Jump to content
Computer Audiophile
  • ted_b
    Sign in to follow this  

    exaSound e12 DAC Review



    The exaSound e12: The little engine that could...change how we think of sub $2k DACs



    George Klissarov of exaSound had an idea that he could take his award-winning e22 stereo DSD DAC and create a lighter smaller (read: less expensive) version that lost none of the sound quality....none! The savings would be in case size, inputs and output parts, removal of the display and an ever so slightly less refined power filter. He told me about it over a year ago and I jumped at the opportunity to hear it. [PRBREAK][/PRBREAK]Did I mention that was over a year ago? His dream of lightening the load did not result in a DAC he was willing to sell. And if you know George at all you know that he is such a man of principles that he is willing to forego markets and user experiences if it means sound quality impact. exaSound DACs get their magic from a proprietary sauce that includes, among other things, very custom ASIO drivers. These drivers preclude exaSound DACs from playing in the plug-n-pray world of Ethernet renderers and Jplay sound engines, among others. One buys an exaSound DAC to marry it to an ASIO direct stream that is hand-wired for sound...period.

    So, back to the e12 project. He is relentless in his quest for an exaSound seat at the table of "cheap and cheerful" sub $2k DACs, but not willing to make the tradeoffs. Design after design fail to impress Mr Klissarov's stubborn benchmarks. The e12 idea languishes for awhile, until one day he pings me that the right combination of processor, FPGA, linear power filtration to each subsystem and whew...that only took over a year!! Now he and I are back in conversation and I take delivery shortly thereafter.


    The e12 is a small form factor (5" x 1.5" x 7") with only USB input and RCA single ended outputs. If you must have more, stop reading and go look for an e22 review, and go find $1500 more dollars (or my e28 one if multichannel is your thing). The front fascia has three small lights (USB lock/PCM/DSD), rather than a large display. Me, I could care less...in fact I am glad; I have written several times that 1) I put my DACs far enough away from my listening chair that the display information is mostly moot; 2) displays make noise; 3) displays cost money.






    The e12 keeps everything else that made the e22 such a winner; galvanic USB isolation, multi-layer ac filtering, the top-of-the-line reference level ESS Sabre 9018S chip, set to stereo (4 parallel channels per side), and a secret sauce (I'm beginning to think a lot of it's the custom ASIO driver) that somehow tames the Sabre chip allergy I was born with (I've written before that the SABRE chipsets are so detailed and fast that the leading edges can often be razor sharp, and that only a few DAC manufacturers can tame this detail beast...namely folks like Auralic, Mytek and exaSound).


    This small form factor is also due to the fact that, like his other slightly less diminutive e22 and e28 cases, the power supply is external. I love this design (and have written about it in exaSound and Chord reviews before) because it allows one to field upgrade the DAC when funds and timing work out for the buyer, and it means I've spent more money per pound on the actual R&D'd pieces (DAC, analog stage, etc). Me? I spend a few days listening to the stock laptop-style ps then quickly replace with my trusty $200 (bought used) Hynes SR3-12 linear 12V power supply. (A newcomer takes the Hynes place eventually; I will explain later).


    The piece de resistance, IMHO, of the exaSound value proposition is that it does raw direct DSD up to DSD256 via a hand-built tube-based (ok, just kidding..making sure you were reading) custom ASIO driver that is not only Windows capable but also MAC OSX. Furthermore, if you couple it with JRiver, there is a plugin that will synch JRiver's internal volume control with the exaSound volume control (SABRE chip's own digital volume), so you can sit at your listening chair, turn JRemote's volume up and down and you are not playing with a poorly developed OS-based volume option, you are running volume directly off the darn chipset, regardless of PCM or DSD bitstreaming. Pretty kewl.


