I wouldn't call the improvement rendered by the DTI v2.0 "large"; rather, the difference was marginal but noticeable.
Moving on to the Theta Data Basic transport, I heard a bigger improvement in the Data Basic/GDA-600 pair than I had with the No.31 transport The Data Basic's excellent bass presentation improved in the ways described above when auditioning the No.31, but the biggest difference was in the soundstaging. The music opened up and became more expansive, with a greater sense of air and ease.
The Theta's treble, which was already smooth and clean, took another step forward in liquidity and timbral rightness. Interestingly, the treble was more incisive and detailed with the DTI v2.0, yet it didn't become overly aggressive or etched. I should note that both the Adcom GDA-600 and Theta Data Basic have somewhat laid-back treble presentations, which benefited from the DTI v2.0's ability to increase definition in the treble.
With the Sumo Axiom transport, the DTI v2.0 produced changes similar to those heard with the other transports. The Axiom sounds considerably less good than the Theta Basic, and benefited the most from the v2.0. Even with the DTI v2.0, the Axiom didn't sound nearly as good as the Data Basic without the v2.0. My experience with the Axiom reinforced my previous opinion of jitter-reduction boxes: they're no substitute for a good transport.
Interestingly, the DTI v2.0 didn't eliminate the different transports' sonic characteristics. Instead, the No.31 still sounded like the No.31, the Data Basic was still the Data Basic, and the Axiom was unmistakably the Axiom. The v2.0 must therefore pass the transport's jitter signature to the v2.0's digital output. A perfect re-clocking device would be immune to the transport's jitter, and its own output jitter signature would dominate the sound. This was also suggested by my finding that the input cable to the 1X1 v2 0 made a difference to the sound.
I - SQUARED-S
Next up was Audio Alchemy's DDE v3.0, connected alternately to the DTI v2.0 through coaxial link or the I-SQUARED-S bus. The DDE v3.0's remote-controlled input-switching let me stay in the listening seat and listen back and forth between coaxial and I-SQUARED-S connection.
Starting with the coaxial connection, the DTI v2.0 rendered a much greater improvement to the sound than I heard with the GDA-600, regardless of what transport was used. Where the DTI v2.0 made a clear but marginal improvement to the GDA-600, it significantly improved the DDE v3.0 via the coaxial S/PDIF connection.
But switching to the I-SQUARED-S connection, the improvement was nothing short of dramatic. There was a huge increase in soundstage transparency, depth, space, air, and expansiveness. Low-level detail seemed to come alive; what had been indistinguishable sounds became tangible instruments. There was just more music there with the I-SQUARED-S connection, and I could hear more of what was going on in the arrangement. Small musical details only hinted at without the v2 0 were suddenly vivid and immediate Instruments occupied
more precise locations in the soundstage and had sharper image outlines The v2.0 was like a focus control on the soundstage, snapping everything into a dearly defined and separated whole.
I also heard a newfound sense of air and bloom around instrumental outlines, which made the soundstage more natural and less synthetic. On Robben Ford's vocal on his Robben Ford and the Blue Line (Stretch STD-l 102), I could hear more reverberation around the vocal, and the voice was better focused between the loudspeakers. Further adding to the soundstage performance with the v2.0, the recorded acoustic seemed more expansive. I could hear Eric Marienthal's sax on Mike Garson's The Oxnard Sessions, Volume Two (Reference RR-53CD) light up the hall, and I could actually hear reflections from the walls. Without the v2.0, the soundstage was smaller, the sax was less far back in the hall, and the instrument had less air around it.
While the improvements I've described with the other processors were moderato the DTI v2 .0/D DE v3.0 combination with the I-SQUARED-S connection was a big leap forward in sound quality. In fact, the DDE v3.0 linked to the DTI v2.0 had many of the qualities I hear in the most expensive processors&emdash;particularly resolution of spatial detail. Once you hear the v3.0 with the DTS v2.0 and the 12S connection, you won't want to go back.
SPECTRA
Finally, I tried the DTI v2.0 between the No.31 transport and the Spectral SDR-2000 Pro processor. Despite the SDR-2000 Pro's highly sophisticated re-clocking circuit, its sound is still affected by transports and digital interconnects. Part of this variability is probably due to the Pro's extraordinary resolving power: it acts like a microscope on the input data. With the v2.0 in front of the Spectral, the sound changed, but I couldn't say it was for the better. Contrary to my experiences with the other processors, the Spectral's bass became fatter and less defined with the v2.0. There was also an odd feeling that something wasn't quite right, although I couldn't describe a specific aspect of the sound that gave me this impression. The difference in sound was, however, less significant than with the other processors auditioned with the DTI v2.0.
