Mini-Review: Mojo Audio Mac Mini Media Server

Spitting Bits

Power Mac G5 — Super Audio Server Extraordinaire!

I’ve been using my trusty (dusty) MacBook Pro as a media server since I dunno when. Oh, right — since I retired my PowerMac G5, which means about 18 months or so. Way back in the day, I remember a lot of fuss and bother about the audio superiority of the PowerMac G5 floated by the Computer Audiophile crew. Namely, that it was the best thing since sliced bread.

Some folks, perhaps those that consider themselves more “objectivist” among the propeller-head crowd, may consider the very notion that a server would make an audible contribution at all as some sort of heresy. I mean, bits is bits, right? Yeah. Look — the whys and wherefores may be mysterious but as I’ve argued elsewhere, my lack of understanding doesn’t entail any particular ontological status.

So, before I go any further, let’s take a second and talk about why the computer might matter. Or not. Well, it’s not that the server doesn’t matter — it does — it’s just that some things have happened in the computer audio space that tend to make it all a little less important.

One is Asynchronous USB. Never heard of “async”? Hmm. Well, after you’ve done a quick tour, let me welcome you back by claiming categorically and without any serious justification that just about everyone has agreed that the spooling data into a DAC using the high-quality oscillators (clocks) in the DAC itself, instead of those in your crap-ass computer, is a better solution than relying on the clocks inside a computer. Async USB does just that, and a proper implementation does do wonders for the elimination of jitter, something DAC chips are notoriously sensitive to. Why this makes the server a little less important may not be terribly obvious, but here it is — a good computer design, with quality power supplies that regulate the CPU and other hardware — including the on-board oscillators — means that it will probably sound better than one without that good, quality power delivery.

My first DAC, a Cullen-modded PS Audio DAC III, was not asynchronous, it used adaptive USB (clocking came from the computer), and with that DAC, my G5 ruled the school. The combo was just more compelling than what I was able to get from my work machine, a MacBook Pro laptop, in pretty much every measure — scale, drama, detail, tone, blah blah blah. The difference in audio quality was clear and painfully obvious.

When I picked up my first asynchronous USB converter, a Legato from Analog Research Technology, the performance gap between the G5 and the MacBook narrowed considerably, though I pretended that the G5 was still clearly superior. It was easier that way. This became more difficult, however, when I upgraded to the Berkeley Audio Alpha USB, arguably the best-in-class USB to S/PDIF converter on the market. With the Alpha, distinguishing computer sources immediately became an exercise in frustration, especially when I added a non-stock USB cable, in this case, a fancy one from Acoustic Revive with the dual-leg/separate wire for power plugged into a USB-battery from King Rex. This completely removed the computer’s power supply from the connection with the rest of the system. As long as I had enough RAM on the system, I was good. I usually targeted 8GB — I don’t know if I needed that much, but Mac OSX always seemed to thrive with more, so why not. As for the why, well, I wanted enough to buffer a lengthy playlist (I’m lazy) of hi-res files.

Why the focus on buffering in RAM, instead of, say, spooling off the hard drive? Well, it has to do with “memory play” — all that stuff about power supplies also affects data retrieval off the hard drive. More paths, more CPU interruptions and more potential delays, to say nothing of the disk subsystem itself. With everything in RAM using the Pure Music audio player, I could ignore all that, and to good effect. Again, using the Pure Music “Memory Play” feature just sounded better. But using it also seemed to minimize the differences the two servers made in the system’s sound quality. Of course, I really only noticed the difference with some very large (hi-res) files, but whatever. So, with everything preloaded into the lowest-latency output portion of the computer (RAM) and fed to an external device that regulated all the timing, it turned out — for me — that the MacBook Pro was better than “good enough”. I honestly couldn’t tell the difference anymore, so I was ready to power down the big server.

As I started thinking about a more permanent replacement, my main concern was finding one that didn’t sound like landing aircraft (the G5 fans were … noisy). I settled on a MacBook Pro, in part because I was familiar with Mac OSX, in part because Gordon Rankin told me to, and in part because it didn’t need all the external bits that a Mac Mini needed: a keyboard, video/monitor and mouse — at least for configuration and troubleshooting. A MacBook Pro can be configured to kill the monitor output pretty quickly, in case that makes a difference, and I figured, compact is better than extra shit strewn all over the place.