    The newest exaSound ASIO driver also allows the user to set a maximum volume limit (just in case) and set a timer when the unit will power itself completely off (note: that last feature is not a favorite of mine since I constantly preach that digital things that have digital clock things in them should never be turned off, lest you want to wait anywhere from 6-48 hours for the clocks to thermally reset to perfect equilibrium...luckily the timer has a "never" setting). George was asked to provide that, and certainly a manufacturer needs to comply with CE, etc.


    So....can the reduction of parts and features and costs be defined as cutting corners on sound quality? Or is this one of those addition-by-subtraction things that occur randomly in the universe (and sound like justified bullshit most of the time).


    I began my evaluation by reducing my dual pc setup to one mighty i7 (16GB RAM, battery powered SSD OS, Hynes powered mains) server. The pc uses a JCAT USB card (if 5V needed, which is true for the exaSound, then linear power is provided via Red Wine Acopian ps), either a JCAT or TotalDac D1 USB cable, JRiver 20 daily build, Jremote and floats in a warm bath of ...(still reading??). I then ran sound tests on it via my plethora of other DACs in the house (Directstream, Hugo, etc) in order to get a good understanding on what a Jplay-less JRiver sounded like on this single pc setup. During this time (2 weeks) I ran the e12 on my Mac Mini second system, with outputs connected to a load but no speakers, etc. It was simply run on a test tracks playlist (various test tracks and music tracks) on repeat for 14 days straight. "Hi, my name is Ted and I'm an audiophile."


    The e12 is capable of PCM to 32/384k and DSD to DSD256, so I threw all those sample rates at it. I started with good ole' DSD64. If Mari Kodama's Pentatone Beethoven piano sonata cycle doesn't sound like a piano is in my music room then we have our first blemish. But in fact, her piano never sounded better, with a beautiful blend of attack and decay, bloom and then brake. Wonderful stuff (some don't like her performances on this cycle but I don't know well enough to care :) ). The unusual jazz trio of trumpet, guitar and bass on fone's La Notte (SACD126) is another of my DSD go-to albums of late, and the e12 reproduced it with very nice burnished tones and razor sharp edges, but not too sharp. The depth of stage is very important in this minimalist recording, and the e12 created plenty for my enjoyment. Even when switching to PCM, the e12's combination of detail and airy delicacy made great redbook sound great and even lousy compressed redbook sound bearable (life is too short to have a DAC that can't play your favorite compressed hits of the 70's). OK, stadium Arcadium is unlistenable with all the money in the world, so that is where it ends.


    All of these comments above were with the stock SMPS laptop style power supply included with your purchase. When I replaced it with my trusty Hynes SR3-12 12V linear power supply the noise floor dropped, the weight of the instruments increased, the dynamics improved and the image specificity (finding the edges of instruments in space) got ever so slightly better. However, these improvements were in the neighborhood of 10% or subjectively "better but not can't live without better", whereas with other DACs that allow external PS swaps these improvements were downright substantial and required, even the e28 (see review).






    UpTone Audio JS-2


    I was intrigued by the linear power supply that upstart UpTone Audio (John Swenson and Alex Crespi aka superdad) was building, and they loaned me one to try in this evaluation. John had a huge hand in the design of the I2S interface on the Sonore Signature Rendu, which impressed the heck out of me, so I knew these guys had some great ideas in them. This $925 dual rail power supply (5A total, choke-filtered supply and expensive R-core transformer) allows the user to dial in voltages on both rails (large thread here on CA for more details); dual rail power supplies are a great choice for powering separate equipment on a subsystem (like a pc, for example, one goes to mains, one to SSD). In this unique case, where the e12 has galvanic isolation after the USB, I used only half of it (one rail, set to 12V...equivalent to my Hynes SR3-12 functionality) so as to not risk bringing common noise (ground) to something separate in my signal path. The noise floor dropped slightly lower again (music emerged from a void, startling at times), bass was slightly more rythmic, and all other aspects matched the Hynes point for point (IOW, started at a great place and got a bit better!). And this from a ps that was using only half its capability, was readily available for purchase (Hynes PS's are very hard to come by) and supports the folks here on CA. Incredible first product; nice job guys!