MEASUREMENTS
Although we have a much better understanding of jitter's effects than we had just a few years ago, we're still a long way from completely comprehending the subject. For example, is the determining factor in sound quality the jitter's spectrum or the overall RMS jitter level? How does jitter of specific frequencies (periodic jitter) affect the sound compared to random (white) jitter? And how much jitter is audible?
Only by a combination of listening and measuring can we hope to answer these questions and attempt to correlate the listening experience with technical performance.
With that goal in mind, I measured the DTI v2.0's effect on the word-clock jitter in two digital processors: the Adcom GDA-600 and Audio Alchemy's DDE v3.0. I measured the DDE v3.0's jitter without the DTI v2.0, then with the DTI v2.0 and coaxial connection, and finally with the DTI v2.0 and the I-SQUARED-S link.
The jitter test instrument was the Meitner LIM Detector fed by a 10x probe attached to the 8x word-clock pin on the respective DACs of each processor. An Audio Precision System One Dual Domain performed the RMS jitter readings (read as an RMS voltage, from which the time deviation is calculated), and also calculated the FFT-derived spectra on the jitter component of the word clock. The measurement bandwidth for the RMS measurements was 400Hz-20kHz.
Fig. 1 is the GDA-600's jitter spectrum when fed a 1kHz, -90dB sinewave from the Sumo Axiom transport with no DTI v2.0 in the signal path. The RMS jitter level was 230 picoseconds. With the DTI v2.0, the RMS jitter level actually increased to 410ps, and the spectrum had more correlated jitter components and higher-amplitude periodic components (fig.2).
These measurements were surprising in light of the unmistakably better sound I heard from the Axiom/GDA-600 with the DTI v2.0. Note, however, that this was the worst-case performance: there was no difference in RMS level or jitter spectrum with the DTI v2.0 when the test signal was all zeros, and only a minor change (240ps Us 270ps) when the test signal was a 1kHz, full-scale sinewave. Although these very low signal levels are more revealing of jitter performance, music has a much higher signal level that may better correspond to the high-level jitter performance. Further, jitter over a bandwidth of DC-40kHz is sonically important, but the measurements taken here were only over a bandwidth of 400Hz-20kHz. There may be things going on outside our measurement window that we can't see.
Conversely, the DDE v3.0's measured jitter performance was significantly better with the DTI v2.0. Fig.3 is the DDE v3.0's jitter spectrum with a 1kHz, full-scale sinewave input from the PS Audio Lambda transport. The RMS jitter level was 240ps. With the DTI v2.0 and coaxial connection to the DDE v3.0, the RMS jitter level dropped to 200ps, and the jitter spectrum became cleaner (fig.4). Note how the strong signal-correlated periodic jitter components at 1kHz, 2kHz, 4kHz, and 5kHz in fig.3 are suppressed by the DTI v2.0 in fig.4.
Now look at the jitter spectrum taken under identical conditions, but with the 12S bus connecting the DTI v2.0 to the DDE v3 0 (fig 5). The jitter spectrum is nearly completely free from signal-correlated jitter components. More over, the RMS jitter level dropped with the 12S bus to just 95ps, down from 240ps with no DTI v2.0, and 200ps with the DTI v2.0 and coaxial connection. This dramatic decrease in RMS jitter and much cleaner spectrum correlate with my listening impressions: the DTI v2.0 improved the DDE v3.0 through the coaxial connection, but the real magic happened with the 12S link.
The same improvements were rendered by the DTI v2.0 with a 1kHz, -9OdBFS sinewave input. Fig.6 is the DDE v3.0's jitter spectrum with this test signal. The RMS jitter level was 365ps. With the DTI v2.0, the jitter level dropped to 190ps, and the spectrum became cleaner (fig.7). With the I-SQUARED-S bus, the jitter dropped even further, to just 95ps. The DDE v3.0's jitter spectrum under these conditions is shown in fig.8. Note the total suppression of the 1kHz component and the fewer periodic jitter components with the I-SQUARED-S connection.
I can't explain the v2.0's measured performance with the GDA-600, but the correlation between listening and measurement is very strong in the case of the DTI v2.0 and the DDE v3.0.
CONCLUSION
The Audio Alchemy DTI v2.0 is a significantly better product than its predecessor, the original DTI. The v2.0 improved the sound of nearly every transport/processor combination I tried with it, but the degree of improvement varied greatly between products. In the case of the modestly priced Sumo Axiom transport and Adcom GDA-600 processor pair, the DTI v2.0 did improve the sound, but not as much as replacing the Axiom with the Theta Data Basic transport in other words, the DTI v2.0 is no substitute for a high-quality transport. The DTI v2.0 really came alive when used with Audio Alchemy's DDE v3.0 processor. Even with the coaxial
connection, the DTl v2.0 rendered a significant increase in soundstage depth, layering, bloom, and air. The bass tightened, and timbres became more liquid and natural. But when connected via the 12S connection, the difference was dramatic, lifting the DDE v3.0 several notches higher in performance. In fact, the DDE v3.0/DTI v2.0 package ($1573 as configured) provided a level of musical performance that rivaled much more expensive digital front ends.
