Mercury V3 Streamer/DAC Impressions

Big thanks to @Michael Kelly and SBAF for the wonderful loaner.

I'm not going to go too deep into the weeds but my hot take is the Mercury V3 is an evolutionary improvement to the Pi2AES hat in almost every way.

Chain: Roon, wired ethernet to optical over ethernet conversion, Pi2AES/Mercury V3, Yggdrasil MIB. The V3 is clearer, cleaner and more dynamic than the Pi2AES, but not in the typical "veils were lifted" kind of way that usually results in a "details at the sake of realism/musicality" type of listener fatigue. Yes, it seems like there's more detail, but my guess is that the overall lower noisefloor has more to do with everything than the typical fake detail from a slight lift in the treble and some increased 3rd order harmonics. There's only a slight hint of etch in the upper midrange of the V3 that, for me, could easily be mitigated by some careful tube rolling in the driver section of my integrated. I do perceive the V3 to be an across-the-board upgrade to the Pi2AES hat.

Compared to my Audiolab 6000CDT the V3 is less liquid (space between notes kind of thing) but beats it in outright dynamics. Tonally I think I still prefer the 6000CDT. Compared to the Shanling ET-3 its a bit of a wash; I preferred the authority in the lower midbass/bass that the ET-3 has, but from the midrange on up they trade punches equally. Yeah, they're a little different, but we're splitting some thin hairs here.

Anyhow, I'll be placing an order for the V3 soon as it's definitely wonderful value proposition!

 

I have been listening to the Mercury v3 all weekend. I have been comparing the Mercury AES output and a Pi4 USB into my Yggdrasil A2/Unison. Both are running RoPieee.

Setup was connect it and power it up. Doesn’t get any easier.

A big shoutout to Michael - I had a question about the front panel lights and he answered it in a couple of hours.

Chain is ROON/NAS/NUC through wired Ethernet to optical with the optical modem connected to a linear supply on the streamer end. USB power is isolated and the USB Pi is run by a linear supply. A little audiophile nervosa there but I got it sounding the way I like it.

Rest of the chain is Yggdrasil>McIntosh C 1100>McIntosh MC 2301’s > Sonus Faber Amati Futura’s.

Headphone setup is the same except C1100 > McIntosh MHA 200 > ZMF Atrium or Veritie.

My system has never sounded better.

The Mercury has a deeper, wider soundstage with more defined instrument placement than USB. Timbral accuracy is better. Mercury is more forward but not in a hyped up way. It’s like being in the 5th row instead of the 15 row at an orchestra performance.

I think it’s hitting a synergy sweet spot for my system. The Amati Futura’s are dark and tilted slightly down in the treble.

All in all a fantastic value.

Music was:
Saint Saëns Sym #3 Ormandy/Murray/Telarc
Blue Joni Mitchell 192/24
Beck Morning Phase 96/24
Sun on Sand Joshua Redman/Brooklyn Rider
Bartok Concerto for orchestra Reiner SACD
Allman Brothers Eat a Peach SACD
Ruth Gipps Symphony #2 BBC Orchestra

P. S. I experimented with a linear supply and the supplied switcher. I could discern zero difference.

Tom

 

What interface did you use?

 

Digital output? AES from both the V3 and Pi2AES, and ET-3, but coax for the Audiolab.

 

I’ve owned an older Pi2AES for a few years and I was curious if the newer V3 model would make any noticeable improvements in my system.

Chain - Roon -> Ethernet to Fiber to Ethernet to Pi2AES/Mercury V3 -> Yggdrasil A2
Note, I’m running a Jameco power supply to the Pi2AES instead of the stock Meanwell and using AES from both units to the Yggdrasil.

Right away I noticed the Mercury V3 is cleaner and clearer with more focus. Instrument separation and placement is improved with added depth in the soundstage. Notable, 90% of my listening time was spent on speakers. One thing that stood out was the V3 sounded louder. I ran a few dB tests with my phone using both pink noise and a 1k signal and it was the same between units. I can only guess the more focused image presented as being louder.

It’s always fun to try different equipment and hear things you’ve never noticed before. I’m not sure if the V3 has better detail retrieval or less noise that lets you hear the music more clearly. Favorite moment was listening to Fogh Depot’s “Quicksilver Spoon” where there is a drum beat bouncing back and forth from speaker to speaker. It was only after listening with the Mercury V3 that I could hear it more clearly and you could track it as it slowly tracked from one speaker to the other. Hard to describe but I could hear more notes in the center as it transitioned through to the other speaker. Swapping back to the Pi2AES, you could still hear it, though not as well, and it didn’t have the same effect.

