Music in the Round #93: miniDSP & Ripping SACDs

As I took my valedictory lap around High End, the immense audio show held each May in Munich, Germany, it was clear that this year's event was an exuberant flowering of mature technology. I witnessed the dominance of hardware for LP playback, as well as analog amplifiers, many of them based on tubes, and passive loudspeakers with traditional cone-and-dome drive-units. And there was no shortage of excellent and impressive musical demonstrations. Still, I experienced no revelations, and heard no announcements of any new technology that might trigger a hopeful anticipation of the near future. It was as if HE2018 were reflecting on the past with reverence and commitment, rather than striving toward the future with innovation and adventure.

I know that I'm no longer a mainstream audiophile and that, compared to the bulk of the products shown at Munich—or, indeed, reviewed in the pages of Stereophile—I've gone off the rails. But the older I get, the more I prefer to listen to music without the constraints of old tech. The things that grab me are products that are the results of fresh ideas, critical reexaminations of accepted technology, and/or direct responses to market demands, whether obvious or as yet unrecognized. For the playback at home of multichannel music files, I saw nothing at High End 2018 that was as exciting as what follows.

miniDSP U-DIO8 multichannel USB to AES-EBU or S/PDIF interface There has been a serious impediment to the spread of multichannel playback among audiophiles. We have players for multichannel discs and servers/streamers for multichannel files, but at the outputs there is a fork in the road. One road—the HDMI inputs of audio/video receivers and preamplifier-processors—is an option generally avoided by audiophiles, though there may be no good reason for the rejection—there are indeed such products with high sound quality. The other road leads to a paucity of multichannel DACs. To repeat a rant from past columns, there are only three purpose-designed multichannel DACs: the inexpensive but capable miniDSP U-DAC8 ($299), the excellent exaSound e38 ($3850), and the impressive Merging Technologies NADAC+ ($12,000). But considering the prices, how much real choice is there for any individual potential buyer?

I understand the reluctance of the makers of audiophile gear to commit to developing a multichannel DAC. It's a problem of chicken and egg: In part, the market is small because there are not enough such products out there to create interest—which is why there aren't enough interesting products. A logical solution would be to make a workable setup from three independent stereo DACs—something Mytek HiFi took a stab at a while back (footnote 1). They offered proprietary solutions in which three of their stereo DACs (originally the Stereo 24/192, and more recently the Brooklyn) could be fed from a USB source via a USB hub to function as a six-channel DAC, as long as the master clock of one was daisy-chained through the others to keep all three in sync (footnote 2). Brilliant and successful, but wouldn't it be cool if you could press into service any of the many two-channel DACs now on the market? The problem is keeping the three DACs synchronized—and the vast majority of consumer DACs have no word-clock inputs and outputs.

Benchmark Media Systems has come up with a solution. They recommend a configuration in which a Lynx AES16e PCIe sound card ($699) provides 16 channels in and out at up to 192kHz. This card outputs two channels from each of its multiple AES/EBU outputs, and since each output carries data driven from the same onboard clock, all connected DACs are effectively synced to each other. Clearly, Benchmark thought it good enough to use in demos at hi-fi shows, and I'm told the results were impressive.

Benchmark warns that while this works with their own DACs, it's not a hard-lock sync. Due to differences among the client DACs, and thermal effects that accompany long-term playback, there is the potential for drift. In addition, a professional soundcard is not the practical solution I'm looking for because it must be installed in a PC; it can't be installed in a dedicated streamer or player.


Now there is such a solution: the miniDSP U-DIO8 ($299, including one multichannel output cable), which is available with either S/PDIF or AES/EBU output, both up to 24-bit/192kHz, and will connect with the USB output of any computer running Windows, Mac OS, or Linux (footnote 3). It's no surprise that it's the brainchild of Tony Rouget and miniDSP, who've been responsible for getting fiendishly simple and effective equalization products, some including Dirac Live, into the hands of adventurous audiophiles (footnote 4).

