Show us your capacitors

A brief history of DEQX by Kim Ryrie

We’re often asked how DEQX came to be and, especially, how it came from Down Under…

I remember being amazed, many years ago, that a piece of rolled up paper or cardboard, moving backwards and forwards inside a magnetic field, could reproduce sound with seemingly uncanny accuracy (I suspect that fascination remains strong for most of us trying to create the most realistic sound experience possible). In 1970
I co-founded the ‘electrics’ magazine Electronics Today (ET), that later became ETI (ET International) after we spun off a French and UK edition.

It featured regular equipment reviews where we enlisted the services of a local acoustic analysis firm, Louis Challis and Associates, who had their own anechoic chamber and
all the good gear from Bruel and Kerr.

Louis was a long-term audiophile and this fortuitous relationship was to teach me a lot.

A source of much puzzlement to me was that even outrageously expensive loudspeakers measured hopelessly in comparison to amplifiers, pre-amps and even vinyl audio sources.

This intrigue began nearly 40 years ago and, with loudspeakers having only been around
about 40 years prior, virtually all consumer loudspeakers were passive while amplifiers were expensive.

Occasionally, we reviewed a ‘pro’ monitor speaker that was active (where each driver had its own amplifier, driven by an active crossover) and it was these speakers I liked best.

I felt that the active ‘pro’ sounded less constricted and more dynamic, and basically just more realistic, even though some were far cheaper than the best passive consumer speakers

Consumer-speakers today still use the fundamentally passive and uncompensated regime. And though there have been many improvements, particularly with the recent availability of anechoic measuring and analysis software design tools, the addictive, ‘being there’ audiophile experience remains unknown to most consumers. This is due to the high admission cost of audio nirvana using traditional analogue regimes.

A famous speaker of the 60’s was the Altec ‘Voice of the Theatre’, a 2-way horn design from
the late 50’s. The highs came from a 2” aluminium compression driver via a large moulded aluminium horn section, while the huge bass cabinet used a 15” driver in horn-loaded design.

I managed to get a pair and modified them to be ‘active’, using a simple 18dB/octave active crossover (C/O was about 800Hz) driving two DIY 100W amplifier modules we had earlier published in ETI.

These were installed into a box that contained the HF horn and sat on top of the bass cabinet. Suffice to say, the improvement over Altec’s passive crossover was astounding.

Volume went far louder and clarity was improved dramatically in the midrange and high end. How could such a simple change make such a difference and why weren’t all speakers made that way?

The speaker issue sat on my back burner for a few decades while I was distracted by a fascination with analogue synthesisers following the release of ‘Switched on Bach’ by Walter (now Wendy) Carlos.

With the help of ETI’s R&D team, we had managed to design one of the first analogue synthesisers to follow Bob Moog’s original Moog synthesiser. This was a large DIY project published over twelve issues and thousands of the synths were apparently built subsequently in Australia UK and Europe. By the end of it however, I was frustrated with the inability of analogue synthesisers to deliver organic or natural sounds.

With the goal of developing a new type of synthesiser that could make more natural sounds,
I left ETI and started a new company called ‘Fairlight Instruments’ in my grandmother’s basement on the waterfront of Sydney harbour in 1975 with school friend and electronics wiz, Peter Vogel. That basement actually had a history of innovation because it was where Lawrence Hargraves, a pioneer of aviation build his model wings; Hargraves’ head later came to adorn the Australian $100 note!

I had convinced Peter that having designed the ETI synthesiser, I thought I knew how to make natural sounds synthetically, or at least I convinced him to come on board and work for almost nothing. Since my plan in retrospect would not have worked as intended, it was fortunate we soon ran into a Mr Tony Furse, living on the other side of the harbor. Tony had developed a prototype all-digital music synthesiser for the Canberra School of music. It was never intended to make the natural sounds I was after, but it used FFTs (Fast Furrier Transforms) to synthesise waveforms of any specified harmonic content. This was way ahead of what I had in mind so we struck up a deal to take over Tony’s project.

My original plan had been to use a microprocessor to generate numerous attach, sustain, decay ‘envelopes’ to drive lots of traditional analogue oscillators and filters in the hope of making more natural sounds (later that idea became the basis of the popular Prophet-5 synth). Using Fourier Synthesis however, any sound can theoretically be generated by adding tens to hundreds of sine waves, if you happen to know how each one is supposed to behave. These days that’s easy to analyse, but not then. Nor would we be able to produce enough sine waves affordably in real-time.