The Mercury V3 is a great piece of kit. Thanks to @Michael Kelly and SBAF for making this happen.

 

Brief impressions as I just had the loaner. Unfortunately, it came at an absolutely awful time for me for work, school, personal life, and then the flu all converged such that I only had 1 good day with it. I was only able to use AES/EBU for a quick round of listening over a couple hours. Apologies that I did not get to do more with it. I am still running a fever as I type this so I hope it will be coherent!

TL DR

The Mercury V3 is a practically built device with good features and decent build. Some of its features do not align perfectly for me but that should not break anyone else's deal. The price seems reasonable. AES/EBU from the Mercury V3 makes the Spring 3 KTE sound as good as it can on that input. I'd prefer no DAC and only digital outs but the price is the same as the V2 without DAC.

Spoiler: Incoherent Complaining

I used the Mercury V3 AES/EBU to a Holo Spring 3 KTE, Audia Flight FLS-1 Preamp, Audia Flight FLS-4 power amp, and Harbeth SHL5 Plus XD speakers. The sound was quite good and I think it's as good as the Spring 3 is able to achieve with AES/EBU. My reference with the Spring 3 KTE remains I2S as I have now spent literal years comparing USB and I2S and I still find I2S to be superior on the Spring 3 KTE. Unfortunately I did not get a chance to use I2S from the Mercury V3 due to having the flu.

To compare the AES/EBU of the Mercury V3 and the I2S of Holo Red to the Spring 3 KTE, I still say I2S from the Holo Red is superior. There is a much better sense of rhythm and timing and "being there" with I2S to the Spring 3 KTE. I occasionally play guitar along with recordings and I2S always gives me much more of a real sense of "sitting in" with the band than any other input on the Spring 3 KTE. The AES/EBU input is good, but just not as good as the I2S input. I would need a couple weeks, or maybe months to really compare I2S with the Mercury V3 and Holo Red to decide if they compare or not. Unfortunately I did not have time to try I2S with the Mercury V3.

The Mercury V3 certainly is a nice device. I would prefer that it does not have a DAC, but that is a minor gripe. I really dislike dealing with Ropiee. Any config change requires reboot, configuration, etc.. I know others dislike Holo Red OS and I agree it's not perfect, but I want a device that is just going to convert Ethernet packets into digital output without complexity of a DAC. The only value in the Mercury V3 having a DAC for me is if I could use it with my TV with an optical input or ARC input. Otherwise it's just additional complexity without any practical use case.

 

Why would you buy Pi2AES based design and go USB to a DAC if you have three other options of connecting? Pi2AES was created in the firrst place I guess to go..Pi to AES. Just curious.
Maybe it is a question to @Michael Kelly : is USB out from M3 the same as RPi USB out or there is something special about it?

 

It's been stated before that USB has no special treatment on the Mercury V3.

 

The Raspberry PI inside Mercury and its USB ports are powered from a 12V to 5V converter internally. This is a switching circuit so is not optimized for super low noise. The AES transmitter (which creates all the I2S signals) and the clocks are powered from a 12V to 5V low noise linear then reduced to 3.3V using ultra low noise linear regulators the final voltage.

 

One small point, Mercury has the mini-noise nuke which has demonstrable improvement for AES and I2S. Perhaps that helps USB. Not sure since I can't measure. Just thought to mention it for full clarity.

 

I did find that a noise nuke improved usb output from my Pi2AES. I have not done an A/B with the Mercury v3 - partly because it has a noise nuke built in. Despite the built in noise nuke, I use a LPS for the Mercury mainly because it has the same input voltage as my Dayzee dac which is what I built the LPS for. Suffice to say, usb output from the Mercury into both of my DACs (Dayzee and MiB) is very good IMO.

 

I was part of the Mercury V3 loaner tour. Here's my write-up.

TL;DR

Over the past 9 years, I’ve tried a total of 10 different streaming solutions/combinations in an attempt to find a reasonably priced digital transport for my Yggdrasil A1 that is as good as my Rotel RCD-1072 CD player. The Mercury V3 is the best of the bunch—and a tie with the Rotel.

After I sent the loaner model on to the next reviewer, I promptly ordered one for myself.

Barring the release of a V4 that is a major step up, the V3 is my endgame streamer.

***

SLIGHTLY MORE DESCRIPTIVE WRAP-UP

Prior to testing out the loaner Mercury V3, my streamer was the Holo Audio Red running RedOS.

I was happy with the Red, and, if not for this loaner tour, I wouldn’t have thought to try out the Mercury V3, particularly as I didn’t really care for the V1. (I found the highs on the V1 extremely off-putting.)