The miniDSP U-DIO8 is small: 6.3" wide by 1.2" high by 4.3" deep. It is also inexpensive: $299. Its front panel has a USB-B input and a blue LED power indicator. On the rear are a barrel connector for the 5V wall-wart power supply, and a DB25 connector that uses the Tascam-default pin layout for eight outputs and eight inputs of S/PDIF or AES-EBU, depending on the version purchased. Both versions of the U-DIO8 come with a suitable breakout cable: DB25 to eight XLRs, or DB25 to eight BNCs. (Buyers of the S/PDIF version must also buy three or four BNC-to-RCA adapters.) The only visible differences between the two versions—miniDSP sent me one of each—are a label on the rear panel and the connectors on the accompanying cable. No doubt miniDSP puts the input in front and the outputs out back for design convenience and for traditional signal-routing conventions but I think most of us would prefer to have all connectors on the rear panel.


The specs say it all. The USB streaming engine is an XMOS 16-core device that needs no driver for up-to-date Mac OSes and is compliant with Linux ALSA 2.0, but for Windows machines it requires the download and installation of an ASIO driver. This is all as expected, and the user manual provides clear guidance. The eight channels of digital output use four connections, S/PDIF or AES-EBU, and support word depths of up to 24 bits at sample rates of 44.1 to 192kHz. The U-DIO8's specified dynamic range is 128dB, and its THD+noise is –125dB. There are also eight channels of digital input via the same cable snake. These accept input sample rates from 32 to 216kHz, which might be useful for some calibration or measurement applications.

How to test something that has no controls, no options, and does nothing to the signal but convert it from USB data to an audio datastream? First, to determine if the U-DIO8 works as advertised is easy: Feed it a multichannel source, connect it to any three (or four) DACs and listen for audio from all channels. Check! Second, to find out if it is good enough that any user's choice of three attached DACs would perform to their full advantage, I used my three Mytek Brooklyn DACs with the U-DIO8 connected to my MacBook Pro computer. Double check! There was the clean, balanced, multichannel sound I expected.

However, to appreciate the capabilities and contributions of the U-DIO8, I had to try a number of different DACs to reassure myself that it wasn't constraining their sound in any way. Although the products I tend to cover in this column rarely end up on John Atkinson's test bench, I got him to measure the U-DIO8. Here's what he said:

"To put your mind at rest, I have attached two graphs. Fig.1 is a digital-domain FFT analysis of a loopback measurement from the Audio Precision's S/PDIF output to its S/PDIF input, with the test signal 24-bit J-Test data sampled at 44.1kHz. Fig.2 is with my MacBook Pro feeding the same 24-bit J-Test data to the miniDSP via USB and the latter's S/PDIF channel 1/2 output feeding the Audio Precision's S/PDIF input."


Fig.1 Audio Precision SYS2722, high-resolution, digital-domain spectrum of 24-bit J-test data (11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz): S/PDIF output fed to S/PDIF input (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.


Fig.2 miniDSP U-DIO8, high-resolution, digital-domain spectrum of 24-bit J-test data (11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz): USB input fed to S/PDIF output (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.

I saw no differences between the two graphs. John added that "the results were the same at 96kHz," and that the miniDSP "produces a bit-accurate output." That's good enough for me.

To appreciate the impact of the U-DIO8, beginning with the S/PDIF version, I decided to try several DACs spanning a wide range of prices. In addition to the trio of Mytek HiFi Brooklyn DACs (original version, $1995 each) already in-house, I borrowed from Schiit Audio three samples of their Modi 2 Uber ($149 each), from Cambridge Audio three samples of the DACMagic Plus ($349.99 each), and from Benchmark three samples of the DAC3 HGC ($2195 each).

The MacBook Pro was running JRiver Media Center (v24.0.30) and Roon (v1.5, build 323), each of which identified the output device as "USBStreamer," with audio-processing attributes up to the maximum resolution of the U-DIO8: 24/192. I used my Mac's Audio MIDI Setup to check that the channel mapping was correct, and clicked to play a 5.1-channel file. It was alive!

I compared the sound of each set of identical DAC triplets with that of my resident exaSound e38 and miniDSP U-DAC8, taking notes throughout. I wanted to determine whether, within the U-DIO8's range of resolution of up to 24/192, it would obscure any differences among the various trios of DACs, especially the sorts of differences that would make a shopper choose one DAC over the others. I didn't do this in the spirit of a DAC Shoot-Out, because that might demand a proper blinded test. Consistent with JA's measurements, my listening comparisons between a direct S/PDIF connection and sending signals via USB to the miniDSP and via S/PDIF to each DAC (in stereo) revealed no differences, and thus I had no preference as to which type of connection I used. The sonic signature of each model of DAC remained audible.

