P-1A Digital to Digital Processor: US $1099.00
P-3A Digital to Analog Converter: US $799.00
Available Internet Direct
(Reprinted from the Fall 2001 Audio Ideas Guide)
Until recently it seemed that things in the world of digital audio were getting a bit simpler (save for the confusion and complexity associated with the ongoing rollout of DVD-A and SACD). After a long period in which outboard DACs and various other little boxes tasked with eliminating digital jitter were de rigueur for any serious digiphile, the last few years have seen a return to prominence of the single box CD player, and even very good sounding single box DVD/CD players.
Prepare to regress. If you want to get the most from your CD collection in the here and now you’d better not be adverse to some clutter in your audio rack and the cost of a digital cable or two. Little boxes are back, and for good reason.
Some of the brainpower responsible for the little boxes in question here, fittingly enough, is the very same gray matter responsible for the myriad of small, affordable components which flew out the doors of Audio Alchemy before its untimely demise a few years back; Mark Schifter and Peter Madnick to name only two. Some of the technology developed at Audio Alchemy also appears in these products, albeit in a more advanced form (see the discussion of the P-1A below).
At first glance newcomer Perpetual Technologies may appear to be an industry neophyte, but the depth of experience in both design and marketing amassed by the company’s principles is impressive. You may have never heard of Perpetual until recently, but the people involved have been associated with, at one time or another, dozens of prominent high end companies (the Perpetual Website has some bios for your perusal). The business model may be new, in this case being based largely on direct sales over the internet rather than through a dealer network, but the technology on offer, especially in the P1A, represents the culmination of years of work with companies like Audio Alchemy.
The Dynamic Digital Duo
Of the two devices in question here, the P-3A is the simpler product. While it’s designed to complement the P-1A, it can be used as a stand-alone DAC in any number of system configurations with or without the interpolation, upsampling, and other features offered by the P-1A. At its heart is Crystal Semiconductor’s 24 bit CS4397 D to A chip, capable of handling 44.1, 48, 96 and 192 kHz sample rates. Feeding the D to A is Crystal’s CS8420 digital input receiver and sample rate converter chip, which is capable of handling bitrates from 16 to 24 and sample rates between 32 and 96 kHz. In the P-3A this chip is configured to upsample every incoming data stream to 96 kHz for processing by the DAC. With future-proofing in mind, the input receiver can be bypassed via the I2S input (there are also coax (spdif), toslink, and AES/EBU inputs), allowing the P-3A to process a 192 kHz signal by giving it a direct path to the DAC (this is assuming, of course, that a consumer digital player will provide a 192 kHz digital output sometime in the future). Jitter is by no means ignored, the P-3A using a “high-precision, low jitter, quartz oscillator which is used to generate a stable reference frequency clock upon which all conversion processes take place.” Once these conversion processes have taken place, the newly minted analogue signal passes “differentially into a two-stage, class A, high-current audio output stage with five pole split passive/active anti-aliasing filter surrounding the high-performance Burr-Brown OPA134 which drives the [single ended] outputs.”
Controls are simple and minimal, provided by two silver buttons on the front of the unit. At left the input button simply toggles through the inputs. The program button at right allows the user to invert absolute phase and to select I2S direct mode. Two rows of coloured lights adjacent to each button indicate the selected input, phase, and incoming sample rate. Like the P-1A, the P-3A uses an outboard, high current, wall-wart style power supply, described as having “eight individual voltage regulators, and two active buffers with capacitor multipliers” (an upgrade power supply which will drive both the P-1A and the P-3A is available from Monolithic Sound for US $349, and is said to improve performance significantly). Dynamic range and noise floor specs are impressive even with the stock power supply, at 120 dB and -140 dB respectively.
DSP is your friend�
When it comes to the design and functions of the more expensive P-1A, things get a little more complicated. The P-1A is part of a small but growing product category perhaps best described as digital to digital processors. Like the Assemblage D2D-1 (Summer/Fall 2000) it’s an upsampler and jitter reduction box. This statement, however, only hints at all the technology packed into the P-1A. Perpetual calls the P-1A “a digital signal processing platform optimized to host any number of digital audio applications.” DSP is indeed the engine of the P-1, embodied by an Analog Devices 21065 SHARC chip operating at 60 MHz within a 32 bit architecture.
For many two channel purists the acronym DSP may set off a number of alarm bells. Yes, this is the same technology (and in the case of many high-end home theatre receivers, the same chip) which makes garish and ridiculous “DSP Modes” like “stadium”, “jazz club”, and “padded cell” possible (okay, I made the last one up, but you get the idea). But blaming DSP for these vulgar sonic aberrations is like blaming calculus for the hydrogen bomb. A DSP chip like the SHARC is a little like a digital audio tool box, allowing a designer to manipulate audio data in any number of ways, both for good and for evil.