    This multi-platform software player has quite a reputation on these forums. I had tried it a few years ago and wasn't prepared to deal with the rudimentary GUI and plethora of settings, so I let the trial expire. However, in my life as technical editor for NativeDSD I would assist Tom (tailspn) with readying files for production and we had come across numerous needs for new DSD tools to adjust things like bit rate, multichannel gain trims, etc...all within DSD, and all of the best ones were coming from one source...the desk of Jussi Laako, Signalyst (aka Miska). I feel like Jussi thinks in one-bit 100Mhz :), and when you combine that with the growing group of supporters for his ever improving HQplayer I decided it was time to buckle up and take it for a ride (free trial).

    When one foregoes such wonderful user friendliness as JRemote one must have one's head examined, right? I mean, life is too short...that is a phrase that kept creeping into my needs-examining brain. So I put on the above-mentioned Pentatone and fone Records demo tracks just to get a quick summary of what life would be like without JRemote. Well.....let's put it this way, I have not fired up JRemote in over two months! Do I miss her? A lot; she was my first real love. Will I get over it? Already have.

    I am a huge proponent of finding a DAC's sweetspot. It is almost always a single sample rate (or small range) where the music sounds best. For those DACs that process both PCM and DSD it is almost always one format over the other (leading me to ruminate that maybe we are just going to have to live in a world where there are two DACs in our musical life, a great DSD one and a great PCM one). Now, however, with HQplayer (and possibly other great software) I am understanding why so many otherwise-right-thinking audiophiles would dare to resample their music to a certain format or rate. I always thought that philosophy was full of gimmicky DSP justification, a sort of "mine goes to 11", regardless of how "11" even sounds. But in the hands of professional quality algorithms and filter sets, resampling to a sweetspot may, in fact, be a great bridge to the solution of needing to deal with two DACs.


    HQplayer is, at its most basic, a set of pro quality PCM and DSD filters, PCM dither choices and DSD modulators....all put in front of the user with a rudimentary GUI. It includes an even more rudimentary library manager. All that being said, It should not be taken lightly; it is a major musical tool that any DSD DAC should not do without! Once the GUI is improved (Jussi is writing APIs to allow 3rd party development of control points) this utility becomes a major player in DSD (and PCM) playback.


    OK, so how does the e12 sound with HQPlayer running the show? Let's start with redbook. Gillian Welch's 2011 alt-bluegrass masterpiece (and one of the best albums I've heard in years) Harrow and the Harvest is not an audiophile recording, but it is a beautiful interplay of tonality, starkness and emotional twang. With HQplayer set to 192k and poly-sinc-mp filtering the soundstage was lit perfectly for her center-fill voice, and her soft playing was nicely grounded in a more yellowish light; Dave's contrasting often-seemingly off key guitar accompaniment was magical. The album is a genius mix of 21st century alt-folk songs that seem as though I grew up with them, yet each is original and different enough to keep one guessing. I admit I never grow tired of this album, but with the e12 and HQplayer it takes on a more performance-in-my-house immediacy that needs a slightly darkened room, a late night glass of wine and say goodbye to 45 minutes. It's a beautiful ride. And this is redbook!


    Another 16/44 album that must be heard with the e12/HQP combo is the tremendously entertaining live recording Live a Fip from the French avant-garde-ethno-jazz-world trio (usually made up of at least four players!! ) The Hadouk Trio. On some systems this redbook recording (also available as 24 bit but almost why bother) can sound too dynamic. The Trio use an aggressive blend of percussion, woodwinds like the ancient duduk (beautiful other-wordly sound), some modern effects and a unique invention by the bass player (it's always the bass player), a gumbass...a combination of electric bass and an ancient guembri (look it up). The resultant sound is soft world music put on steroids! With the e12/HQP you are transformed to the live event; mics pick up the sound of the venue and load your room (or headphones) with it, even prior to the first note played. The textures and timbre that are hinted at with most software players are released into the wilds of your room with the e12/HQP. This ain't your world-travelling-South American-now-living-in-Holland uncle's world music...it's now plainly heard as coming from somewhere near the third moon of Saturn. Oh, and it's two discs...so put aside an hour or more! Nice ride..again.


