Schiit Gungnir Multibit impressions

Further experiments down the jitter rabbit hole...

Removed the entire cover of the CD player and more gains present themselves. Anyone ever try this?

EDIT: bleh, terrible form factor. Changed it back.

 

I guess I will be taking my CD player apart today....

 

I just installed the Gungnir in my system to replace the Topping D50 after @purr1n's insistence that I get rid of it. I bought the D50 based on the ASR review, that was my first exposure to Amir's website, and I've generally been one to believe there isn't much to gain upgrading from a DAC that measures well like the D50 to the higher end stuff, in part thanks to the circle-jerk on Reddit. After some snooping around on here I lost trust in Amir's reviews, and I kept seeing SBAF members saying they were converted to DAC believers by Schiit Gungnir/Yggdrasil. So i bit the bullet and literally the second I clicked play I could hear a difference.

Set up is Foobar ASIO (-20 db volume control) > Gungnir A2G5 > Vidar> He 6 (4 screw)

It's not warmed up yet, so apparently its going to get better, but as it is I'm already happy with the upgrade. Everything sounds much more natural, bass is more defined, and just like everyone else has said it throws details in your face like nothing I've heard before (I haven't heard very good source components, D50, Bifrost, NFB 11). Big drums have more body to them and even without a preamp there are more dynamics. The D50 boasts a non-destructive digital volume control, I now realize this is because those dynamics were out of the D50s abilities to begin with. I've been listening for maybe half an hour so I'm still getting to know the sound, but it was such a clear bottleneck I wanted to express my appreciation for the advice.
Thanks, I'm a Schiit shill now, where's my friend tag?

 

Try listening to Kodo - O-Daiko. Set volume for comfortable level with the flute during track opening. Don't have beverage in mouth or attempt to play bladderbust while listening. You will most likely lose the challenge when the big drum hits.

Topping D30 in my lab renders sound as a cartoon mush version of the recording.
Gungnir MB A2 v2 firmware will provide a startlingly realistic rendition of the big drum. Transient leading edges hit hard as heard when standing in front of similar drums setting mics. Depth of drum fundamentals produces a phantom bass in head and neck on my HD800-Jmod as if subwoofer were present.

 

Wait, HD800 Jmod? Do you mean 650, or has Jupiter branched out?
Either way, I'm going to Google that track...

 

No, not Jupiter. Local persona @johnjen modification. After hearing several of his works I had him modify one of my HD800 and kept one HD800 stock for A/B comparison. This has become my goto can of choice. Fits my personal preference profile better than any other electro-acoustic transducer outside of Klein-Hummel O300.

 

Currently downloading

Once I can snag a used Saga and I get a chance to make a good comparison between the Vidar and FirstWatt F5 @bazelio has been so gracious to hold for me I won't be changing anything in my set up for at least a couple years (probably some point after graduating). As it is all my money is tied up in assets that I need to get rid of, or being held by Paypal .

 

After some more time with the Gungnir I think I've gotten a good idea for all the areas that it improves on over the Topping D50. Soundstage is bigger in every dimension, greater instrument separation, precise imaging, notes throughout the spectrum have a fuller more natural tone, a more dynamic presentation, and details galore.

Gungnir pushes details forward in a way that has given me a lot of aha moments. Things like the guitarist moving his hand around the guitar while not playing, or realizing that some low notes in a song were actually a bunch of instruments playing at the same time. I went from hearing the averaged sum of the instruments to being able to tell that there's a piano, bass, and a tuba all playing the same low note in Bloom by Between the Buried and Me. I'm also noticing that a lot of songs have piano and keyboard playing at the same time when I had only noticed one or none of them before. Song for the Man by Beastie Boys does an amazing job of showcasing the Gungnir's ability to keep everything in the mix even when the song gets busy. I don't know what the instrument is called but there is some percussion in the beginning that involves someone moving something up and down the instrument to make sound, and I can clearly tell where on the instrument the sound is coming from, and If he's going up or down.

impact and decay on snares and cymbals have also become dramatically more pronounced, revealing a bunch of ghost hits on the snare that I hadn't realized were there before. The shimmering of cymbals are much more pronounced instead of just being a monotone pitch (this adds a lot of musicality to me, and has me developing a newfound appreciation for cymbals in general), and decay lasts longer before fading away. There's substantially more dynamics in general.

I'm using SE outputs for now but I ordered XLR to RCA cables from Redco that were linked by someone around here. Probably not something I will be able to notice making an impact but worthwhile to get rid of any nagging Audio Nervosa.

 

Got my Gungnir Multibit 13 days ago and have been playing it almost non-stop (never turned off).
Over the last month I've been listening to the Modi Multibit, Bifrost Multibit, and now the Gungnir Multibit, all feeding the Jotunheim.

