This review page is supported in part by the sponsors whose ad banners are displayed below
Tech support! After a recent more complicated 3rd-party forwarding call-tag pickup by DHL—all domestic shippers know 6moons well—I got an unexpected call. How would I rate the conduct of their driver who'd picked up the parcels and handled the paper work? I gave him an honest 10 out of 10. The caller was clearly pleased. And I'd do the same if Ed Meitner called to inquire about their tech support. Since it's germane to our topic, let's get more specific.


Updating the system and XMOS firmwares involved nothing more than unzipping two folders. Tech support had promptly emailed those and then talked me through their installation by phone. The folders contained the files at left. While my iMac didn't need the Thesycon driver for playback, the update at the time was accessible only from a Windows computer. And that did need the driver from whose control panel the firmware upgrade would be made. That simply meant pointing the firmware image file option window at the proper dsd-out-xmos_345.bin file. This quickly overwrote the temporary version my loaner had come with to bridge the gap.

If this wasn't sufficient demonstration on how today's digital designers must be fluent in code to keep up with the ongoing unpredictable changes in the operating systems or files which their products get associated with, here's another. On my year 2011 iMac running OSX 10.7.3—Meitner's tech still ran 10.7.1 which required updating so he could duplicate my issue—turning off the MA-1 and server meant that upon power-up the next morning, AudioMidi couldn't find the Meitner as selectable audio device. I had to unplug/reseat the USB cable for the MA-1 to reappear: "This is a bit speculative but our embedded designer says that in theory the only way for this to happen is if OSX sends audio packets to the USB port while the MA1 is already off but before OSX has gone to sleep. This is not a problem for most USB audio devices. In our case with the MA1 off the XMOS has no clock and won't respond. It's as though a cable were unplugged. OSX decides the device is no longer there. AudioMidi defaults to something else. Normally when a DAC comes on and enumerates, a system device manager reverts and apps are not affected. That seems to be what Linux and Windows do. Not AudioMidi.


"We likely didn't notice this earlier because we rarely ever turn DACs off while connected to a host. Since we have hundreds of MA1s in the field and not one such report, most users probably leave their converters on or run Windows or Linux. In OSX we mainly use Audirvana. There Damien decided to routinely re-initialize the audio device port no matter what. This seems to ameliorate this behavior." Since I too was using the latest Audirvana, that wasn't it. A day later: "Testing continues but we have already confirmed your issue. It only occurs in OSX 10.7.2 or 10.7.3 when it sends audio packets to an MA1 already turned off. In Windows and Linux it doesn't appear to be a problem. We've asked the Coreaudio folks what if anything they suggest. We'll try different solutions to see what's best. For now I can only recommend that when shutting down the system, turn off the Mac first. And when powering up, start the MA-1 first." The upshot for listeners who neither need nor want to understand how any of this works as long as it does? Meitner employs a proper in-house engineering team. It stays on top of this ever-shifting sector to keep customers happy.



DSD. After downloading the song "Looking for a home" from Blue Coast Records in three formats, I received this email: "By your purchase it looks like you're going to compare format differences. The 9624 is made from the conversion of the DSD master. The 44.1 was compressed a little during mastering for CD release. It's a modified conversion of the DSD master. Unfortunately the CD layer used compression for better playability in the car (now we realize this was a bad idea) and so is not a fair test. Instead a better example for comparisons is Art Lande's While She Sleeps. I no longer compress the CD layer on SACD masters. All formats are mastered the same. You should be able to distinguish formats by listening to the percussive decay of the instruments and the stereo image (which grows larger with more samples). DSD is incredible. You might also be interested to check out DSD-Guide.com, our newest information and free DSD download site. Also DownloadsNOW.net has a list of high-resolution audio we control. Enjoy - Cookie Marenco". This was a good reminder. Don't come to any 'ultimate' format conclusions whilst comparing different files unless you know exactly how they were derived and what was done to them.


I also downloaded two free tracks from BluePort Records in 44.1kHz, 96kHz and DSD versions. The above shows how once the tracks were acquired—a 5:57 track in DSD was 473MB and took ca. 14 minutes to download—they needed to be book-marked in iTunes which had them appear as ALAC files. Downloading an entire album in DSD will definitely keep you waiting. The PCM files arrived as WAV. I instantly rewrote them to AIFF to attach artwork (I dislike missing art even for solo tracks). I also rewrote meta data to attach resolution figures to the track names and organize them in common folders. I then contacted Morton Lindberg of Norwegian record label 2L for additional material and to ask how his various files are created.


"The outputs from our microphones are amplified by a Millennia Media mic-pre and fed direct to a Digital Audio Denmark AD converter. Sampling within the AX-24 is actually a hybrid between DSD and PCM via a 5-bit delta sigma modulator at 5.6448MHz. From this raw stream we can decimate to either 1-bit 2.8MHz DSD or 24-bit 352.8kHz DXD. The latter is branded Digital eXtreme Definition by Philips and Merging Technologies. The samples on our free Test Bench are produced entirely within the DXD resolution. Then DSD and the various PCM files are each derived from the original DXD with the Weiss Saracon software to be perfect for direct comparisons. DSD has excellent qualities as a delivery format to end users. It gives a fine balance between file size and playback quality. But to me high-resolution PCM is what most accurately represents the original analogue output from our microphones. We have recently broadened our horizon with an acoustic Jazz project. As this album is not due for release until September, I have prepared a special sample for you. These five resolutions of the same track should make for a good comparison."



By how DSD files import into iTunes, their bookmarks do not display true data density. They merely contain non-musical data which point at the real files located elsewhere on the drive. For the real resolution figures refer to Morton's download links. Those I embedded in the lower left of the above collage. From the 24/176 packaging spec we already knew how a standard DSD file's density falls right between a 24/96kHz and 24/192kHz PCM file. In this example the 96kHz version carried 124MB, DSD64 had 144MB and 192kHz showed 236M. DSD128 at twice the DSD64 rate increased to 289MB whilst 24/352kHz packed a whopping 433MB - well in excess of even DSD128 and thrice higher than standard DSD.


PureMusic and Audirvana both processed dsf/dff files without hiccup whilst the Meitner jointly lit up its 176 and 192 LEDs to offer hardware confirmation. PureMusic's display confirmed DSD over USB with 64fs Native DSD Playback, Audirvana with DSF 1bit/2.8MHz (DSD64). Because the MA-1's XMOS USB transceiver is limited to 192kHz, both software players automatically downsampled 24/352 PCM files to 24/192. (When I opened one of these files with another program, my iMac displayed its code like shown in the above insert.) Long story short, the streaming DSD code really had been cracked.
Enlarge!