If you own an Audio Alchemy processor with 12S bus input, the DTI v2.0 should be a significant upgrade to your system, and well worth the $599 price. Owners of other processors should audition the DTI v2.0 in their systems before committing to a purchase. Although I thought the DTI v2.0 improved the sound of the various transport/processor combinations I tried, you may get better overall results by investing in a higher quality transport first.
With those cautions, I can enthusiastically recommend the DTI v2.0&emdash;the latest product to reduce jitter's effect on digitally reproduced music.
AFFORDABLE HDCD
Larry Greenhill auditions Enlightened Audio Designs' DSP-1000 Series III and Adcom's GDA-700 and compares them with Audio Alchemy's DTI v2.0 and DDE v3.0 combination
The availability of the Pacific Microsonics High Definition 1 Compatible Digital (HDCDs) PMD100 decoder chip, manufactured by San Jose's VLSI Technology, has brought about a minor revolution m Compact Disc playback. It brings sonic improvements in imaging, sound-staging, and resolution of detail. In the past six months, Stereophile has published a number of reports on the HDCD decoder's operation, what HDCD recordings are available, and the improvements brought by the HDCD chip to specific digital audio processors.
High-end manufacturers are incorporating the $40 HDCD chip in their newest decoders, including the $4695 Sonic Frontiers SFD-2 Mk.II D/A processor (reviewed in Vol.18 No.3, p.127), the $15,950 Mark Levinson No.30.5 (Vol.17 No. 10, p.205; Vol.18 No.3, p.133; VoL18 No.4, p251), and the $8195 Spectral SDR-2000 Professional HDCD D/A Processor (Vol.18 No.5, p.85).
"The difficulty is that conquering the top end of the global hi-fi market is hardly a way to get fabulously wealthy," said one observer ("Pacific Microsonics &emdash;Before the Gold Rush," The Economist, April 15, 1995). For HDCD-encoded CDs to be enjoyed by more listeners, relatively inexpensive processors need to be produced, like the $799 Audio Alchemy DDE v3.0. Similarly, the $1495 Enlightened Audio Designs (EAD) DSP-1000 and the $1000 Adcom GDA-700 are equipped with the PMD 100 HDCD decoder and are considered more "affordable" D/A processors.
The good news is that, by the end of March 1995, six of the 25 manufacturers licensed to use the HDCD process were offering units selling for less than $1500 (Adcom GDA-700, Audio Alchemy DDE v3.0, Electronic Visionary Systems DAC-2, EAD's DSP-1000, Parasound's D/AC-1600 HD, PS Audio's SL Three), and three were offering price-effective upgrades (Counterpoint's D/A-10 HDCD Upgrade, Pink Triangle's DaCapo HDCD Up-grade, and Theta Digital's DS Pro Generation III HDCD Decoding Module).
Another aspect of "getting the technology down market," states Michael Ritter, Pacific Microsonics' President (in The Economist) is to produce more HDCD-encoded CDs. Only a limited number of HDCD CDs had been issued by late summer 1995. Pacific Microsonics' just announced summer 1995 shipments of production versions of the HDCD professional encoder to major producers should increase the availability of encoded discs. Even if the HDCD discs are slow to appear, both RH and Lawrence B. Johnson of the New York limes have found that all CDs sound better on HDCD-based processors. Other commentators find that HDCD encoded discs sound better when played over standard non-HDCD-equipped processors. For that reason, this review was carried out with both types of CDs.
I selected an Adcom GDA-700 for this review because it is the HDCD up-grade of the company's GDA-600, which RH praised (in Vol.17 No.3, p.112) for its "terrific bass, a great sense of pace, and open and spacious soundstage . . . wide dynamic expression and smooth . . . refined treble." Yet he found that the GDA-600 playing HDCD-encoded CDs "wasn't even close" to a price-matched HDCD equipped Audio Alchemy DDE v3.0 (Vol.18 No.7, p.137), which "has firmly established itself as the processor to beat in the under $1000 price category." He hinted that an HDCD-equipped GDA-600 (enter the GDA-700) could help Adcom regain its proper place in the decoder pantheon.
For comparison with a previously reviewed "affordable HDCD decoder," I used an Audio Alchemy DDE v3.0, reviewed by RH in Vol.18 No.7 (p.137), with many of the Audio Alchemy accessories also recommended by RH. (See RH's reviews in Vol.18 No.7 and Vol.18 No.9 for details about the operation of these units.) That meant setting up its remote option, and attaching the company's digital transmission interface DTI v2.0 jitter attenuator via its IS bus. Although the total suggested retail for this full blown Audio Alchemy system is $1722, the DDE v2.0 and DTI v2.0 were being offered a few months back as a "summer special" for $999, bringing the cost to the audiophile down to $1148, well within the price constraints of this review.
