A.R.T. scored a huge hit a couple of years back with the Tube MP desktop mic preamp. Now the company has married digital converters to a 12AX7a tube to create the DI/O, a 2-channel, 24-bit A/D/A conversion box with variable tube warmth, switchable sampling rates as high as 96 kHz, and intelligent digital synchronization (see Fig. 1).
The DI/O is designed primarily to provide a high-quality alternative to the digital converters found in computer sound cards — a worthy idea if only because it takes the conversion process out of the noisy, interference-prone CPU environment. The unit can be used for various applications, including warming up digital audio, improving monitoring from digital devices such as DAT recorders (particularly older models), and acting as a tube stage in analog signal paths.
With its sturdy metal enclosure and easy-to-read silk-screen lettering on the front and top panels, the DI/O bears a strong resemblance to A.R.T.'s other desktop processors. Like those units, it is designed to sit on top of something — there is no easy way to rack it. What's more, information about sampling rate and clocking status is printed on top of the unit, reinforcing the designer's intention that it reside on a level surface.
The front panel provides a pair of continuously variable rotary knobs: one for Input Gain (from 0 to +20 dB) and one for Tube Warmth (from Clean to Warm). The only other control is a button on the left for selecting sampling rate and clocking source. Two columns of green, yellow, and red LEDs complete the front-panel offerings. Those on the left (beside the button) display the selected sampling rate and whether the DI/O is operating as the master clock or slave; those on the right (labeled 0 dB, WRM, and -20 dB) indicate input level and the amount of signal flowing through the tube.
INNIES AND OUTIES
The DI/O's rear panel provides two unbalanced ¼-inch inputs and two unbalanced ¼-inch outputs; two RCA jacks supply S/PDIF digital I/O. The DI/O can sync to an outside digital clock carried in an incoming S/PDIF data stream, or you can use the unit as a S/PDIF master clock. As with most S/PDIF-clocked devices, the DI/O will slave to nearly any clock rate you throw at it, from 22 to 100 kHz. Unfortunately, the unit does not support word clock, which means you can't use a single clock source to maintain peace among a multiplicity of digital gear.
A wall wart with a fairly short cable supplies power. The review unit arrived with a mismatched power-supply plug and jack that caused the unit to short if the cord was jiggled. So I called on one of the four elemental forces in the universe: duct tape.
MEET THE METERS
Reading input levels on the DI/O takes practice. At first glance, the LEDs seem to act as a simple ladder meter: there's a green LED at the bottom, a yellow one in the middle, and a red peak-warning LED at top. The green LED glows when the DI/O detects a signal of at least -20 dB; however, the middle light doesn't indicate signals between -20 and 0 dB. To quote from the manual: “The yellow (Warm) LED comes on whenever the tube is adding warmth or character to the signal. This is a function of both the signal level and the Tube Warmth control.”
I would have preferred three-stage input metering along with a separate indicator for the tube circuit. I also would have liked an output-level control; it was sometimes difficult to balance the Input Gain and Tube Warmth pots to set output level. Even with some built-in digital headroom (the peak LED lights at -3 dB Full Scale, not 0 dB as indicated by the manufacturer), I still managed to fry the audio going to my computer more than once.
I have to say a bit about judging the sound of digital converters. In the first place, other factors are involved besides the quality of the converters (see the sidebar, “Sounds Good to Me”). Also, it's often difficult to hear subtle differences in the less-than-pristine environment of the typical desktop studio. For that reason, I want to stress that my analysis is subjective.
When testing the DI/O in my studio, I always listened with the unit set both as master and as slave, monitoring adjacent channels on a Mackie 1604-VLZ mixer through a pair of Alesis Monitor Ones driven by an Alesis Mattica 500 amp. At other locations, the listening environment was similar.
WHAT'S THE WORD (LENGTH)?
I decided first to see how the DI/O's D/A converters stacked up against those in a proven workhorse: the Panasonic 3800 DAT recorder. I cabled up with fresh S/PDIF cables, popped in a recently finished mix of a blues band, and prepared to be unimpressed. After all, the 3800 cost me a lot of dough, and the DI/O lists for less than $250.
Boy, was I surprised. Here are my notes from that first listening session: “3800: Big. Bold. Bass kind of washed out? Some splashiness in cymbals. DI/O: Tighter bottom. Cymbals more distinct. Sound field wider, more transparent. What's going on here?”
The DI/O rocked; when I dialed in tube warmth, I really sat up and took notice. As levels approached saturation, I heard telltale tube compression and grit, just the thing to capture a '60s vinyl vibe. I repeated the tests with a different demo; again, the DI/O impressed me. At the Clean setting, it flattered the female singer and did delightful things to the spaciousness of the reverb.
I did both tests with 16-bit, 44.1 kHz material — no problem when using the DI/O's 24-bit D/A converters for monitoring. However, when using the unit for recording to a 16-bit medium (such as a DAT recorder), the final 8 bits of each sample necessarily get truncated, effectively eliminating many of the advantages of the longer word. Likewise, when recording to a 20-bit medium, the final 4 bits get truncated. The DI/O's lack of dithering prevents the usual solution of dithering samples down to desired lengths.
Even so, the DI/O's A/D converters compared favorably to the 3800's. For example, on an Appalachian mountain dulcimer with the DI/O's Tube Warmth control set to Clean, the DI/O recording better captured the sparkling transients of the pick hitting the strings, without losing what little lows the instrument produces.