I began with the three samples of Cambridge DACMagic Plus, a well-received DAC that's relatively old but still in production. The sound, in the familiar context of my Manhattan system, surprised me: somewhat warm and spacious, with lovely detail and dynamics. The DACMagic Plus was much more relaxing to listen to over long periods than the miniDSP U-DAC8, which could develop a tiring brightness. The exaSound e38 didn't sound as warm as the DACMagic, but its highs seemed better resolved. At a total price of $1348.97 for three DACMagic Pluses and a U-DIO8, this combo struck me as the multichannel DAC the market has been waiting for, and one that I'd consider running through the multichannel analog input of my Marantz pre-pro. This will bypass the Marantz's digitization and processing while taking advantage of its volume control.

I got a similar impression from Schiit's three Modi 2 Ubers, but this time the pleasing warmth tended to soften details and dynamics more than did the exaSound e38 or the Cambridge DACMagic Pluses. It was a reasonable trade-off for the sharper detailing of the U-DAC8, and I could say that the differences between the Modi 2 Ubers and the miniDSP U-DAC8 were so distinct that personal taste in sound will be the most important factor in choosing between them. The total price of three Modi 2 Ubers and a U-DIO8 is $746, about halfway between the prices of my base reference U-DAC8 and the Cambridge-miniDSP package. I can easily imagine that many would find its sound and price just right for their taste and wallets.

Footnote 1: There was a third route, based on using the multichannel analog output of a disc player, but that road has been increasingly less traveled due to the extinction of players with more than two analog output channels—or, for that matter, any analog outputs at all. Oppo Digital's recent decision to cease making disc players was the final straw.

Footnote 2: Playback Designs offers a similar but more elegant and expensive arrangement using their Designs USB-XIII Digital Interface with three Merlot DACs. See my September 2017 column.

Footnote 3: miniDSP, Hong Kong. Tel: (852) 2358-2066. Web:

Footnote 4: While writing this column, I learned that miniDSP will release its own multichannel-capable SHD-series (Streaming High Definition) streamer-players with Dirac Live and Volumio built in.


rt66indierock's picture

The interface sounds like a simple answer to a problem. I can't wait to try one. Thanks

Did you consider you may be cruising along enjoying the scenery and everybody else at Stereophile is off the rails?

See you at RMAF.

JR_Audio's picture

Hi Kal. A nice write up. I enjoy reading your multi channel efforts. At AES conventions, I have listened to some really good surround setups, but at HiFi shows, you will find mostly (or nearly only) 2 channel playback (leaving AV setups aside). Multi channel audio is (when done right), so easy for the ears (your brain has by far less work to recreate a sound stage from phantom sources). And one further step into realism would be wave field synthesis, that I have heard at some AES and at some “Tonmeister Tagungen” and they deliver a very realistic sound stage. Juergen

AB's picture

Thanks for telling us about this new product. It opens up a lot more options. I am currently using an Oppo BDP105D for MC streaming from NAS and although it is good, I feel the need to improve on its sound and convenience, my old and temperamental Sony SCD-XA9000ES stills gives me the best sound from SACD or CD, giving me more detail and air than the Oppo, whether from flac, DSF or SACD.

Can you give an opinion where the Oppo sound fits in to the various DACs you listened to with the miniDSP U-DI08?

Kal Rubinson's picture

Can you give an opinion where the Oppo sound fits in to the various DACs you listened to with the miniDSP U-DI08?
I have not used any Oppo as a DAC in quite a while but it is a relevant question. When there is an opportunity to try it, I will.

AB's picture

Thanks Kal.
A lot of people must be familiar with Oppo players, so they make a good reference point when trying to get an impression of the performance of DACs and other equipment in reviews.
Thanks for all the reviews over the past years, they have been invaluable in helping me with my multi-channel music system.

Kal Rubinson's picture

Here's a link to instructions for ripping SACDs with certain Sony SACD players.

jeffhenning's picture

When I was assembling my latest system a couple years ago, I was using an Oppo BDP-103D as a front end into a Parasound A23 amp and KEF LS-50's. While the gain staging wasn't perfect, it did sound really good.