LG SUMS UP
All three of these HDCD-equipped decoders have wide dynamic range, which lends pace and impact to digital music. Now it's possible to have large dynamic contrasts without the glare and harshness too often associated with the CD medium. These improvements are available in all three of these under-$2000 decoders &emdash;good news for audiophiles. The EAD DSP-1000 and the Adcom GDA-7000 offer HDCD decoding and excellent sonics, and should join Audio Alchemy's DDE v3.0 and DTI v2.0 on Stereophile's "Recommended Components" list.
Differences were apparent, however. The EAD DSP-1000 was smooth, liquid, and natural, and could never be pushed to be aggressive or hard, even during huge orchestral climaxes. Its ability to recreate non-specific hall ambiance gave it a realism sometimes missing with the other units. The EAD's defeatable attenuation on non HDCD CDs is an asset, considering that the majority of music is available on the non-HDCD format. On the other hand, the point-to-point wiring and use of hot melt for stabilizing critical digital circuits in the DSP-1000 was a concern.
By contrast, the Adcom GDA-700 has the best build quality of the three units. Sonically, it was more dynamic, more transparent, and more adept at retrieving low-level detail, and thus more musically involving than the EAD DSP-1000. The Audio Alchemy system had the best bass response, excellent resolution of orchestral outlines and space, the best rendering of rhythmic pace, but did not have quite the transparency of the Adcom GDA-700. Yet used with the DTI jitter filter, it yielded the clearest vocals and allowed me to follow lyrics lost on the other two decoders. The Audio Alchemy remote is terrific, and adds great value and convenience not available in the other two decoders.
So you pays your money and takes your choice. These three affordable HDCD decoders differed in subtle but meaningful ways in the areas of smoothness, transparency, and convenience. None of them was perfect, but each had its strong points. If you want smoothness, a liquid midrange, and retrieval of nonspecific hall ambiance, the EAD DSP-1000 should be high on your list. On the other hand, if transparency of sonics, retrieval of low-level detail, and bass slam and heft are critical, the Adcom GDA-700 may be for you. The Audio Alchemy shares the bass response, the dynamics, resolution of low-level detail, and excellent soundstaging found in the other two, but is the only one with the remote option.
Because I'm partial to any audio appliance that includes a remote, my personal favorite was the Audio Alchemy system. However, all three of these decoders yield full HDCD-decoding, bringing wide dynamic range and excellent sonics from CDs. For these reasons, I recommend you audition all three of these "affordable" D/A decoders. &emdash;Larry Gneenhill
Robert Harley STEREOPHILE, MARCH 1996
Jitter filter and resolution-enhancement device. Inputs: S/PDIF coaxial on BNC, S/PDIF TosLink optical, I-squared-S bus (ST-type optical optional). Outputs: S/PDIF coaxial on BNC, I-squared-S , AES/EBU, ST Type optical. Output word length: 16.18.20, 22,24 bits (user-selectable). Dimensions: 8.5" W by 2" H by 7" D. Weight 7 Ibs (shipping). Serial number of unit tested: 83317. Price: $1595. Approximate number of dealers: 180. Manufacturer: Audio Alchemy Inc., 31133 Via Colinas # 111, Westlake Village, CA 91362. Tel:(818) 707-8504. Fax:(818) 707-2610.
With the introduction of Audio Alchemy's Digital Transmission Interface (DTI) more than three years ago, the company created an entirely new category of hi-fi product: the jitter filter. The original DTI was a good start, but didn't always improve the sound of the better-quality digital frontends.
The DTI-Pro, released 18 months later, bore little similarity to the rather simple DTI. The Pro added a more sophisticated dual Phase-Locked Loop (PLL) input receiver for much lower jitter, and also performed a new type of Digital Signal Processing (DSP) that Audio Alchemy called "Resolution Enhancement." This processing reportedly increased the resolution of the compact disc's 16-bit digital words, approximating 18- or even 20-bit resolution. It did this by looking at the 16-bit words over a time window, and calculating the additional bits that would have been present had the signal been originally encoded with higher resolution than 16 bits.
In my November 1994 review of the DTI Pro.l I called it a "breakthrough in digital audio reproduction" and "a musical revelation." When inserted in the playback chain between a transport and digital processor, the DTI-Pro significantly improved the sound of my system. Moreover, experimenting with the output word length suggested that the Pro's benefits were indeed rendered by Resolution Enhancement, not just jitter reduction. Shortly after the DTI-Pro was launched, Star Semiconductor, the manufacturer of the Pro's DSP engine, temporarily went out of business. Although Audio Alchemy scoured the globe and hoarded all remaining Star DSP chips, it was clear that the DTI-Pro's days were numbered.