Notoriously difficult to record, the Irish bodhran (a frame drum) often comes out sounding like a pair of tennis shoes rumbling around in a dryer. In this case, the results were even more pronounced. The DI/O won hands down, better capturing the thwack of wood against skin and the wonderful low end the drum is known for. Dialing in the tube added the final touch — it made the drum sound big, warm, and fat. I certainly want a DI/O for my next hand-percussion session.
Impressed by how good the DI/O's A/D and D/A converters sounded, I was keen to find out how well the unit functioned in its intended role: as the digital front end for a computer sound card. Used in that manner, the unit essentially replaces the sound card's converters, so the results depend on the card.
The DI/O provided a distinct improvement in sound to a popular, inexpensive sound card I compared it with. The kick drum and snare sounded pleasingly fatter, and a crash cymbal's decay lacked the graininess I noticed when listening through the card's internally mounted digital-to-analog converters. If you're stuck with a wimpy sound card, the DI/O may be the cheapest improvement to your sound you can make.
I also put the DI/O head-to-head with some respectable 20- and 24-bit converters from Mark of the Unicorn and Digidesign, among others. I listened to material recorded at several studios and at various word lengths and sampling rates. Things started to get fuzzy: sometimes it was all but impossible to tell the difference between the DI/O and the more expensive converters. Considering that the other units cost much more than the DI/O, that is no mean achievement.
OTHER (CHEAP) TRICKS
I also checked out the DI/O as a digital direct-injection (DI) box. Again, lack of an output-level control made things difficult. The unit does not have sufficient gain for some signals — my acoustic guitar's pickup, for example, scarcely registered.
Electric guitar, bass, and keyboards fared better, however. In each case, the tube provided character and, well, warmth. The DI/O won't replace a full-fledged Class A tube DI, but it does a nice job. I particularly liked the added roundness it brought to Fender bass tones.
For my last alternative application, I took a tip from the manual and ran a keyboard through the DI/O's tube path with the unit's analog ins and outs. (That requires a jumper across the S/PDIF input and output.) The unit provided an improvement, albeit subtle. Stick the DI/O between your keyboard and the P.A. and you, too, can say, “Gee, Dad, it's a Wurlitzer!”
Admittedly, I was prepared not to like the DI/O. Call it audio snobbery, but I just didn't believe something this inexpensive could sound so good.
The A.R.T. DI/O does everything it is supposed to do and does it quite well. If you have a sound card with surface-mounted S/PDIF connectors, the DI/O will almost assuredly give you better-sounding audio. What's more, it can improve monitoring from older DAT and other digital recording devices; can double as a credible DI (with digital out); and can warm up analog signals, whether in the studio or live.
I remain underwhelmed with the DI/O's unbalanced analog ins and outs and the lack of a dedicated output-level control. Moreover, I would like to see support for multiple word lengths and word-clock sync, though clearly those amenities would drive up the unit's price. Just the same, the DI/O is one remarkable little box, especially considering its low price.
Mark Nelsonlives and records in southern Oregon. A lifelong acoustic musician, he remembers when “sampling rate” meant how fast you could scarf the shrimp at the wrap party.
tube A/D/A converter
FEATURES2.5EASE OF USE4.0AUDIO QUALITY3.5VALUE4.0
RATING PRODUCTS FROM 1 TO 5
PROS: Compact. Affordable. Easy to use. Surprisingly good-sounding converters. Tube adds warmth to analog or digital signals. Can function as DI box with digital out.
CONS: Unbalanced analog I/O. No output-level control. Does not support word clock. Word length not selectable. Wall-wart power supply with short cord.
SOUNDS GOOD TO ME
CD audio was standardized at 16-bit word length and 44.1 kHz sampling rate because it was believed that those specs adequately reproduced the full spectrum of human hearing. Almost immediately, though, audiophiles began carping, and the search for something better was on.
Although much of CD audio's original bad reputation can be traced to inferior digital-to-analog conversion technology, you can now take advantage of 20- and 24-bit depths for exponentially more resolution and headroom. Likewise, increasing the sampling rate to 48, 88.2, or 96 kHz greatly increases detail, which is especially noticeable in the higher frequencies.
But no matter what the resolution, all digital signals are susceptible to jitter, an unpleasant distortion caused by an unstable clock. That's why professional installations slave every piece of digital gear to a single clock source.
Fortunately, converters have improved dramatically since the early days of CDs. Yet most professionals agree that a high-quality converter at 16-bit, 44.1 kHz frequently sounds much better than a consumer-grade model at 24-bit, 96 kHz. Why? Aside from the quality of the converters, just about every design feature and component inside the box can have an effect on the audio quality. External conditions can also affect audio quality. For example, place the converters near a source of heavy electromagnetic and radio-frequency interference — inside a computer, say — and you're asking for trouble.
Digital Converters (A/D/A)24-bit, 128× oversamplingSampling Rates44.1, 48, 88.2, 96 kHz (selectable)Frequency Response10 Hz-30 kHz (± 0.5 dB)Dynamic Range (20 Hz -20 kHz)>100 dBA (A/D); 105 dBA (D/A)Maximum Input Gain+20 dBMaximum I/O Level+20 dBuAnalog Inputs(2) unbalanced ¼"Analog Outputs(2) unbalanced ¼"Digital I/O(2) S/PDIF on RCA connectorsExternal Sync Range22-100 kHzTube12AX7a dual-triodePower Supplywall wartDimensions5.375" (W) × 2.0" (H) × 5.25" (D)Weight1.5 lb.