MacBook Pro 13″ — My Audio Server

I opted for the 13″ model, because it’s small, and got the cheapest one I could find. I swapped out the drive for an SSD, upgraded the RAM to 8GB, connected my old Firewire drive with all my music, downloaded and installed my then-and-still favorite player, Pure Music. Pure Music let me run my DAC in the then-cool “integer mode”, so I was off to the races. And all was sunshine and light. At least for a while.

Sometime shortly thereafter, I started running into the Widealabs Aurender at dealers and then at audio shows. Everyone was very impressed with its fancy oscillators, optimized OS, bright display and hunky case. I started to wonder if maybe, possibly, my off-the-shelf “server” wasn’t the all-that. Maybe I was missing something. Maybe I was missing out.

Doubts. They’re a pernicious bunch. Ah, the life and mind of an audiophile.

That was a year ago, and no, I never did get my hands on the Aurender. Honestly, I’m not sure I’m going to bother. The price on the entry-level S10 is $6,500+, rather daunting for what could be a relatively minor incremental improvement. I mean, I’ve heard that song and dance before and found that good software and a good converter pretty much made the server irrelevant. In light of that, $6,500 seems a little spendy — especially for a box that isn’t specifically designed to support USB as a connection to the DAC (the forthcoming W10, at $15k, is the Aurender that is optimized for USB). But who knows. Maybe it’ll be a “game changer”.

But the Aurender is simply the fanciest server-only solution I’ve run across. There’s a lot of other options out there, from Weisss, from Auraliti, and more — in fact, this segment looks like someone kicked over an anthill stuffed with aspiring entrepreneurs.

I’ve also heard reports of users getting big gains from modding what they already have, instead of ditching for an all-new front end. Like, say, swapping out the Mac Mini power supply for a linear PSU. Again, good clean power feeding the board could only yield improved performance and therefore improved audio quality, right? Right — maybe. But this is interesting. What if it did matter? What if you could take an off-the-shelf Mac and “do something” to it to level it up, not just a little, but to “world-class”?

Well, that’s what I wanted to find out. Enter, Mojo Audio.

Power Play

Mojo Audio Mac Mini Media Server, stacked with Joule II PSU

Mojo Audio Joule II linear PSU

Guts of the Mojo Audio Joule II linear PSU

I met Ben Zwickel of Mojo Audio at CAF this year. Ben makes power cords, power supplies and fully configured media servers leveraging that most perfect of computer brands, Apple’s Mac. Having spent most of my professional life having to cope with Windows machines, I have a soft spot for Macs. Love ’em. Can’t get enough of ’em. I’ll never do Windows again, if I can help it — and no, Linux isn’t all that much better. Anyway, that’s why all the computer servers I mentioned above are Macs. Mac = awesome.

Okay, fine, yes — I’m overstating it for a reason. The reason? Apple’s iTunes. Is iTunes the world’s greatest music player? Probably not. But it is, arguably, the most important thing that has happened to audio since … well, since Napster. iTunes was a game changer. Think it was the iPod that made Apple the most highly valued company in history? Wrong. It was iTunes. iTunes is king — and it’s what even the most computer-illiterate wannabe/future computer audiophile is already familiar with. If you’re a prospective audio server modder, tossing iTunes out is, at best, a great way to miss out on the bulk of your prospective customer base. Yes, there are better clients. Audirvana, for example, makes your system sound better. But Pure Music is better still (my opinion, of course), and part of that happy-happy joy-joy on my part is due, perhaps, to not having to do anything to the terabytes of already-ripped CDs currently in my audio library. Forcing a happy, comfortable consumer into a new paradigm is, historically, a loser. So, why bother? I got iTunes!