So my main testing of the V3 was through a head-to-head matchup against the Red. First I ran both streamers on RoPieee just to verify it was an apples-to-apples comparison, but then I repeated some of the comparisons running the Red with RedOS, which I think performed better.

The results: In terms of technicalities, they were very close (particularly running the Red with RedOS as opposed to RoPieee), and I would be happy with either one of them. In terms of sound signature, the Red is very vibrant, very detailed and clear, but slightly edgier in the highs. The Mercury V3 is more relaxed but in comparison sounds the tiniest bit more grayed out and recessed.

If your downstream chain is very relaxed and warm, or if you need top-notch USB implementation, the Red might be a superior choice. And if you stream mostly jazz or laid-back pop, the Red might give you a bit more color and texture—especially with RedOS.

But I do the vast majority of my jazz listening on vinyl. When I’m streaming, it’s generally rock, metal, or grunge via a Yggdrasil A1 using my Grado RS1i. Not what you’d call a warm and cuddly combination. With this chain, the Red is just a little too much. Plus, I never use the USB input on my Yggdrasil (cause it’s dead, and it’s a Gen 3 anyway).

So, I’m going with the Mercury V3 as my streamer. As noted above, ever since I got my Yggdrasil in 2016, I’ve been trying to find a streamer that could match my Rotel CD player both in terms of technicalities and listening enjoyment. The Red could match the Rotel on technicalities, but it just wasn’t as pleasant to listen to. Given the option, I found I would always wind up reaching for a CD.

The Mercury V3, on the other hand, not only matched the Rotel on the technicalities but also was as pleasurable to listen to. I never felt the need to switch to a CD. The presentation between the transports was different—the Rotel was a bit sturdier sounding, whereas the Mercury V3 was airier and more delicate—but I never felt like I was missing anything. Maybe—mayyybe—I liked the Rotel’s presentation a tiny bit more. But the Mercury V’s ability to handle hi-res files tips the scales back to even.

So I think I’m done with my streamer quest. The Mercury V3 is it.

Now, as for the Mercury V3’s DAC . . . uh, I have to admit that I totally forgot it existed until just after I shipped the loaner on to the next reviewer. Oops. But, as noted above, I just purchased my own Mercury V3, so I’ll evaluate the DAC when it arrives. Obviously, it won’t match up to the Yggdrasil, so I’ll do a head-to-head comparison with the Schiit Modius I purchased a few years ago for a secondary system. That seems like a fairer comparison. Stay tuned.

***

FULL, VERY LOOOONG WRITE-UP

MERCURY V3 vs. HOLO RED: SONG-BY-SONG LISTENING COMPARISON

Both streamers went to Yggdrasil A1. All streamed via Roon. All listening was via headphones, either Sennheiser HD800 with SDR mod or Grado RS1i stock, both amped by a Zana Deux Super running RCA red base tube.

Unless noted below, both the Red and the V3 were both running RoPieee

***Let’s start with my Sennheiser HD800 (SDR mod) for jazz, pop, and low-aggression rock.***

Kenny Dorham, Quiet Kenny, XRCD 1992 mastering, “Lotus Blossom” 16 bit 44.1 from local library via Roon

There’s maybe just a little bit more air coming from the Red during the long ride symbol introduction and maybe just a little more bite to the trumpet tone. Maybe. This is really, really close.

Lee Morgan, The Cooker, Blue Note 2020(?), “New-Ma,” 192kHz, 24bit, from Qobuz via Roon

This album is now my go-to treble torture test because it was recorded so very hot and miked waaay too close to Morgan’s trumpet. His tone can be piercing and distorted at points. So, how did they do? Pretty well, actually. (And certainly better than the first time I listened to this track on vinyl after my new cartridge had been set up for me with far too little VTF. Ouch!) Again, maybe a little less abrasive on the Mercury V3, but, dang, it’s close.

Natalie Merchant, Tigerlily, “San Andreas Fault” original 1995 CD mastering 44.1kHz 16bit from local library via Roon

The Red gave me a tiny bit more of a leading edge to the vocals, but it also seemed that the Mercury was a bit smoother. But, again, just not a big difference.

Natalie Merchant, Tigerlily, “San Andreas Fault,” 2007 96kHz 24bit mastering from Qobuz via Roon

On both the Red and the Mercury V3, you can definitely hear the difference in the mastering and resolution, both the good (greater spaciousness and reverb/decay to vocals) and the bad (the bass drums was pushed farther back in the mix and lost much of its oomf). The Red did a slightly better job with the sustained decay on the vocal but was just a tad harsher. Again, though, very small differences. I honestly don’t know if I could tell them apart in a blind A/B test.