After using that set up for a few months, I got an Emotiva XMC-1 surround processor. The sound improved. A good part of that may be the lower noise from improved gain staging. My impression was that there was more going on. Recordings seemed better resolved by the XMC-1.

Bottom line: a BDP-103 is really good, but it's no replacement for a dedicated DAC or processor. Not surprising since most of the money went toward the video processing.

Whether I'd feel the same about their BDP-105 with its beefed up analog outputs is another story. I bought the 103 with every intent of hooking it up to a processor. It made no sense to spend the extra $700 for the 105's audio circuits when I wasn't going to use them.

Kal Rubinson's picture

Bottom line: a BDP-103 is really good, but it's no replacement for a dedicated DAC or processor. Not surprising since most of the money went toward the video processing.

I will not dispute your observations but I do not accept that argument at all. One might say exactly the same about the XMC-1 referring to its video processing, DSP and HT codecs.

jeffhenning's picture

Kal, I have no wish to get into a flaming session so I'll keep this brief.

An Emotiva XMC-1:

• Has no video processing - all of its HDMI facilities are pass-thru so, on that count, it's a giant HDMI hub and the Oppo units are small hubs (I started out using the 103 as the center of my AV system)

• Has no more or less support for HT codecs the 103

• Has no DSP other the Dirac Live room correction and tone controls and all the heavy lifting for room correction computation is done by a computer, not an on-board CPU

• Has balanced outputs while the 103 has none

• Costs 5 times as much as a BDP-103 and, even though it has way more outputs, it doesn't surprise that they sound better

My observation was made after listening to the old configuration during unpacking, system planning and a 5 minute break to add the XMC-1 so the sound of the 103 was fresh in my ears.

Kal Rubinson's picture

OK. I guess I should have known that since I reviewed the XMC-1 a while back.

OTOH, I generally do not buy that argument anyway, especially with regard to significantly different types of equipment, e.g., a universal player vs. prepro.

doak's picture

An FYI for those interested in SACD ripping:
The Pioneer Elite BDP-80 can be used to rip SACDs with the same instructions and software cited for the Oppo machines.
I’ve done this successfully and know first hand that it works well.

stevebruzonsky's picture

For years now I've ripped SACDs both 2Ch and MCh tracks and I've got over 800 Mch albums alone on my server. About 5 years ago I had Small Green Computer build a custom CAPSv3 HDMI (and USB out for 2Ch) music player, initially using JRiver MC, and for past years using ROON. My current setup is a Sonic Transporter i7 ROON Core, Sonore Signature Rendu SE, Berkeley Audio Alpha USB to Digital AES/EBU, to Theta Digital CBIV-A for 2 channel; and the Sonic Transporter i7 ROON Core with CAPSv3
(with an AMD Sapphire video/HDMI card) for MCh music. Although MCh music from my server out through CAPSv3 sounds very good, my 2Ch USB music simply sounds much better almost live at times!

Right now the Trinnov SSPs (pricey) are ROON Ready and my understanding is they will handle at least 5.1, perhaps 7.1, simply with the ethernet interface thanks to ROON. And Theta Digital is just starting to work on upgrading for the next Casablanca version to be ROON Ready and once I have this I will be able to ditch the Sonore Signature Rendu (2Ch USB), Berkeley Audio Alpha USB Converter, and CAPSv3 in favor of straight ROON Ready ethernet connection from SonicTransporter i7 (ROON Core) to hopefully some day sooner than later Theta Casablanca V (next model to come out). At CEDIA I talked with DCS, as their new 2Ch DACs all are ROON Ready, and they tested ethernet vs USB and said it was NO CONTEST, ethernet is much better sounding!

This is the future! EXCITING! When this occurs, not only will the ethernet connection to the CBV Roon Ready sound even much better for my already glorious sounding 2Ch, but as MCh will ditch the poorer sounding HDMI connection (no one manufactures a more audiophile output HDMI card) my MCH will sound as good as my sterling 2 CH!

Of course, blu ray and 4k discs are still stuck with HDMI. If only someone would develop a ROON for 4k blu ray to handle both audio and video with the appropriate hardware, all over ethernet. I imagine it will happen some day but father into the future. Even Kaleidescope and its hard drive/streamer for 4k blu ray uses the antiquated HDMI connection!

This is the future! EXCITING!