did not like how it staged with my 2 channel system (headstage was not an issue with headphones though)
     

Share This Page