Eliminating the Weakest Link

Listening to and evaluating audio products in the CES trade-show environment is usually an utterly useless exercise. But every once in a while, a demonstration will clearly prove an exhibitor's point. PS Audio was able to do this with a convincing introduction to their Power Plant a couple of years back, as was Ray Kimber with his DiAural technology. This year, the "proof of concept in a hotel room" award would likely go to a new Australian upstart, ClarityEQ.

As we reported from the show, the company unveiled their $3500 Model PDC-6.6 (due in May) active DSP correction system for loudspeakers, which has been in development for three years. Using a $350 pair of NHT Super One speakers driven by mass-market consumer gear, the PDC-6.6 noticeably improved the midrange tonality and imaging we were hearing each time it was switched into the circuit.

Sensing that ClarityEQ might be a company to watch, we wanted to learn more, so we contacted Kim Ryrie, managing director for the outfit. He traces the company's origins back to Fairlight Instruments, which he co-founded, and which is credited with creating one of the first computer-based music production work-stations (a favorite with Jan Hammer for producing the soundtrack music for the Miami Vice TV show). Ryrie is still a director with Fairlight, which garnered an Oscar this year for services to the film industry.

Ryrie says that Clarity's goal is to "bring the beauty of high definition audio to lower-cost speakers." Noting that the likelihood of someone's using expensive DSP on an inexpensive pair of speakers is practically nil, Ryrie explains that the demo was intended to show that the loudspeaker is the weakest link in the audio chain and that the "Clarity-Calibrated" processing clearly provides audiophile-quality performance. "The CES demonstration was to compare the original (passive) Super One speakers with a second pair that had been 'Clarity Calibrated'—our certification for speakers with our correction that provides inherent time alignment and a frequency response within one decibel (±0.5dB) over the design frequency range of the speakers.

"For this demonstration we went a step further by actively amplifying the tweeter and woofer, which also allowed the processing to incorporate our 300dB/octave digital crossover. The advantage of going active is that further improvements in clarity can be realized by reducing the energy going to the woofer and tweeter above and below the crossover frequency respectively. This allowed us to lower the crossover frequency from 3kHz to 2kHz (the tweeter's resonance is 1.3kHz), eliminating the woofer's original contributions above its crossover frequency (the woofer in the Super Ones, for example, was contributing significantly up to 10kHz using the passive crossover).

"The DSP is performed on all audio signals going to the amplifiers. The frequency and time-domain correction used is unique for each speaker model, or individually for higher-end speakers. It is computed on-line by our Internet-based server, after measurements are made anechoically, or pseudo-anechoically. The prototype unit demonstrated uses Analog Devices SHARC 32-bit floating point Digital Signal Processors (240 Megaflops). The A/D and D/A are both 24-bit/96kHz devices, although we were only running at 24-bit/48kHz for the demonstration. Although not demonstrated at CES, our standalone processor product will optionally provide room correction, which has different requirements to our far more detailed speaker correction process.

"For example, Clarity-Calibrated speaker correction works to within a sample's resolution across the spectrum, amounting to thousands of correction stages for both phase and frequency-response. When room measurements are made around the listening area, care must be taken to avoid coloring the purity of the original audio content by taking account of complex room reflections. For this reason we deal with room correction to a resolution of only 120 bands, ie, one band per musical semitone for 10 octaves. This allows the user to tailor the room response according to circumstances.

"While Clarity-Calibrated speaker correction must provide a flat frequency response to be faithful to the standard used for all recording and reproduction equipment, a perfectly flat frequency response at the listening position is not necessarily desirable when our psychoacoustic expectations are subconsciously set elsewhere due to the room's natural acoustic. We did not measure or attempt to correct the room for the CES demonstration, as we wished to show the benefits of detailed speaker correction alone."

Ryrie says that he is hoping to bring out a $1000 consumer version of their speaker correction device this year, based on the same grade converters and DSPs as in the PDC-6.6. "It seemed to be an attractive price point from feedback we had at the show. The $1k product I had in mind is a set-and-forget unit, perhaps with several buttons and indicator LEDs allowing one of three programs to be selected, plus a bypass. This unit would be two in and four out, unbalanced. This allows it to be used for single (passive) speaker correction, or as a two-way digital crossover with Clarity-Calibrated correction. I doubt that we could include a calibrated mic at that price, although I am looking. We would hope to sell this unit through the Internet, although a margin would be possible for dealers. I expect that both the high end and the $1k unit would have digital inputs."

In addition to consumer audio buyers, Ryrie says that target markets for his company's products include professional music, post-production and broadcast studios, mobile audio, and most importantly, speaker manufacturers themselves. According to Ryrie, the best results will be possible when speaker designers work in a software/hardware development environment that allows them to try out a variety of different approaches. "We are already in discussions with about five speaker manufacturers. Once they have decided on the features they want for various models, we will customize an OEM module that will be manufactured for them."

Because Clarity's OEM business will focus on active speakers, Ryrie says that they are evaluating three digital amplifier technologies with a goal of integrating these as options with the DSP in each speaker's OEM module. He explains that "this has the advantage that the signal from DSP to speaker remains in the digital domain. Speaker systems with a digital input allow the signal to remain digital from, say, a CD to the individual speaker drivers.

"The other advantage to digital amplifiers is that the power supply is a significant cost in active speakers, and the amps we are looking at are about 90% efficient. This degree of efficiency halves the size, cost, and heat dissipation of the power supply, while heatsinking for the power amps is about 20% of that required for analog equivalents (ie, a 100W digital amp dissipates about 10W while a 100W class-AB amp dissipates 50W).

"The entry level, two-way OEM units planned include a mono analog input (speaker and line level), DSP with RAM/ROM, and 50Wpc digital amp. Our target OEM price in 1000 units for this is about US$200, or $260 for the 100Wpc amp (plus mains transformer). This should translate into about $800 and $1000 respectively at the end user price. However, it replaces the normally required passive crossover and external amplifier. One manufacturer has suggested making our system an optional add-on. This has the advantage that a 'ClarityEQ-READY' speaker could have had its component drivers previously measured, and the necessary ROM can be retrieved at a later time since we would have the serial number on file."