Jot/Modi Multibit: The only situation where I would use the Modi Multibit is maybe in a work situation when I'm only half listening. It sounds closed in and compressed.

Jot/Bifrost Multibit (w/ Singxer SU-1): Dynamic, excellent bass, fast. No complaints.

Jot/Gungnir Multibit: Wow, whole 'nother league above the Bifrost Multibit. Hard hitting, finely textured bass, big/deep soundstage, impressive separation of instruments and vocals.

So Jot/Gungnir Multibit is now my daily driver. I've tried a number of headphones with this setup and came to a few conclusions:
HD800 (unmodded) is too bright/dry.
LCD2C sounds a bit dull and unengaging. I'm not sure why but it sounded better from Jot/Bifrost Multibit
Sony MDR Z1R not surprisingly given its slightly dark nature, is an excellent match with Jot/Gungnir Multibit: musical, dynamic bass, just a fun sound.
Beyer T1 (2nd Gen) is surprisingly good with sweet female vocals, 3Desque soundstage, and good bass (doesn't slam or extend as much as the Z1R).

 

This post—at the request of @Hands, @rlow, @famish99 and others—is going to upset the apple cart a bit. My intent is not to shit on Schiit for shit's sake, but to point out a quality of their DACs that has bothered me and at least a few other people. Hopefully this will help other readers make better-informed decisions instead of jumping blindly on the Schiit Wagon.

In a recent demo of a well warmed up Gungnir Multibit (admittedly fed from Gen5 USB), I mentioned it sounded etched / unnaturally bright. Chain was MacMini->Gen5 USB->Gungnir Multibit->Pass B1 custom->Pass XA30.8->B&W 801 Series III.

My spouse, who is very good at this stuff and is the one who turned me into an audiophile, also listened to Gungnir Multibit and said the same thing. I first thought it was tonality, but I covered the tweeters and that very obviously did not fix the problem—it wasn't a tonality issue. Similarly, @Thenewerguy009 and I have been chatting, and he also disliked Gungnir Multibit, saying it sounded "digital" to him. Fair—that's another way to describe what I'm hearing. So that's three people who feel there's something unnaturally sharp about Gungnir Multibit. Now, I realize this stands in contravention to what most people say about Gungnir Multibit. They say it sounds like a cloud, it's dark, it's smooth, etc. But I'm not echoing the popular trope. I'm telling you what *I* hear on *my* system—and I will later explain why I think others like it. FWIW, I've heard similar unpleasantness in the Yggdrasil Multibit, though it was long ago.

After a week of demoing Gungnir Multibit, it dawned on me that I had heard this kind of crispy sound before... when fixing a push-pull class-A amplifier that had mismatched stages. It's the sound of odd-order distortion! So, that's was my working hypothesis, and I was going to take my own measurements with my QA401 and my HP 3561A—but then the Gungnir Multibit sold to a local buyer who came and picked it up right away. But just as well, since nothing better than to use measurements done by others who don't have a dog in the fight to help make my point.

Let's start with the below, from @atomicbob. It shows a Gungnir Multibit sine wave at 1kHz, but unfortunately at -90dBFS...



The little blips at the zero crossings in the green trace should create 3rd order distortion. The red & blue FFT along the bottom isn't labeled. It could be the power-supply noise. But if it's the audio signal, you can see the fundamental as the big spike, and the next spike to the right is presumably the 2nd harmonic, and the next (and bigger) spike after that is the 3rd harmonic. If indeed the third, fifth, etc. odd-ordered harmonics are all higher than the even, then (1) that's exactly what we expect to see based on the green waveform, and (2) this will cause the DAC to sound crispy. The problem with this measurement is that -90dBFS fundamental is .... not exactly relevant—but more data below. FWIW, he repeated these measurements on BNC and optical, and the DAC produced the same pattern of distortion. It's not the input.

Here's another FFT measurement by AtomicBomb. Usefully, this one is at a higher level. Here we can confidently see that the odd-order harmonics dominate. In this case, the individual harmonics are ~65dB down off the fundamental—which I would consider just at the edge of hearing. FWIW, Nelson Pass has said in the past that distortion is audible at levels below -66 dB, which is why his company specs their amplifiers at 0.05% THD.



Ideally I would have taken measurements at higher output levels, but too late for that. Yet, further digging reveals this odd-order distortion is seen in other Schiit Multibit DACs. Here is the sine wave on Yggdrasil measured with an Audio Precision SYS2722 analyzer by John Atkinson. Lo and behold, blips at the crossovers (again, most easily visible at low signal). [EDIT: This was done into a 600 ohm load, which @purr1n says the Yggdrasil couldn't drive. The 3rd harmonic distortion at -65dB harmonic in Atomic Bob's measurement of Gungnir Multibit, above, stand. However, as discussed in responses to this post, this was a crossover glitch that is apparently fixed in updated Gungnir Multibit.]