Audio Alchemy took the opportunity to redesign the Pro with a 32-bit DSP chip from Texas Instruments, hence the product's new name, the DTIPro 32. The new chip reportedly has 50% more computing horsepower to run the Resolution Enhancement algorithm. The Pro 32 also includes some operating improvements over the Pro. Fortunately, I still have an original DTI-Pro on hand for comparison with its successor.
DESCRIPTION
The DTI-Pro 32 is housed in Audio Alchemy's standard 8.5" wide by 2" high chassis. From the front panel, the DTI-Pro 32 looks identical to the DTI-Pro. The front panel's 10 LEDs indicate the input selected, power to the 32's three main sections, when the unit is locked to a source, and whether the phase or polarity has been inverted.
The lock indicator actually requires two LEDs, marked "Primary" and "Secondary." The Primary lock shows that the Pro 32's first PLL stage has locked to the source, and the Secondary LED indicates the low-jitter "double-lock" condition. The Pro 32 should double-lock to most CD transports. More on this dual PLL later. The tiny rear panel is consumed by input and output jacks. Input is via coaxial (BNC jack), TosLink optical, or Alchemy's I-SQUARED-S bus. The addition of the I-squared-S bus input is new on the Pro 32, and is included to accept I-SQUARED-S output from Alchemy's forthcoming CD transport. For an additional $179 you can swap the TosLink input for an ST-Type optical jack. The Pro 32's output appears on a BNC jack, AES/EBU connector, ST-Type glass optical jack, and I-SQUARED-S bus. A DC input connector, which accepts a plug from the Pro 32's small external power supply, finishes the rear panel.
An I-SQUARED-S connector looks like an S-Video jack, but has five pins rather than S-Video's four. The five pins carry left and right audio data on one line bit clock, word clock, master clock, and the emphasis flag. By separating the clock and transmitting it independently of the audio data, the unit receiving an I-SQUARED-S formatted signal need not "lock" to the incoming datastream and "recover" a clock. The result is nearly total immunity to interface induced jitter. Indeed, the whole idea of extracting a clock signal from the audio data is a fundamentally flawed approach to transmitting digital audio. Note that when one of Alchemy's processors with I-SQUARED-S input is connected to the DTI Pro 32's 12S outputs the familiar and comforting "lock" LED doesn't illuminate on the processor's front panel. Don't worry; the processor will still decode the I-SQUARED-S data.
Seven user-selectable operating modes optimize the Pro 32's output for your digital/analog converter. These modes set the output word length (16, 18, 20, 22, or 24 bits) and other output conditions. One mode turns off the dither, but performs Resolution Enhancement. The output word length in this no dither mode changes with the program source, with the algorithm deciding how many additional bits it can reliably interpolate. Turning off the DTI-Pro's dither can be a benefit with some digital processors that are already dithered internally. The seventh mode turns off the Resolution Enhancement processing for HDCD playback. HDCD decoding requires that the control code burned in the least significant bit of the CD's 16bit words arrive at the HDCD filter in your digital processor. Any change in the data&emdash;such as that introduced by Resolution Enhancement&emdash;corrupts this control code and prevents HDCD decoding. This is why the Pro 32 offers the HDCD bypass mode. Note that this HDCD bypass option wasn't offered on early versions of the Pro.
Choosing the correct output word length is crucial to getting the best performance from the DTIPro 32. Lets say you have a digital processor with 18-bit DACs, and you set the Pro 32's output word length to 20 or more bits. The DACs will simply truncate (cut off) any bits below 18, introducing noise and distortion. As JA described in his review of the Meridian 518 Mastering Processor in the January '96 Stereophile, truncation also hardens midrange textures and reduces the sense of space.
Conversely, setting the pro 32's output word length to 18 bits if you have a true 20-bit processor prevents you from experiencing the full benefits of Resolution Enhancement.
Just because your digital processor has 20-bit DACs doesn't mean that it will pass 20-bit data from input to output. The NPC 5813 digital filter, for example, truncates the incoming data to 18 bits. The older Yamaha YM3623 input receiver chip will pass only 16-bit data. For comparison, the Crystal Semiconductor CS8412 and Ultra-Analog AES21 input receivers will pass up to 24-bit data, as will the Pacific Microsonics PMD100 HDCD decoder/ filter.