Which is one reason I find the Mojo Audio solution interesting. It uses an off-the-shelf Mac Mini. It then takes that Mini apart — ripping out the stock switching PSU and replacing it with an on-board power-filter and an interface to an off-board, super-duper, linear PSU. And that’s pretty much it. Pretty much. The Mojo Mini is minimally configured — Audirvana is pre-loaded, for example, and pops up automatically after a reboot. So, a user simply selects some tracks and Bob’s your Uncle.

Ben sent me a brand new unit that I promptly hooked up (after I went and bought a $150 HDMI/DVI monitor — I already had a spare mouse and keyboard) and ran in for 2 weeks. After that point, I ran through my usual demo tracks. After some pretty exhausting (if not exhaustive) testing, I can jump right to the punchline — the Mojo Audio Mac Mini Server is the best computer source that I’ve heard in my system to date.

I’m annoyed by this. Honestly, I was hoping the differences, if any, would be small enough that I could pretend I didn’t hear them. I don’t want to have to get another server …. Anyway, that’s my problem.

So, before I get into details on the Mojo, let me say this — the level of improvement over my MacBook Pro is audible, even if it isn’t exactly “night and day”. This is becoming something of a common refrain, it appears, but then, marketing and hyperbole tend to have lunch on a daily basis, so perhaps that isn’t surprising. No, the difference is a bit more obvious than, perhaps, the difference between two power cords, but the differences do tend to fall along the same scale.

Tone is more fleshed out. It’s like that picture of the two girls I referenced recently. In comparison to the system fronted by the Mojo, when fronted by my laptop, the system seems bleached, a little lean, thinner. This is what was most annoying — I mean, it sounded fined — great, even — last week. Now, it’s meh, because with the Mojo, the sound feels more organic. Resonance and harmonics all seem more realized. Again, before you get all hot and bothered, I’m compelled to say that this is a subtle thing, and something that I didn’t key on immediately.

Reaching for specifics, the Stockfisch Records release of Chris Jones’ Roadhouses & Automobiles album is, in a word, wonderful. It has nice dynamics and is very clean, open and a pile of other adjectives that apply to recording and mastering (feel free to pick some at random, and chances are, they’ll apply). Stockfisch, as a label, really does a wonderful job with the sound quality of their product, so I keep coming back. The fact that I like the music helps, too, I guess. Anyway, I’ve talked many times about the low-frequencies that the Chris Jones disc develops, and how what I guess are the harmonics — when the loudspeakers are able to present them — really round out that low-end, producing a nice, threatening, moaning kind of rumble. It’s not just deep, it’s really full and deep. “Room filling” is one way to put it. “A 200lb Rottweiler with the neighbor’s cat” was another that I thought colorful, but useful. When the system get’s it right, that’s kinda what it’s like. Meaty. Threatening. With a little bonesnapping on the side.

When fronted by the Mojo, the bass just gets firmer. Not deeper, not fuller, firmer. Bloom, not boom. As if the strings on the bass were being struck a little more precisely, if that makes sense. On “Roadhouses & Automobiles”, those sweet intro guitar decays last a little longer, the cymbal strikes have more metal, and the crickets come “on song” across more of the song, and a little more obviously when they do. It’s not much. But it’s noticeable, though. Kinda like comparing the focus from a photograph taken with an f-stop of 2.8 and one with an f-stop of 4. The detail just goes a little deeper into the image, or rather, a bit more of it is in focus. Perhaps there’s less blur in the background. Perhaps the image stands out a little less. But, hey, is that a forest? I thought it was just a tree!

What I’m really trying not to say is this: “it’s like a veil has been lifted”. Gah — what utter nonsense. But I wouldn’t be surprised to read another reviewer put it just that way. There’s less noise, less grunge, less whatever obscura you tend to favor in your analogical soup kitchen of audio clichés. But, yeah. Just like that. The Mojo fronted system was just more resolved than the system was without it. It wasn’t life changing, but it was there. The question then becomes, is the mod worth it to you?

Mojo Audio has some options, though the website is a bit confusing, so let’s see if I can chart it out.

