dCS 972 D/D converter Page 2
Fast forward a bit. (A bit is an undefined unit of audiophile time in which a reviewer waits and a vendor endlessly repeats "Any day now...") Expectations were high as I sent the 24/96-capable 972 and Elgar (serial number ELG289) I'd been auditioning back to the UK for the 192kHz upgrade. Once they'd returned (a "bit" later), I warmed them up and, raising the bar as high as possible, cued up some vinyl on the SpJ/La Luce. Settling into the Ribbon Chair, I turned my attention to a document entitled "A Suggested Explanation for (Some of) the Audible Differences between High Sample Rate and Conventional Sample Rate Audio Material," by Michael Story, Managing Director of dCS.
The long and short of it seem to be that the gentler slope of the anti-aliasing digital filter at 192kHz avoids the ringing endemic to the sharp filtering required at 44.1kHz: 100dB down at half the sampling frequency, or 22.05kHz. Furthermore, according to Story's white paper, "It is worth noticing that the ringing sets in at quite a gradual slope. It's not the extreme severity of the sharper filters that causes ringing, but the move away from a very gentle filter function."
Interesting. Then this:
"One can get oneself into a bit of a twist thinking about the energy in the ringing. After all, if it is in the audio band, allowing extra energy at higher frequencies through the system surely cannot cancel out some of the energy that is in the audio band? It does, though—so although we may not be able to hear energy above 20kHz, its presence is mathematically necessary to localize the energy in signals below 20kHz, and it is possible (and our contention) that we can hear its absence in signals with substantial high-frequency content. A high-sample-rate system allows it through (fact)—and allows the high-frequency signals to sound more natural (contention) but allowing better spatial energy localization (fact)."
The localization "fact" is derived from remarks of those auditioning the effects of the upconverted signal. The paper, replete with supporting graphs, should be required reading for anyone interested in the technology.
A bit of description
The 972 Digital to Digital Converter is the Swiss Army Knife of dCS's pro audio line, and looks very much the pro gear that it is. Basic boring black, no fancy casework, a plethora of inputs and outputs at the rear, and an informative, backlit, menu-driven LCD display front and center. There are two rows of four pushbuttons to the left of the display, and a rotary encoder with a series of red LEDs to the right. The upper rank of buttons is used in conjunction with the rotary encoder for navigating and selecting items from the menu. The lower rank is used for storing and retrieving setups.
Understand that the 972 does a lot of stuff, including noise-shaping and dithering, gain, sync source selection, and phase and balance control. However, what concerns us for high-end audio applications are sample-rate and format conversions. The unit provides five digital input and output formats: AES/EBU (XLR), Dual AES (XLR), S/PDIF (RCA, BNC, TosLink), S/PDIF (optical), and SDIF-2 on BNCs. The Dual AES interface allows a 24-bit/192kHz signal to be encoded as two 24/96 AES datastreams and decoded back into a single datastream. I tried Single and Dual 96kHz AES; Dual sounded better.
Learning to navigate the menus and select operating modes took no time at all. As both the Elgar and 972 have an Auto feature, little is required of the user once basic choices have been made. In fact, the only time I futzed with the unit was when changing inputs from the Theta DaViD transport or comparing sampling rates and bit depth.
While the 972 is available now for $6995, Gary Warzin of Audiophile Systems, the new US distributor of dCS audiophile products, tells me that a less expensive consumer version, the Purcell, will be introduced in spring 1999. Expected to come in at around $5000, it will be more focused on the needs of the high-end audiophile market.
The naughty bits
So what is it about 16/44.1 that improves with upsampling? Many visitors chez-10 first noticed the liquidity of the presentation. Getting (far) away from that sense of chopped-up and reconstituted digital data we've come to (mostly) grin and bear imparts a remarkable ease and musicality to the presentation. My intellectual "body" always untensed when listening to the 972/Elgar combo, relaxing me completely into the utter clarity and detail, the enveloping liquidity and transparency of the music. Others were struck by the superb imaging and sense of air.
But it's the midrange that's most vivid in my mind. I just couldn't believe how fine it was on a wide variety of recordings. The mids were so fully developed, so complete, so utterly engaging, attractive, and musical, that I couldn't pull myself away. The inspiring richness put paid to the conundrum, "Can digital ever be truly musical?" Because musical it was, in spades.