David Bowie, Rise and Fall of Ziggy Stardust, “Suffragette City,” 2015 remaster, 192kHz, 24bit, from local library via Roon

This is the track where I could tell them apart the most clearly on the Senns—the snare via the Red was just a tiny bit sharper and less pleasant than the Mercury.

***Changing up to my Grados RS1i now for some metal and grunge.***

Soundgarden, Superunknown, “Drown Me,” original 1994 CD mastering by Bob Ludwig (muuuuch better than the recent reissue), 44.1kHz 16bit from local library via Roon

Wow, maybe I should’ve started here. Much easier to tell the difference with this music through these headphones. It’s the same thing that I was hearing on the Sennheisers playing “Suffragette City,” but much more pronounced. The Red is giving me more information–a bit more color, maybe a tiny bit more slam–but the sound is overall less pleasing than with the Mercury V3. The Mercury sounds just the tiniest bit more grayed out in comparison, but I think that helps here, particularly with extended listening.

Iron Maiden, Piece of Mind, “Where Eagles Dare,” 2015 remaster 96kHz 24bit from local library via Roon

Here’s an interesting test: This 2015 remaster is the only version I own (or that is available via streaming services), but it’s pretty dreadful. It’s kind of muddied, and the vocals are buried and not particularly distinct. So this should play to the Red’s strengths, right? Well, yes, the Red does a better job of separating out the instrument and bringing the vocals forward. But I’d still rather listen to the Mercury V3 for all the reasons I mentioned above. It’s easily the more pleasing of the two.

Judas Priest, The Chosen Few (2012 compilation) “Exciter,” 44.1kHz 16bit from local library via Roon

This is another case where the only version I own is poorly mastered. In this case, it’s too shrill. You can probably see where this is going… Yeah, the Mercury V3 is just a more enjoyable listening experience. It's notably more relaxed.

***Changing the Red from RoPieee to RedOS***

Just to make sure I wasn’t holding back the Red from playing at its best, I then reinserted the original RedOS and listened to “Exciter” again. I know some of the folks on the “Deathmatch” thread thought that RedOS is the worst-sounding operating system for the device, but I gotta say, I strongly disagree. There’s a real sense of openness and clarity that wasn’t there with the RoPieee. Plus, some of the edge I heard with the RoPieee is reigned in with the RedOS. So it does a better job with what the Red does well and less of a bad job with what it does poorly. It makes the trade-offs a better bargain. The bass has additional slam, too. Ultimately, I still slightly preferred the Mercury V3 on “Exciter,” but it was much closer running RedOS than with RoPieee.

So, next I tried the RedOS with Iron Maiden’s “Where Eagles Dare” and found that . . . nope, the Mercury V3 is still much more pleasant than the Red, even with RedOS.

And then, working further backwards, I listened to Soundgarden’s “Drown Me” again with the RedOS. My experience here was much as it was with the Judas Priest track—the Red is less offensive (and a bit more slamming) with RedOS than with RoPieee, but the Mercury V3 is still more enjoyable (but the delta wasn’t as large as with the Iron Maiden).

Winner: Mercury V3
***

MERCURY V3 vs. THE REST (MERCURY V1, UNISON, LYNX CARD, ROTEL RCD-1072)

All transports went to a Yggdrasil A1 except for the Schiit Unison, which went to a Yggdrasil A2. All streamed via Roon except the CD player.

I spent a little time comparing the Mercury v3 against the Mercury v1, but I didn’t need more than a few songs to realize that the V3 was simply head and shoulders above its granddaddy. The V3 was clearer, more open, and much, much more relaxed and pleasing in the treble. Even in comparison with Red running RoPieee, the V1’s upper register was much less relaxed, but at the same time it lacked the Red’s vividness, detail, and slam. Winner: Mercury V3.

In comparison to Schiit Unison USB (working from memory and previous head-to-head comparisons with other streamers as my A2 with Unison is now packed away for sale), I’d say that the Mercury V3 was technically superior: better detail, clarity, spaciousness. Unison offered more color, but it was maybe too much color and not entirely natural sounding. Winner: Mercury V3.

In comparison to my old Lynx card enclosure (again, working from memory and previous head-to-head comparisons with other streamers, as I haven’t had it set up since my Mac mini died about a year ago), I’d say that the Lynx was even more relaxed overall than the Mercury, but it was significantly less vivid and detailed. Winner: Mercury V3.