    My final redbook example is Keith Richard's solo album Main Offender, likely a drunken weekend recording session where they made the songs up during one party and then laid down the tracks during the second night...and then accidentally released the master tapes as redbook. The impact of Keith's rhythm guitar genius is so evident in this little diddy that it is infectious. The e12/HQP might even be able to identify the amps, but I'm not expert enough. Growl, buzz, impact. Even his pathetic vocals come across as real and full of rock and roll. Main Offender is clearly evidence that 16/44 can be used to create wonderful popular/rock music. Leave the compressor alone please!


    Let's jump ahead to DSD64. Now we set HQplayer for SDM (Jussi hates the term DSD :) ), 5.6Mhz rate, and set your favorite modulator (mine being ASDM7) and away we went. That's not so difficult, is it (I could have used these settings for PCM-DSD conversion but I found I liked separate format settings better)? Throw fone's La Notte on the playlist and be unprepared for the depth of soundstage when the trio of trumpet, guitar and acoustic bass start playing in your room (you starting to get the consistencies here). My Aerial 20T's excel at soundstage depth, and this recording has it to burn. Heck, with the filter settings one can almost "tube roll" one's soundstage depth and width from one's armchair (I use remote desktop app on iPad to control HQplayer running on the i7 server).


    The e12 is a Swiss army knife for sample rates (as are other modern DACs) but I got the distinct impression that George had made sure each selection was voiced properly..this wasn't just a "quad rate" feature list item. DSD256 (left played "direct" by HQPlayer) native recordings (and tape transfers) had an effortless quality that reminded me of vinyl. Flow and continuity were the feelings that came to mind. Be careful, late night DSD256 listening can be lulling; I want DSD256 recordings playing in my Soylent-Green induced final resting setup....etherizing melodies that calm and soothe so effortlessly that I................ :)


    At $1999 msrp (throw in a hundred bucks or so for HQPlayer) the exaSound e12 challenges other sub-$2k DACs to get serious about deserving a place at the table of your main high-level audio system. We now need to insist on USB galvanic isolation, high-end power filtration, raw DSD playback capabilities..all at this new price point. This DAC should be considered by anyone who needs only single ended connections and has a USB Windows or Mac source. Anyone! Imagine the monies saved to use for recreating hundreds of performances in your music listening room.


    Thanks to George, Jussi, Alex, John, Josef, Marcin, Vincent and the movers and shakers of CA's forum to push me into new territories. I'm loving the music.


    NOTE: Although I volunteer my time and effort to NativeDSD.com I am not financially in conflict with their hardware sales program, which includes DACs such as exaSound. I do not get any material or financial reward from my work (except great friendships and inside release information :) ).






    Ted Brady



    CA Profile ex.png







    Product Information:


    • Product - exaSound e12 DAC
    • Price - $1,999
    • Product Page - Link ex.png
    • User Manuals - Link ex.png
    • Measurements - Link ex.png














    Sign in to follow this  

    User Feedback

    Recommended Comments

    Read some reviews on the E22. Pretty much the same sound. Upgrade the clock and sound may even be better due to lack of noise causing display and headphone amp circuitry. Upgrade the power supply and take it to even another level. Another awesome thing about it is, it only consumes 0.5A vs 2A on the E22. That means a higher quality for less money linear power supply can be used with it! I wouldn't even consider any other DAC under $3000. With the above mentioned upgrades it would probably stand shoulder to shoulder to $5000+ DAC's all day long.


    Hey, how am I going to upgrade the clock for the exasound e12 as you mentioned? any thoughts?

    Share this comment

    Link to comment
    Share on other sites

    This is now closed for further comments