There's yet another trap to be aware of. Digital processors with the Crystal or Ultra-Analog input receivers, the PMD100 filter, and 20-bit DACs still may not pass 20-bit data. Some processors that have been retrofitted with the PMD100 weren't redesigned to pass the PMD100's full resolution to the DACs. Two such examples are the PS Audio UltraLink Two and Enlightened Audio Designs DSP7000 Sonic III. Each of these processors has only a 16-bit data path from the digital filter to the DACs. Consequently, they'll truncate 20-bit input data, even though they have a 24-bit input receiver, 24-bit digital filter, and 20-bit DACs. Any processor designed from the ground up around the PMD100 (as opposed to an existing design retrofitted with the PMD100) should pass at least 20-bit data. The only way of finding out which "20-bit processors truly pass 20-bit data is by the measurements included in reviews. Reports from readers suggest that some processor manufacturers either aren't aware of their products' capabilities, or provide misleading information.
Setting up the DTI-Pro 32 and choosing its output word length requires the owners manual; there's no direct readout of the 32's operating status. You must instead interpret a code provided by the three power-indicating LEDs. For example, 20-bit output is indicated by illumination of only the bottom two LEDs. without the owner's manual (or a great memory), you won't be able to correctly set up the Pro 32 for your system.
To change the output word length, press and hold the "Phase" button until the three power-indicating LEDs switch from power indication to operational status indication. Still holding the Phase button, press the Input button to scroll through the output word length options. Note that you must go through this procedure for playing back HDCD-encoded CDs (if you have an HDCD-based processor), then reset the Pro 32 to its previous settings for conventional CDs. The process is a little clumsy, but workable. Fortunately, the new Pro 32 remembers the input and output settings in nonvolatile memory when powered down, meaning you don't have to reset the unit after it has been turned off. This important feature was lacking in the DTIPro. The Pro 32's output mode selection and display are also improved over those of its predecessor.
The Pro 32 uses the same jitter-reduction technology as that of the original DTI-Pro. A Crystal CS8412-based input receiver locks to the incoming data stream with a Phase-Locked Loop. A second, low-jitter PLL locks to the first PLL's output, then generates a low-jitter master clock for output to your outboard digital processor. The first PLL recovers the data stream, the second attenuates jitter. A Voltage Controlled Crystal Oscillator (VCXO) provides the timing reference. The quality of the VCXO's output largely determines the quality of the output signal. The Pro 32's jitter attenuation cutoff frequency (the frequency above which the circuit attenuates jitter) is reportedly a low 5Hz. For comparison, a Crystal CS8412 without a dual-stage PLL has a JACF of 25kHz. The Pro 32's second PLL has a slightly narrower bandwidth than that of the Pro, a haracteristic that reportedly improves the product's jitter performance.
The recovered signal is then processed with Audio Alchemy's Resolution Enhancement algorithm. A Texas Instruments TMS320C31 Digital Signal Processing (DSP) chip running at 40MHz executes the processing. The Texas Instruments chip is a 32-bit floating-point device, while the Star SPROC DSP is a 24-bit fixed-point processor. Although the Resolution Enhancement algorithm remains unchanged from the original DTI-Pro, the new Pro 32's greater computing horsepower and longer interpolation filters mean that the algorithm can be executed with greater mathematical precision. A Programmable Read-Only Memory (PROM) chip contains the Resolution Enhancement software that tells the DSP chip what to do. This software is easily up-gradeable in the field simply by replacing the socketed PROM.
The DTI-Pro 32's build quality is classic Audio Alchemy: no cosmetic frills or lavish construction, just solid engineering in a spartan package.
LISTENING
I've had the DTI-Pro 32 in my system for a few months, removing it temporarily while auditioning the Avalon/ Spectral/MIT system reviewed in the January Stereophile. I did, however, listen to the DTI-Pro 32 in that system before dismantling it for return to the manufacturers. Most of my experience with the DTI-Pro 32 was in my usual system, consisting of a Sonic Frontiers SFL-2 preamplifier driving a pair of Audio Research VT150 monoblock power amplifiers. Loudspeakers were Genesis II.5s, whose woofers were driven by an integral servo power amplifier. Interconnects in the Sonic Frontiers/ARC/ Genesis system were AudioQuest Lapis Diamond X3, and Wire World Gold Eclipse; loudspeaker cables were short runs of AudioQuest Dragon II. Digital transports with which I tried the DTI-Pro 32 included the Mark Levinson No.31 and Sonic Frontiers SFT-1 transports, and Audio Alchemy DDE v3, Classe DAC-1, and Spectral SDR-2000 Pro processors. Digital interconnects included Alchemy's excellent DST powered coaxial cable, Illuminati's true 75 ohms coaxial interconnect, and Audio-Quest Diamond X3 AES/EBU. Note that the DTI-Pro 32's BNC input and output jacks force you to use BNC-to-RCA adapters unless you have specially terminated digital interconnects.