Marketers of audio equipment probably avoid associating their products with the word “computer” as much as possible, but the P-1A bears a greater resemblance to a PC than any run of the mill DAC. Specifically, the computational horsepower of the SHARC wouldn’t be much use without some software to chew on. Just like your PC the P-1 runs software applications, simultaneously or one at a time. This not only makes the P1-A able to do a variety of different things, it also helps make it future friendly, allowing software upgrades (via the USB port) which support new formats and technologies. Perpetual claims, for example, that with its 32 bit floating point architecture the P-1A could support up to eight channels of digital audio.
Perpetual’s first software application for the P-1A, the one which currently ships with all units, is their DSP based upsampling/interpolation algorithm. Like other upsampling boxes, the P-1A will take a 44.1 kHz/16 bit signal and convert it to a 96 kHz/24 bit data stream for processing by a DAC. Most upsamplers stop here, realizing better sound by using the gentler anti-aliasing filters made possible by conversion to analog at 96 kHz, and through jitter reduction. In other words, this is a hardware approach to upsampling which essentially doubles the data rate and, in turn the speed at which the data are processed. Having spent time with the Asemblage D2D-1 and the MSB Link III with upsampling, both examples of this kind of processing, I can attest that this method does have very perceptible benefits, producing uncommonly smooth and detailed sound from CDs (The Assemblage does interpolate bit depth to 24, but not sample rate).
The perception of resolution may be higher via this kind of upsampling, but bumping up the sampling rate and bit depth alone does not create musical information where there was none before. There’s no denying that we’re stuck with the number of samples per second in the original data. Copying each sample twice and processing the signal twice as fast is a little like photocopying a book, inserting each copied page after its original, and then reading every second page. You might have twice as much data, but you’ve still got the same amount of information.
This is where the Perpetual’s software-based approach differs. In creating a 96 kHz/24 bit signal from a 44.1/16 source the P-1A’s interpolation algorithm examines the incoming signal and, instead of just doubling the original data, it fills the larger digital words and extra samples with new data by calculating intermediate values between existing samples. It does this dynamically, based on the musical content by analyzing trends in the incoming data and feeding them into “a proprietary algorithm resulting from [Perpetual’s] eight year research project into psychoacoutics and effective digital audio enhancement.” More simply, using software that appeared in early form in products like Audio Alchemy’s DTI Pro 32, the P-1A attempts to extrapolate what a conventional CD would have sounded like if it had been recorded and/or mastered at 96/24. For a remarkably clear and thorough examination of the differences between upsampling and interpolation have a look a Perpetual’s website.
More arrows in the quiver
This resolution enhancement application, however, is only one of several emerging and planned applications for the P-1A. Over the past several months the company has been busy rolling out SOCS (Speaker Only Correction System), a DSP based algorithm to correct frequency and phase anomalies in loudspeakers. The goal here is to flatten the frequency response and perfect the phase characteristics of loudspeakers by making adjustments in the digital domain, altering the signal to compensate for these problems. Logically enough, the next frontier is room correction, the P-1A compensating in the digital domain, again through a similar custom algorithm, for flaws in the user’s listening room. While SOCS is in its early stages (the small list of supported speakers is growing steadily), full room correction is still some distance away from release. Not having played with either SOCS or the room correction features at the time of writing, I’m going to concentrate on the P-1A in resolution enhanced mode only. In future I hope to follow up on the P-1A’s other intriguing features.
While the P-1A’s software/DSP capabilities are impressive, it’s no slouch on the hardware side either, especially when it comes to attacking jitter. Incoming data are phased locked to a crystal reference source to remove jitter before resolution enhancement. On the way out of the P-1A the data get another frisking for jitter as it’s buffered in “system memory to remove temporal errors.” It’s then “removed from system memory and phased locked to a highly stable master clock source to remove jitter from the output stream.”
In keeping with the themes of flexibility and upgradeability, inputs and outputs are also plentiful, featuring AES/EBU, Toslink, Spdif via coax, and I2S ins and outs. There is also the aforementioned USB port, simplifying the process of future software upgrades. Like the P-3A, the P-1A is housed in a compact, but very solid little chassis, with two chrome buttons controlling all its features and settings. As a result, the P-1A is not as easy to use as most audio equipment without resorting to the manual. Setting the sampling rate, bit density and resolution enhancement requires coordination and patience as you press, hold, toggle, press, set, toggle, hold and press your way through the menus. Once set, mercifully, the P-1A stays that way until you feel the need to fiddle with it again.
The Sound of Enhanced Digital
After having the MSB Link III with upsampling and the Assemblage D2D-1/ DAC 2.6 in my system over the past few months, I was excited by the idea of having another upsampling combo in for a longer period of time, having grown used to a very high standard of digital playback. The Perpetual gear certainly rose to that standard, and with the P-1A in the chain, surpassed it handily.