And here is the corresponding FFT with fundamental at 50Hz—and finally a measurement at full scale.



Here we clearly see (1) odd order distortion dominates, and (2) 3rd order spike is only 50dB below the fundamental -- that is totally within the audible range.

The same distortion also appears in Bifrost Multibit, albeit the balance of distortion is a little better. The Bifrost delta-sigma actually measures better on this particular point --- but probably has the stage of a melted ice cube.

All that is just to say I heard what sounded like odd-order distortion and when I later looked at measurements, I saw plenty of evidence of odd-order distortion. You can disagree on the audibility of it, but my hearing it came first.

Now—I really believe these Schiit DACs are remarkable pieces of engineering. They have fantastic stage (thanks probably to their discrete output stages). They are incredibly detailed. Schiit is to be congratulated for achieving what a bunch of companies have failed to achieve at much higher price points. But why did they retain audible levels of distortion? Well, either (1) they did it intentionally because there is a market for third-order distortion, or (2) my theory is they did the best they could given that the chips they used. The Analog Devices 5781 and 5791 were designed for MRI, not audio. They care more about monotonicity and low noise, and in exchange they have nonlinearity and glitch errors. Indeed, the chip designer is on record saying "code-change-dependent glitch energy would make THD unacceptable to audio buffs." There are deglitching techniques and I think Schiit has done pretty well in the implementation—but the distortion is still audible.

So who cares if there's audible but low levels distortion? Most of our gear has lots of distortion, much of it at higher levels. And if you're like me, you know the cost of squeezing out the last bit of distortion is usually a loss of stage.

So, this is where it becomes subjective. Nelson Pass has a really good article on distortion that I commend to every SBAF reader. Basically, it boils down to this: You're going to have some distortion, so just pick your flavor. Odd-order distortion tends to sound better with complex music (e.g., pop, metal, electronic). It gives the perception of detail and a sense of separation. It has long been associated with a certain "hifi" sound. Even order distortion sounds is more euphonic and "liquid." It sounds better with simple music.

The rest of the chain must be considered too. If you have a SET tube amp set-up, you may have a lot of even order distortion down the line that balances out the odd-order distortion of the Schiit DACs. No surprise Schiit sells a bunch of tube stuff now. My amp, the XA30.8, is so low on distortion that I don't get the euphonic band-aid to fix it. It's important to know your chain.

I hope this helps explain why I don't like Schiit DACS.... but you might.

 

Oh boy the zero crossing glitch... Fasten your seatbelts gentlemen.

I suggest you give some indication on the dates, M. Moffat stated this glitch has been fixed like 2/3 years ago. Unless it remained on the gungnir.

 

What are some of the other Schiit DACs you had in the past? I wasn't a fan of the Bifrost MB but I enjoyed the Modi 3.

 

Fourth from bottom paragraph just cuts off without resolution, and less importantly you misspelled Yggdrasil in paragraph 3. Might have missed a few others, but interesting read. Looking forward to a civil discussion on the same— my listening skills are sucky beyond headphones and amps IMO which probably explains why I don't find the Bifrost MB offensive, but might be a fun intellectual exercise.

I was always under the assumption, based on reading and extremely limited experience with sources, that yes odd-order distortion is less desirable for pleasure listening than even-order; I can say that I think the Bifrost isn't a relaxed DAC at all despite the weird warm, very slight v-shape I think it is (again, I'm a DAC noob with not a great deal of experience here) so this might well explain my perception of it. Thanks for the post.

Minor nitpick, the headline reads like a Buzzfeed article. I don't mind, just felt like pointing it out.

 

Ok, so you basically confirmed this existed.

These were all DACs with older A1 output stage DACs. If this is fixed, I'm glad to hear it. But I'm not keen on buying a fourth Schiit DAC and paying @schiit 's 5% fee just to find out.

 

I think it's extremely reasonable when you consider a lot of companies charge a 15% restocking fee

 

I didn't, but the designer did himself. Long after it was fixed. It was a firmware update that was implemented fairly early iirc, which means you might not have had it on the dacs you tried. Hence why I was talking about being specific with dates.

 

I don't know when the DACs were made. My Yggdrasil test was about 3 years ago. The Gungnir Multibit ... I dunno. I bought it used.

Better yet, why doesn't someone with a current Schiit DAC take fresh distortion measurements and we'll see if that actually fixed it.

 

https://www.superbestaudiofriends.o...rmware-v2-update-technical-measurements.6600/

https://www.superbestaudiofriends.o...rasil-v2-upgrade-technical-measurements.5820/