  • Base price: $1,800.
  • Full Monty: $3,200 (includes a new external linear PSU, the highest level of filtering available at each critical juncture, an audiophile-grade power cord, and all the RAM and SSD space your Mini can eat)
  • External PSU only w/ basic filter package: $800
  • External PSU w/Full Monty filter package: $1,200
  • Remove Apple stock switching PSU, replace with basic ingress power filtering: $200
  • Remove Apple stock switching PSU, replace with best ingress power filtering: $500

So, if I’m reading the site right, if you send in your fully configured Mini, for $1,700 you can get the best power delivery to that Mini that Mojo can make ($500 for the tear down and $1,200 for the new PSU). Alternatively, you can have them buy a new one and build it for you for more or less the difference in cost of the Apple hardware, with some nominal fee to do all the modding. Obviously, any of this makes your AppleCare policy pretty much useless.

Just an FYI, but as of this writing, Apple still doesn’t offer an SSD on the Mini — an upgrade that some audiophiles consider non-optional — so if you’re terrified of converting your new server into a shapely aluminum brick by following some online video DIY guides, this might not be a bad option.

If I were the kind of guy looking to get an edge over the current off-the-shelf offerings, a linear PSU modded Mac Mini is my kind of upgrade. I get to keep Mac OSX. I can use iTunes (if I want), or use Pure Music or Audirvana or whatever other player captures my imagination — and either use it or iTunes to manage my library. There are no fans, so the Mojo System is always dead quiet (something I can’t say about my MacBook Pro) and the thing never gets warm. You’re still going to need a monitor, a keyboard and a mouse, at least up until you get comfortable with doing some kind of remote control software — and there are plenty of options for when you feel like exploring. The Mojo is pretty much a no-brainer.

There are some other options out there, like the Paul Hynes PSU, but the mod from Bolder isn’t — the owner has stopped modding in favor of brewing craft beer (how awesome is that?!?). Anyway, I’m sure there are others out there doing this, but I haven’t any of their work to hand to do any kind of meaningful comparison. This is really the only thing that’s keeping it off my Most Wanted list at this point — I just haven’t heard enough of the competition to know its value. All I know at this point is that it is clearly better than what I have. And that’s saying a lot. Call it a “Challenger“, then, for now.

Okay, so that wraps my mini-review of the Mojo — two thumbs up! Wish I’d had more time with it, but it’s now has to wing its way to another reviewer. Just between us, I licked it before I packed it. That’s how I roll.

Associated Equipment

Mojo Audio Mac Mini Media Server, as tested

  • OS X v10.7.4
  • 2.3 GHz Intel Core i5
  • 4GB RAM
  • 120GB SSD
  • Level 2 AC Filter
  • Level 2 DC Filter
  • Level 2 Internal Filter
  • Enigma power cord
  • Cost: $2,390.70

Testing gear

Mojo Audio Joule II Power Supply

Mojo Audio Mac Mini Media Server, with Joule II power supply, on my AudiAV rack

Full system w/ Mojo Audio Mac Mini Media Server w/ KVM, next to my Raven AC-3

20 Comments on Mini-Review: Mojo Audio Mac Mini Media Server

  1. I’m thinking of buying a Mojo Audio Mac Mini Server, however, I’m not sure if it will still make a difference if I buy a USB cable with a separate leg for power (like Acoustic Revive or KingRex). What was your experience in that regard?

    • Part-Time Audiophile // May 15, 2013 at 8:15 AM //

      There are a lot of different ways to build a USB cable. I think it matters most to find one that is “well made”, and after that, it’s a matter of taste. Pretty subtle, though.

  2. This is for Dave. Any insights on why the Soulution 590 chooses to go completely bus powered?
    I have both the Soulution 590 and the BADA Alpha USB, and frankly I do not hear much difference.

    • Part-Time Audiophile // October 4, 2012 at 2:29 PM //

      YMMV, but it seems — to me, at the very least — that there are these plateaus of performance. There’s variation, sure, but the terrain is more or less flat once you’ve made it to that level. The question is, is there a path to the next plateau, and if so, where is it/what is going to open that way up? Not sure we’ve seen that — or will.