Power to the system was conditioned by the MIT Z-Center and Destabilizer, with MIT Z-Cord II AC power cords throughout the system (except on the Alchemy units, which have integral AC cords). The MIT power-treatment package has become an essential part of my system.
I'm glad I had a chance to hear the DTI-Pro 32 with the Avalon/Spectral/ MIT package before sending the system back to the manufacturers. This system's soundstaging, image specificity, and amazing ability to resolve spatial cues provided great insight into the DTI-Pro 32's effect on spatial presentation. If you think that a Mark Levinson No.31 transport driving the Spectral SDR-2000 Pro processor couldn't benefit from an Audio Alchemy processing device inserted between them, you'd be wrong &emdash;as I was. The Spectral's sound was taken up another full notch in quality by adding the pro 32. The system had a greater sense of air, space, and soundstage size with the Pro 32. The presentation opened up, with more resolution of the recorded acoustic surrounding instrumental images.
This character was illustrated on the acoustic bass opening of the tune "Leather Cats" from the superb new CD Beyond Words by Oregon (Chesky JDIXO). Glen Moore plays a restored 1715 Klotz bass, recorded with a stereo microphone arrangement in St. Peter's Episcopal Church in New York City. Putting the Pro 32 in the system expanded the size of the recorded acoustic and better resolved reflections from the church's walls. With the DTI-Pro 32 in the chain, I could hear the reflections from the rear wall, which deepened the already spacious soundstage. And this was with what I consider to be the finest digital front-end extant: the mighty Mark Levinson No.31 transport and Spectral SDR-2000 Pro digital processor.
In the Sonic Frontiers/ARC/Genesis system, the Pro 32's effect on the spatial presentation was less pronounced, though still significant. The Pro 32 widened and deepened the soundstage, revealing more space around the presentation. Concurrently, image focus tightened up, and the sense of air around images was better resolved. The result was a greater impression of individual instruments hanging in three-dimensional space. The Pro 32 also made the background sound "blacker," with greater intertransient silence and heightened dynamic contrast.
The Pro 32's improvement varied with the program material and the converter with which it was used. When connected to the DDE v3 via the I-SQUARED-S interface, the Pro 32's effect was more significant. All the benefits I heard with other converters increased in magnitude with the DDE v3, making this already high value processor sound like a much more expensive converter. The jitter-reduction abilities of the I-squared-S interface combined synergistically with the Resolution Enhancement processor for a greater improvement than either benefit alone. When the Pro 32's resolution enhancement was bypassed (meaning the Pro 32 was acting only as a jitter filter), I thought that the Pro 32 slightly degraded the sound of the SDR-2000. The presentation was overall cleaner and more highly resolving without the DTI-Pro 32. I heard a slight loss of low-level detail with the Pro 32, and a trace of hardening in the mids and treble. Engaging the resolution enhancement turned the tables, however, making the Pro 32 an asset to the system's sound. This experience also suggests that the resolution enhancement processing does provide benefits. I preferred the Pro 32 set to Resolution Enhancement, but no dither when using it with the Spectral. With all other converters, the Pro 32's dither was a decided advantage.
No matter what the converter, the Pro 32 improved the bass presentation. The sound took on more weight and solidity, and seemed to have more power and articulation in the lower bass and mid-bass. with acoustic and electric bass, I could hear more dynamics in the strings' attack and decay, more body and weight, and greater pitch articulation. There was no question about the Pro 32's overall improvement in bass performance.
Compared to the original DTI-Pro, the Pro 32 sounded only slightly better. The difference with and without the DTI-Pro was much greater than the marginal improvement the Pro 32 offered over the Pro. I thought, however that the Pro 32 produced smoother midrange textures and a more liquid treble. The Pro 32's bass was also some-what tighter and deeper, but not significantly different from that of the Pro.
MEASUREMENTS
I started by looking at the DTI-Pro 32's effect on low-level signals alternately in bypass mode, and then with the resolution enhancement processing engaged. I performed wideband third-octave spectral analysis, waveform capture, and FFT derived spectral analysis on a variety of signals, but saw no difference in the noise level, noise-floor spectrum, or even the presence of dither. I also tried to measure the DTI-Pro 32's effect by looking at the output of the DDE v.3 processor with and without the DTI-Pro 32 in the signal path. Again, the DDE v3's low-level waveforms, noise-floor, and noise spectrum were identical with and without the DTI-Pro 32. So much for seeing what Resolution Enhancement does. Perhaps Audio Alchemy can provide some measurements in their Manufacturer's Comment, or suggest tests that would reveal the effects of Resolution Enhancement.
