A Complete 4-Part Series
Issue 71



June 12, 2014


Korg has been miniaturising its analogue gear. Now its digital devices are getting the downsizing treatment.

Review: Brad Watts

Welcome to an audio can of worms: DSD audio. A number of years ago I reviewed Korg’s MR-1000 recorder, giving me the opportunity to make a dual recording; one to the MR-1000 and the other via a Digidesign 192I/O at 192k, and make a subjective judgement on each. The DSD recording was indeed ‘better’ to my ears, with less harsh high frequencies, more depth, and an (overall) more ‘relaxing’ experience than the high resolution PCM-based recording. But before we bandy subjective appraisals about, it’s worth reading the box on what DSD actually is — and isn’t.

So where and when does DSD become a useful medium? Recording units are touted as being the highest resolution available for two-track recording, and I’m inclined to agree, having heard the difference. DSD presents as an ideal archival format as it offers super high fidelity in an economically viable data platform, i.e. the files take up way less drive space than a 24-bit/384k PCM file. The other, perhaps more prevalent advantage is it sounds better. Really. It does. Audio encoded via DSD seems more natural, more stable and detailed. It’s smoother, richer, and more hospitable to the ear than PCM encoding. I find it far more inviting, almost addictive as you listen to any recording rendered to hard drive over standard PCM encoding. Even to the point where 24-bit/44.1k recordings sound ‘better’ when converted to DSD. I know this may seem like audiophile bunk and verging on the superstitious, but it seems that by pushing quantization noise out of the audible spectrum and adding a little distortion in the conversion, PCM-to-DSD upconversions can be more pleasing to listen to. Which is where Korg’s latest DSD devices come into play.


Interestingly, Korg’s most recent DSD offerings are straight DA units — they’re simply for monitoring. Sitting on my bench are the DS-DAC-100 and the miniature (read: portable) DS-DAC-100m. Both units are USB devices and are powered via the USB bus. I’ll focus initially on the ‘Area 51’-style DAC-100, which is an elliptical shaped unit of black anodised aluminium with three brass pinpoint feet. It even comes with brass seats for the three feet for even extra audiophile-ness. I’ll admit to being sceptical of this type of thing and after listening to the unit with and without the brass seats I remain so. Still, they’d look the shiz in any setting.

The rear of the DAC-100 is quite spartan with a USB port, left and right outputs as balanced XLR and unbalanced gold-plated RCA connectors. It doesn’t get much simpler. On the near-side are eight blue LEDs to signify the sample rate in use, all the way up to 5.6MHz, a 1/4-inch headphone output and a large volume knob for the headphone level. This is very smooth in operation and seems to be a DSP-assisted affair judging by the lack of ‘lean’ in stereo balance as you raise the headphone level from zero. Inside is Cirrus Logic’s top CS4398 DAC chip, used in the likes of Prism Sound’s Orpheus, the Universal Audio Apollo, and Korg’s own MR-1000.

Korg seems to be positioning the unit as a kind of transitional device between both audiophiles and the recording fraternity, hence both XLR and RCA outputs, and it’s because of this I wish the headphone level control could be switched to provide attenuation over the rear outputs rather than the headphone output only. Considering the units both offer a CPU-based control panel I feel Korg has missed a trick here — especially if it’s attempting to endear themselves to recordists and high-definition obsessed engineers. For now, you’ll require an attenuating device between the output and your monitors, adding further connectors and components before the final sound output.

The DS-DAC-100m is tiny by comparison, and uses a single 1/8-inch stereo output connector, and 1/8-inch headphone output connector. USB connection is also of the mini variety. Volume control over the front panel headphone out is via a pair of ± buttons. At 92 x 129 x 20mm it’s about the size of a ‘phablet’-style smartphone, so you’ll throw it in your every daily carry kit without a hassle. Sound quality-wise it’s identical to the DS-DAC-100 according to spec, and indeed, I sensed no difference between the units sonically.


Hell, yes! Recordings made in DSD and played back via appropriate (DSD) converters are inspiring, to say the least. In order to audition these units I rummaged around various contacts and was supplied a number of recordings made primarily with Merging Technology Pyramix systems. Listening via this medium is, as I previously espoused, more organic, with far greater depth, detail, and ‘quality’ than one could expect from PCM-based mediums. Even the indulgence of converting 24-bit WAV PCM files to DSD resulted in audio quality I found to be almost addictive in nature. I simply wanted more and more.

You can use Korg’s Audiogate software to do it in real-time, up to 5.6448MHz. Though it uses computer juice, and the conversion can’t be flown-in between DAW software outputs and the device’s driver.

No doubt these units will be indispensable for Pyramix users as a quick access point for monitoring DSD projects. And for a travelling DAC/headphone amp, to get you away from your laptop output, look no further than the mini version. Again, anyone seeking state-of-the-art audio playback owes it to themselves to take a listen to either of the Korg units.



DSD, or Direct Streaming Digital, when compared to Pulse Code Modulation (PCM), is a completely different method of encoding analogue audio into digital data. Instead of registering the amplitude within the bit-resolution of 16-bit for example, and sampling that amplitude a number of times per second according to the sample rate (in the case of CD that’s 44,100 times per second or 44.1kHz), DSD samples at an extremely high number of cycles per second — 2.8224MHz or 2,822,400 times per second (5.6448MHz is also possible but most DSD recordings you’ll come across will be 2.8224MHz). It also uses only a single data bit and instead of measuring the amplitude from zero each time DSD measures the change in amplitude from the previous sample. If the amplitude is greater than the last sample then a 1-bit is registered, less than the previous sample and a 0-bit is registered. Consequently a rising portion of a waveform would consist of many consecutive 1-bits, a falling waveform would consist of many consecutive 0-bits and a flat section of waveform would register as alternating 0 and 1-bits. It’s a form of quantum measurement, and therefore relies on ‘knowing’ where the previous point exists before the next point can be stipulated — each point is measured from the previous, with all points relying on the others (alternate universe anyone?). The advantages of the concept are a high sample-rate representation of the waveform without requiring scads of drive space — 2.8224MHz DSD uses about four times the storage as 16-bit/44.1k data, and only about 20% more than 24-bit/96k. Equally advantageous is the fact the 1-bit sampling conversion used for DSD doesn’t require the use of digital filters. In a PCM-based system there’s some sophisticated filtering going on, both on input and output. This can mess with the audio quality greatly and is a prime attraction for proponents of the DSD format. However, the primary disadvantage with DSD is the waveform can’t be edited like PCM data. As soon as you ‘cut’ the waveform the system no longer knows where zero amplitude is, and the waveform ceases to exist as it was originally captured. 1-bit data is also a complex proposition when it comes to digital signal processing, consequently it’s going to be quite a wait before DSD finds its way into your garden variety DAW.


Leave a Reply

Your email address will not be published. Required fields are marked *

More for you

A Complete 4-Part Series
Issue 71