  3. It was fun to find this review the day after I purchased the Mojo Audio Server and Joulle II Power Supply. I enjoyed reading the review and these responses.

    I own the AMR DP-777 Digital Processor, and am curious to hear what Dave or Scot might have to say about this statement on AMR’s Features page.

    Under “Asynchronous 24/192 USB Input”:

    “Crucially, the USB signal is completely decoupled (i.e. isolated) from the DP-777. So any power supply noise, RFI and ground noise from the PC is not able to enter and contaminate neither the DP-777 nor its signal.”

    Having previously read about the USB data/power/ground issue, I initially looked at cables by Locus Design and Acoustic Revive. But when I read the above AMR statement, I hesitated going further down that path.

    Thus far I’ve used the supplied USB cable that came with the DP-777, but I’m still trying to make sense of this discussion regarding…

    “bus power”, “the need for a split USB cable and something like the Upower”, “there still has to be a common ground […] and that means there’s still a noise path”, “the Off-Ramp 5 and the new Channel Islands Audio Transient MKII converters”, “you can cut the USB power line entirely”, “no need for a split cable and battery”, “use a common mode choke such as the Empirical Short Block to keep ground noise at bay”, “the ground is a direct connection to the really nasty noise inside the computer, Galvanic isolation helps, but it can’t solve the problem on its own”, “the 5V DC running on the power leg has buckets of DC ripple and distortion […] and it’s running directly alongside the data lines, which you want to keep as clean as possible.”

    Can you help dial me in? Given that the AMR DP-777’s “USB signal is completely decoupled” — do any of the above quoted statements apply? Meaning, would I potentially benefit by exploring any of those cable and device solutions?

    Thanks in advance!

    And Dave, I would love to have a brief conversation with you regarding the servers you’ve mentioned. If you are interested, I can be reached at g(dot)percy(at)me(dot)com.


    • Part-Time Audiophile // October 1, 2012 at 9:37 PM //

      You could send a note to AMR to ask whether or not the ground/power lines are not connected, shorted to ground, or whatever to give you a better idea of how the electrical disconnect is made.

      Generally speaking, I don’t find my custom no-power/no-ground USB makes much difference when the upstream device doesn’t require power from the computer to run. YMMV.

  4. Norman Williams // September 30, 2012 at 5:51 AM //

    What about the Music Vault server?

  5. Yep, it’s a new world. The tricky thing about USB is that it was never really designed for this, the designers of the spec had no idea it would be used for high-end real time audio streaming. S/Pdif (especially with RCA connectors) is sort of in the same boat, it was designed to be convenient rather than sound as good as possible. In a lot of ways USB is even worse though.

    You’ve got your data +/-, your power +/- and your ground. The ground is a direct connection to the really nasty noise inside the computer, Galvanic isolation helps, but it can’t solve the problem on its own. The 5V DC running on the power leg has buckets of DC ripple and distortion, way worse than any wall-wart, and it’s running directly alongside the data lines, which you want to keep as clean as possible. The amount of ripple and distortion on that 5V line by the way is directly impacted by the quality of the power that the computer is getting. This is why laptops are usually the worst case scenario – they are powered by switch-mode supplies with their own super high levels of DC ripple, and that makes its way through the motherboard and directly to the USB port. Run the same laptop on battery power and the amount of DC ripple on the USB port will drop dramatically, and it should sound better.

    Another reason why laptops tend to sound bad is the CPU clock. Laptops just aren’t thermally designed to run their CPUs at a fixed clock, it’s always shifting depending on the work load to keep heat and noise levels down. For computing tasks that’s great, for streaming audio that’s not great. Asynchronous mode is definitely a big help in dealing with this as opposed to adaptive mode, but still, ideally the CPU should always be at a fixed frequency.

    Locus Design helped pioneer audiophile grade USB cables by using high quality metals, but just as critically, isolating and shielding the data and power lines from each other. Just as with S/Pdif cables, USB cables don’t transmit a string of imaginary 1s and 0s like something from “The Matrix”, there’s an electrical signal like any other that’s subject to interference. Even better than isolated power is no power though, which is what a self powered converter or DAC lets you use.