I was able, however, to measure the DTI-Pro 32's effect on the word-clock jitter inside a digital processor fed from the Pro 32. Specifically, I measured the DDE v.3's clock jitter, then the clock jitter with the DTI-Pro 32 with a coaxial interface, and finally the DDE v.3's clock jitter when fed from the pro 32's I-SQUARED-S interface. The jitter analyzer was the Meitner LIM Detector operating over a measurement bandwidth of 400Hz-20kHz. Note that the RMS jitter levels shown here are slightly lower than what I measured in the DDE v.3 when I evaluated Audio Alchemy's DTI v2 jitter filter, probably because this was a later production sample of the v3.
Fig.1 is the DDE v3's clock jitter when fed from a PS Audio Lambda transport playing the 1kHz, OdB tone from the CBS Test CD. Note the periodic jitter components at 1kHz and multiples of 1kHz. This signal-related jitter (the clock is jittered at the same frequency as the audio signal being processed) is largely caused by the S/PDIF interface between the Lambda transport and the DDE v.3. The RMS jitter level was a highish 220 picoseconds. The same measurement, but with the DDE v.3 driven by a 1kHz, -90dB sinewave, is shown in fig.2. The signal-correlated jitter components are much higher in level as would be expected with this low-level signal. The RMS jitter level rose to 275ps.
Adding the DTI-Pro 32 to the signal path, and connecting the DDE v3 via a coaxial connection, reduced the RMS jitter levels and slightly cleaned up the spectrum. Fig.3 is the DDE v3's clock jitter when decoding a 1kHz, OdB sinewave with the Pro 32 in the signal path. Comparing fig.3 to fig.1, we can see that the Pro 32 actually seemed to raise the amplitude of the periodic jitter components, although the Pro 32 reduced the RMS jitter level from 220ps to 165ps. Fig.3 (with GI-Pro 32) also shows a reduced number of periodic components; the plot is less dense with spikes.
Fig.4 is the DDE v3 s clock jitter with the DTIPro 32 and coaxial interface, but with a 1kHz, 9OdB sinewave test signal. Compared to fig2 (the same signal, but no DTI-Pro), we can see that the Pro 32 changed the jitter Spectrum, but had little effect on the number or amplitude of periodic jitter components. Again, however, the Pro 32 reduced the RMS jitter amplitude from 275ps (no DTI Pro 32) to 180ps.Not surprising in light of the auditioning, the I-SQUARED-S connection between the Pro 32 and DDE v3 produced a dramatic reduction in the RMS jitter level and a much cleaner jitter spectrum Fig.5 is the DDE v3's clock jitter when the unit was driven by a 1kHz, OdB sinewave via the Pro 32 and I-SQUARED-S connection. The 1kHz interface induced jitter component is completely removed, although we can see some strong periodic components higher in frequency. The RMS jitter level dropped to just 70ps (compared with 220ps with no Pro 32, and 165ps with the Pro 32 and coaxial interface). A more dramatic example of the benefits of I-SQUARED-S is provided by fig.6, the DDE v3's clockjitter spectrum when processing a 1kHz, -9OdB sinewave through the Pro 32 and 12S interface. Note the nearly total absence of signal-correlated jitter components and the clean spectrum (except for the spike at 13.5kHz, also seen in fig.5). The RMS jitter level measured a low 77ps.
These measurements suggest the DTI-Pro 32 is moderately effective in reducing jitter when connected to a processor through its coaxial output, and massively effective when the I-SQUARED-S connection is used. I must caution you against reading too much into these measurements The limited measurement bandwidth (400Hz-20kHz) means we're looking at only a small part of the sonically significant jitter spectrum, which extends from just a few Hz to 40kHz in a multi-bit DAC. There may be things going on outside our measurement bandwidth we can't see.
CONCLUSION
It's hard to imagine that a worldclass digital front-end such as the Mark Levinson No.31 transport and Spectral SDR-2000 Pro digital processor could be improved by a little Audio Alchemy box, but it was. The DTI-Pro 32 tightened and deepened the bass, made the presentation bigger and more expansive, and better revealed space and bloom around instrumental outlines. The DTI-Pro 32 made the presentation sound more like real instruments existing in a real space. I didn't really appreciate what the DTIPro 32 did for my system until I took it out for comparison listening. The urgency to put it back in the system perhaps says more about the Pro 32's benefits than any specific description of what it did to the sound. I simply enjoyed music that much more with the Pro 32.
Moreover, the DTIPro 32 worked its wonders over a wide range of processors and transports. Every combination of transport and processor benefited from the Pro 32, and the improvements were the same regardless of the digital components. The real synergy, however, happened with the DTIPro 32 driving the DDE v3 processor through the 12S interface, making the sound of the v3 competitive with much more expensive processors.
In this age of products that offer marginal musical benefits, the DTI Pro 32 stands out as a significant improvement. You may not have heard what your system can do until you've experienced Audio Alchemy's DTI-Pro 32.
