Perpetual Technologies P-1A D/D & P-3A D/A processors

At the Consumer Electronics show in Las Vegas in January 1999, Mark Schifter, erstwhile president of Audio Alchemy, was handing out a press release announcing what seemed like a groundbreaking product from his new company, Perpetual Technologies. The product was the P-1A, a digital-to-digital processor that would do resolution enhancement, loudspeaker correction (amplitude and phase), and room correction—all for less than $1k. It sounded too good to be true.

perpp1a.jpgFour months later, at HI-FI '99, PT actually had a working prototype of the P-1A, and gave an impressive demonstration of its speaker-correction function with the Vandersteen 2Ce. I was told that I'd be getting a review sample as soon as production began, just a few weeks later.

Weeks passed, then months, but still no P-1A. I was starting to think that it might be an example of "vaporware," existing only as a concept and a prototype. However, Mark Schifter continually reassured me that the P-1A was very much a real product, and that the delays in production were caused by revisions in the design to incorporate the latest generation of DSP chips, and to build in more power to do the job even better. PT was also about to introduce a companion piece, the P-3A digital processor, which would do the basic D/A conversion. When would these products be ready? Soon, very soon—maybe just a couple of weeks.

Finally, more than 18 months after the initial product announcement, the P-1A and P-3A review samples appeared on my doorstep. The loudspeaker/room-correction software is still not ready for release (the prototype software was 16-bit; it's being rewritten to take advantage of the P-1A's 24-bit capability), but the resolution enhancement is said to be far better than the original prototype's.

Of the two products, the P-3A is the simpler design—as much as the design of any digital processor can be said to be "simple." Developed to complement the P-1A, with matching connections and control format, the P-3A is also a standalone D/A converter. As such, it accepts digital data from 16-bit/44.1kHz to 24-bit/96kHz through coaxial S/PDIF, optical TosLink, AES/EBU, and I2S 5-pin mini-DIN connectors, the latter conforming to the Audio Alchemy standard (the I2S will accept 24-bit/192kHz when it becomes available), and outputs a line-level analog signal through RCA connectors. All inputs are automatically upsampled to 96kHz and, in all modes except I2S Direct, output bit density is increased to up to 24 bits by adding dither in the form of triangular-probability-density function white noise. (This is a much simpler form of redithering than the P-1A's.)

The P-3A is based on the latest-generation Crystal Semiconductor CS4397 chipset and CS8420 input receiver/sample-rate-converter chip, the latter also used by the Bel Canto DAC 1 and MSB Link III digital processors. The analog output stage runs in class-A, built around Burr-Brown OPA134 op-amps. A pushbutton toggles the absolute phase.

The P-1A is a D/D processor, so it must be used in conjunction with a D/A processor to produce an analog output, and the D/A processor should be capable of 24-bit/96kHz performance. The P-1A has S/PDIF, AES/EBU, and I2S inputs (no TosLink); each input can accept a different combination of sampling rate and bit density. The input receiver is a CS8420, same as in the P-3A. However, unlike the P-3A, the P-1A uses only the CS8420's sample-rate conversion (upsampling) function, bypassing its built-in dithering capability. Instead, the digital audio datastream is routed to a high-powered programmable DSP (an Analog Devices SHARC 21065L with 32-bit block floating-point math), and subjected to what Perpetual Technologies refers to as Resolution Enhancement. Designed by Peter Madnick and Keith Allsop and based on proprietary algorithms developed over the years for the Audio Alchemy DTI•Pro and DTI•Pro32, Resolution Enhancement involves dynamically tracking the signal and attempting to derive a better estimate of the 24-bit data than what is possible with a static dithering algorithm.

The Perpetual Technologies website explains how this is done: "The frequency spectrum is split up into multiple bands and the energy of each band is monitored. The algorithm looks for certain psychoacoustic relationships of the frequency bands, which are used as indicators to estimate the original data before it was processed down to 16 bits on CD, for example. This is a single-ended process that needs no special decoding. The algorithm is dynamic, and is most effective on low-level signals, where it is needed most....After Resolution Enhancement, the audio signal is dithered down to the selected output word length. No attempt is made to add in higher harmonics."

Perpetual Technologies' description of the P-1A's design goes on to talk about speaker and room correction, but those are stories for another day. The P-1A has a USB port, which will allow software add-ons and upgrades (including upgrades to the Resolution Enhancement algorithm) through downloads from PT's website.

Connecting a standard digital processor is normally a pretty simple matter; only a complete neophyte would need to read the instructions to figure out what goes where and what buttons must be pushed.

Perpetual Technologies
368 S. McCasling Blvd. #189
Louisville, CO 80027
(303) 543-7200