    You should hit up Dusty at Channel Islands, maybe he can get you a Transient MKII and VDC-5 MKII LPS to compare to the Alpha USB. PI Audio (the guys behind the superb MajikBUSS and UberBUSS conditioners) have a new USB cable with no power leg, and there are a few others around as well. The Short Block also has the effect of cutting the power line.

    Auraliti has two standard models, the PK100 which has a conventional digital output from an internal sound card, and the PK90 USB which uses the old SoTM card for USB output. SoTM just released a new version which has a single USB 3.0 output port and a DC input jack which allows it to be easily powered by an external source rather than the computer’s power supply. Sonore’s new power base has two transformers, one for the main server and one specifically for the new USB output card. Unfortunately it costs nearly as much as the server itself, so the improvement would have to be pretty substantial to make it worth while.

  6. A couple of notes to make. First, for the engineer who doesn’t understand how any changes made to the computer could possibility make an audible difference, you’re confusing one type of data transfer with another. When you transfer a Word document to a USB hard drive, the file is transferred in what’s called block mode. The data is not moved or copied in “real time”, and the OS can check the source and the destination and make sure all bits match exactly.

    Real time audio streaming over USB depends on timing accuracy. The Word document just has to get there with all of the bits in the same place, WHEN it gets there is irrelevant. That’s why using a $1,000 USB cable for a portable hard drive won’t make the Word document any “wordier”, it’s the same file.

    When streaming in real time, that same $1,000 USB cable more precisely matches the correct impedance, lessening the effect of signal reflection, just as with high-end S/Pdif cables. It may also do a much better job than a basic cable of isolating the power and data lines.

    More importantly, why does one USB > S/Pdif converter sound better than another, especially if they are both using the same asynchronous tech such as XMOS and the same driver set? Clocks are a big part of it, and the ones inside the Alpha USB are good ones, two Crystek CCHD-957s I believe, for 44 and 48 sample rates. Cheaper clocks mean higher jitter. The other major issue is power, and ground. Clocks like very clean power, and USB converters that are powered by the USB bus simply cannot give it to them. You can clean it and regulate it as much as you like, but you just won’t be able to match the clean DC provided by batteries or a linear transformer.

    This is actually an area where I think the Alpha comes up short a bit. The clocks and the outputs are powered by a transformer, but the input circuitry is all bus powered, hence the need for a split USB cable and something like the Upower. Ideally NOTHING should be bus powered. You can provide the 5V USB power line from another source, but there still has to be a common ground for the whole thing to work and that means there’s still a noise path.

    The Off-Ramp 5 and the new Channel Islands Audio Transient MKII converters can run completely on their own power. What’s critical about this is that you can 1. cut the USB power line entirely, no need for a split cable and battery and 2. use a common mode choke such as the Empirical Short Block to keep ground noise at bay. Ignoring this is a critical mistake that a lot of high-end manufacturers are still making. The $15K EMM DAC2X for example doesn’t even have galvanic isolation on its USB input!

    One of the more interesting custom made music servers available right now is the Sonore by Simple Design. It runs a stripped down version of Linux called VortexBox, which doesn’t need any kind of display or keyboard. Everything can be controlled via the web or mobile phone or tablet apps. It supports up to 32/384 PCM and DSD over USB output, has a 3TB hard drive, and is completely fanless. It’s also the only server I know of to use the latest SOtM tX-USBexp USB output card, which can be externally powered along with the server itself. It costs $2500 which is pretty reasonable, though of course it’s not the bargain that the $750 Auraliti PK90 USB is.

    • Part-Time Audiophile // September 29, 2012 at 9:38 AM //

      Nice write up — thanks.

      I think a lot of this sails over the head of the “average audiophile” (if you even can average across such a small population).

      I haven’t had a chance to play with either the Auraliti or the Sonore, though I will admit that I’ve heard nothing but good things from friends who own one or the other. My understanding is that you can get an Auraliti with the SoTM USB card, but that it might be a special order.