Let me start with my sonic impressions of the P-3A DAC on its own, since many customers may wish to start with it and add a P-1A, or similar device, later on. While I didn’t have the Assemblage or MSB gear on hand for direct comparison, it was clear that the Perpetual box was doing very similar things to the data coming out of my Rotel RCD 951 transport (via coax spdif digital out). The most obvious thing was an increase in perceived resolution. As discussed above, without the P-1A’s interpolation algorithm, the P-3A is not creating any new data, but the combined effect of greater detail (especially low level information), cleaner transients and an overall smoothness is a subjective increase in resolution. As with the MSB and Assemblage boxes, the P-3A had a stunning effect on massed orchestral strings. I have several CDs from the European classical label BIS which date back to the very early days of the format (circa 1983) which are utterly transformed by upsampling DACs. These recordings of various Sibelius works sound murky, veiled and spatially constricted when played back on the Rotel alone. Played back through the P-3A the sound opened up dramatically, the violins suddenly much more distinct and detailed, and the soundstage much more enveloping. The effect was less pronounced on more modern classical recordings, but the P-3A still boasts excellent detail retrieval and smoothness on any “audiophile approved” disc you can throw at it.
Moving on to pop/rock and jazz recordings the P-3A displayed consistently excellent bass, and very crisp transients across the entire frequency band. Despite the extra detail on offer, the Perpetual DAC resisted digital splashiness very well, taming sibilance and high frequency edge which the Rotel on its own seemed to leave un-buffed. Good DACs tend to do microdynamics very well, with great low level performance and jet black sonic backgrounds, and the P-3A was no exception in this regard. All in all, the P-3A is a very musical and involving piece of gear which I can heartily recommend on its own, especially at the very competitive, internet-direct price of US $799 (Like many internet retailers, Perpetual offers a 30 day money back guarantee on its products).
Connecting the digital dots
If my enthusiasm for the P-3A on its own seems a little muted it’s surely because by the time I listened to it on its own, I had already spent lots of time with the P-1A feeding it from further up the chain. I got the best sound with the P-1A set up to upsample to 96/24, with resolution enhancement (interpolation) running, feeding the P-3A via the supplied I2S cable in “I2S Direct” mode (thus bypassing the P-3A’s input receiver/upsampler). In a truly synergistic fashion, the performance of the two boxes working together was an order of magnitude better than the P-3A on its own. The P-1A turns an impressive and competitive DAC like the P-3A into a giant killer, coaxing out the best CD sound I’ve ever had at home. All the strengths of the P-3A on its own were magnified. Bass became even more tightly controlled and cleanly defined, microdynamics and low level performance improved significantly, and top end detail gushed forth with the kind of ease and silkiness only vinyl and the very best digital gear seem capable of.
If you really want to put a digital front end through its paces, throw some cymbals at it (I can hear the digiphobes now, suggesting that this be done literally). Mediocre digital just can’t handle well recorded (acoustic) cymbals, making them sound hashy, glassy, splashy, or, in the worst cases, like bursts of white noise. If you ever want to demonstrate the problems associated with MP3 compression, for example, try encoding some well recorded jazz with lots of cymbals and listen to the results. Not pretty. The Perpetual combo, however, rendered fantastic cymbals, and sounded absurdly good playing the very low-key jazz tracks from Big Sugar’s first record. Airy, lush, deliciously detailed, I’ve never heard their take on Round’ Midnight (featuring Molly Johnson’s super sultry vocals) sound better. Heading back to the BIS Sibelius CDs, the addition of the P-1A improved things even further, this time soundstaging spectacularly well and making the Rotel’s DAC sound deeply veiled by comparison.
When playing some of my best sounding CDs, the Perpetual combo certainly narrowed the digital/analog gap, conjuring up sound with ease and naturalness, rhythm and attack, and freedom from listener fatigue very uncommon in digital gear. Natural and musical are perhaps the most telling adjectives to describe the P-1A/P-3A combo, as they’re probably the best distillation of my long term impression of their sound. My venerable Rega 3/ Audio Tehcnica/ Rotel RQ 970 BX phono system still has the edge in a number of areas (with good, clean vinyl that is), but the distinction surely isn’t as stark as it used to be. Best of all, you could build a cutting edge digital front end around these pieces, kicking in a few hundred bucks for a transport, for US $2500 or less (one might even argue that with all the jitter reduction going on in both of these units, that the transport becomes increasingly irrelevant, sonically speaking). Not only would you have a digital system that might be difficult to better at twice the price, you’d have a high degree of upgradeability and the option to explore speaker and room correction, should you wish, for nominal additional fees (SOCS coefficients are priced at US $399 for example, with no additional hardware required — no word yet on the costs associated with room correction). Even without taking into account the flexibility and future ambitions of the Perpetual combo, the P-1A/P-3A would make a great addition to the systems of all but the most demanding, and well-heeled CD lovers.