So the Mercury V3 has thus far bested all rivals, but how would it stand up against my go-to digital source: my 21-year-old Rotel RCD-1072 CD player? (Here I’m using it purely as a coax transport to the Yggdrasil A1.) The Rotel has always been my measuring stick–and the reason I’ve gone through ten different streaming devices/combinations since I bought my Yggdrasil back in 2016. Because of the Rotel, knew how good the Yggdrasil could sound. Why couldn’t I find a streaming solution that sounded as good as that? I kept throwing one combination after another at it, but none could equal it in terms of detail, color, smoothness, and especially solidity and coherence.

When I bought the Holo Red, I found that it actually was able to match the Rotel on the technicalities–which kind of shocked me after so many failures. But I just found the Rotel to be a smoother, more relaxed, more pleasing listen. When I had the choice, I found myself listening to the CD.

So would the Mercury be the one to deliver the full package? Holy cats, yes it was. It’s either on par with the Rotel or just a hair’s breadth behind it. I tested the Mercury V3 against the Rotel for all the tracks above where I had the same mastering on CD and in my local streaming library, and every time, it was very, very close. This is not to say that I couldn’t tell them apart. They had subtly different presentations: The Mercury V3 is slightly airier and more delicate, whereas the Rotel sounds more solid. But both are very enjoyable, and on the technicalities, the Mercury V3 keeps up with the Rotel at every step. Generally speaking, I very, very slightly preferred the Rotel’s presentation, but when you factor in the ability to play high-res files, it evens out again. The important thing is that I don’t feel like I’m missing anything when I’m listening to the Mercury V3. I’m not pining for the Rotel.

To give you an idea of what it took to get to this happy conclusion, below is a ranking of all the streaming/transport solutions I’ve tried over the last 9 years in my attempt to find a digital transport as good as my Rotel.

***

RANKING OF MY DIGITAL TRANSPORT SOLUTIONS

All transports went to a Yggdrasil A1 except for the Schiit Unison inputs noted below, which went to a Yggdrasil A2. All streamed via Roon except the CD player.

All listening was via headphones, either Sennheiser HD800 with SDR mod or Grado RS1i stock, both amped by a Zana Deux Super running RCA red base tube.

1a. Rotel RCD-1072 CD player via coax
1b. Mercury V3 running RoPieee via AES
2. Holo Red running RedOS via AES
3. Holo Red running RoPieee via AES
4a. microRendu to Schiit Unison USB
4b. Mac mini direct to Schiit Unison USB
5. Lynx E22 Thunderbolt Enclosure (from Mac mini) via AES
6. MicroRendu to Schiit Eiter (i.e., Schiit Gen 5 USB enclosure) via coax
7. Mercury V1 running RoPieee via AES
8. microRendu to Schiit Gen 3 USB
9. Mac mini direct to Schiit Gen 3 USB

I’ll probably start fiddling around with GentooPlayer and maybe Diretta now on the Mercury V3, but in terms of hardware, I feel like I’m set.

Finally.

 

Just read this whole thread and wondering why more people haven’t commented on the sound quality of the i2s outputs?

 

I haven't used the I2S out of the Mercury because I don't own a dac with I2S input and I think a sizable percentage of people on SBAF are in the same "dac boat".

But IIRC, when I did own the Matrix X-Sabre, I2S out of the Pi2AES wasn't any better than the AES output. I don't know if the Mercury makes some sort of leap there.

 

using I2S bypasses the DAC's internal clock extraction circuitry. AES, SPDIF or optical all require the DAC to recover the critical clocks from the input stream. I2S inputs those clocks directly without the need for extraction. Different DAC's do a better job than others. That is why on one DAC you don't really hear a big difference, but on another you might. But let's also not forget the rest of the audio chain, including your ears! From a strictly measurement point of view I2S will always be superior, even if that difference is tiny or essentially inaudible.

Ain't this stuff fun?!

 

Fun to a point I enjoy the researching and learning and deciding what would fit the best. Since using it for streaming 24/196 tidal. Sounds like the AES would be best option.

If later I start getting into higher rez pcm or dsd will have to use the i2s output since the other outlets don't work above 24/196.

 

Well I waited a couple of days to make this post. On Wednesday night I got an update on the V3. After the update I started my usual listening session and noticed immediately that something sounded different to me. So I went to my go to music list I use specifically for critical listening.
Whatever that update did is insane! There's this new dimensional plane that wasn't previously there. An expansion in depth and the biggest change was in the height. Imaging seems more coherent and better layered.
I spent the better part of 5 hours going through music to confirm this revelation.
Has anyone else experienced this with their V3 after an update? Please tell me I'm not trippin! Lol!

 

Which OS are you using?

 

Volumio

 

what dac and interface are you using?