      As for the Off-Ramp/Short Block, I have nothing but respect for the work that Steve Nugent at Empirical Audio is doing these days, but again, I haven’t had any opportunity to try them out. Some day. 😉

    • Hi Dave, and thanks for the feedback.

      This is the part where I disagree with you: “Real time audio streaming over USB depends on timing accuracy. The Word document just has to get there with all of the bits in the same place, WHEN it gets there is irrelevant.”

      My understanding is that with async USB, the timing is totally irrelevant: basically the DAC is requesting additional data when its buffer isn’t completely filled up. In my book this is exactly equivalent to transferring data from an USB HDD or any other USB device.

      Are you aware of any article or documentation about USB DAC design that explains this in more detail and would help me clarify the above question?


      • Part-Time Audiophile // October 1, 2012 at 8:30 AM //

        I don’t have a good definitive source that isn’t written in jargon. Here’s a primer, though: USB Audio on Well Tempered Computer.

  7. Your fine review aside, the photos concern me. Stacking the components so closely together, and topping them with an LCD monitor is an invitation to EMI and RFI, IMHO. Just sayin’.

    • Part-Time Audiophile // September 28, 2012 at 5:39 PM //

      A good point. The monitor goes into standby after 30seconds, however, which is nice. And the other components on the shelf are also shut down. The EMI/RFI from PSU to computer is a concern, but the big sheet of ERS paper I had slipped between them (removed for the sake of photographic splendor) didn’t really seem to do anything. YMMV.

  8. This is a great article but it had me raise my eyebrows 🙂

    I’ve been working in software engineering for the telecom industry during the last 15 years and I still don’t get the benefits of that kind of modification for audio playback.

    If you’ve used an USB connection to connect to your DAC (or an USB to SPDIF converter) then I don’t see how any modifications made to the computer could result in a change of the playback results. As you rightly pointed out, using async USB means the DAC (or the USB/SPDIF converter) will play the music off its internal memory buffer and request more data when the buffer needs to be topped up.

    AFAIK, the stored digital musical information has two aspects that need to be accurately transmitted: the data (bits) and the timing (clock). When using USB the timing is the responsibility of your DAC since it does the playback and the Mac software has no influence over that whatsoever.

    What remains is the accuracy of the data sent over USB. Let me illustrate this with an example: when I save my personal accounting Excel file on an external hard drive connected over USB and reload it later on, I’m not expecting to see any lost values or changed cells. Why? Because that will be caught by the various error correction mechanisms (such as CRC calculation) present in the driver stack or the HDD firmware to ensure that what you saved is exactly what you’ll reload. The same applies to your digital music library or when you download a song from HDtracks: you will get the same copy every time. Another example: software downloads would be impossible without the ability to ensure 100% perfect copies!

    I must be missing the point because you seem to have found a significant improvement over the standard Mac mini. Do you have any suggestions as to why changing the PSU would bring a sonic improvement?

    • Part-Time Audiophile // September 28, 2012 at 5:36 PM //

      I still don’t get the benefits of that kind of modification for audio playback.

      I don’t get it either, if that helps any. It shouldn’t make any difference. And in many cases, “little changes” don’t — that is, I can’t tell (reliably). Piled up, well, that may be a different story.

      Anyway, if you held a gun to my head and demanded “WHY!?!”, I’d have to say the reason for just about every audible difference has to do with jitter, what kind you have and how it’s handled.

      Anyway, here’s a reading list:

      About square waves (digital signals) and how they aren’t really all that square.

      More on jitter, the kinds and where the come from. Note especially, the bit on “random jitter” for a hint about power supplies.

      Whether or not any of this is convincing is, of course, moot. Interestingly, it’s relatively simple to put to the test.

4 Trackbacks / Pingbacks

  1. new mac mini audiophile upgrade I came across
  2. RMAF12: Hi-Fi meets Computer-Fi with Mojo Audio, Volti Audio and BorderPatrol | Confessions of a Part-Time Audiophile
  3. What I Want For Christmas | Confessions of a Part-Time Audiophile
  4. My Audio To-Do List: Digital and More | Confessions of a Part-Time Audiophile

Comments are closed.