As a compromise, we could compute the waveforms off-line and then ‘animate’ them in real-time with considerable dexterity, including various attack, sustain and decay envelopes using specially designed microprocessor controlled hardware, very fast (for the day) memory, and variable pitch control. Unfortunately, although this created very interesting sounds, there was
no cigar as far a ‘natural’ was concerned. So as enthusiastic as we were about the leap in technology that Tony’s creation represented, it couldn’t make the complex ‘natural’ sounds we were after… but we would worry about these minor details later!

As a necessary part of this design, Tony had unwittingly prototyped the world’s first parallel microprocessor design - we were later advised by Motorola, where the shared memory could be written to and read from either processor, which ran out of phase with each other. This avoided the much slower process of ‘interrupts’ that essentially stopped the microprocessor for what seemed like an eternity, causing audible ‘clicks’ when a processor had to wait to receive unsynchronised input data. Instead the hardware and software was designed to allow one processer to manage input and output controls such as playing the keyboard and live controls, while the other managed the audio itself.

Tony had also designed a graphics display controller that allowed a light-pen to select graphical areas or points on the video display (this was before the invention of the mouse and before the first Apple PC). Meanwhile our prototype took about an hour to boot up from nearly a foot diameter reel of punched paper tape fed into a Teletype-33. We had to take care not to tread on the paper tape during this process or it had to be started over again! Some years later we had designed our floppy disc controller, bearing in mind that hard disks were not available except in propriety washing machine size cabinets costing tens of thousands of dollars for tens of megabytes.

Undaunted, and having no idea where this was all going to lead, Peter and I had gone ahead with the licensing deal with Tony even though we had not idea of how we were going to be able to make it play ‘natural’ sounds, which remained the real goal. However, we soon realised that with a relatively simple redesign, since we could play audio waveforms out of memory at different pitches, all we needed to do was to actually ‘sample’ a real sound directly to the waveform memory and play it back at the required pitch with the necessary attach, sustain mechanisms, and decay… all very obvious in retrospect!

We used the microprocessors to record real-time keystrokes from our newly designed velocity-sensitive music keyboard, and so the first real-time computer sequencer was born with a real-time graphical interface. It took a few more years to get the design from twenty 8” x 8” hand-wired boards down to about 12 boards that made the system more affordable and serviceable. List price was about US25,000. First customers were Stevie wonder, Fleetwood Mac, Hans Zimmer, Herbie Hancock, Peter Gabriel (who helped set up a company for our UK distribution) and the list of VIP customers including the like of George Martin grew quickly.

Video - Herbie Hancock demos his Fairlight CMI Series 1 (8-bit) to Quincy Jones:

Video - Me 25 years ago, demonstrating the ‘bottle’ on a CMI Series III (16-bit):

Getting back to loudspeakers though… as new products were developed, Fairlight became
a pioneering user of DSP, reporting many of the first DSP chip bugs back to the manufacturer.
Co-incidentally, since we had opened Fairlight offices through Europe, UK, USA and Japan,
I was fortunate to hear many of the world’s best monitoring speakers and rooms in many legendary studios and mastering rooms. Never did I see a passive speaker except where they were used to hear how the “rest of the world would hear it”, so their use were largely reserved for final mastering.

The Fairlight CMI (Computer Musical Instrument) became a standard for music production through the eighties and at one stage we employed 25 software and hardware engineers. While Fairlight introduced many firsts in the world of digital audio and DSP, my serious interest in loudspeaker design didn’t return until the late 90’s, when it became increasingly obvious
that consumer speakers were still not sounding ‘real’, nor for that matter were most studio monitoring speakers.

This realisation coincided with Fairlight taking on Venture Capital funding to allow further expansion, particularity into video where we had introduced the CVI – or Computer Video Instrument some years earlier. This came with the inevitable style of VC management that suggested it might be time for a change. Co-incidentally, Brian Connolly, co-founder of Lake Technology in Sydney, invited me to help them commercialise their ground breaking 3-D headphone technology based on FIR processing. I agreed to work half time for a year, remaining with Fairlight as a ‘consultant’.

What seemed obvious about the combined frequency and time-domain capabilities of this type of DSP processing, was that it could correct the group-delay and frequency-response errors that remained particularly alarming in affordable loudspeakers. I suggested to Brian and David that Lake should investigate speaker correction, but their focus was the headphone technology, which today is marketed as Dolby Headphones™, and Lake is now a subsidiary of Dolby. Instead, Brian suggested, “why don’t you do it” and introduced me to Paul Glendenning who had done contact software for Lake and who Brian felt was one of the brightest software engineers he had ever met. Given the talent in the rest of Lake’s R&D team was an impressive reference.

Paul and I founded DEQX (initially called Clarity EQ) in 1997 thinking this would all be a piece
of cake and take several years, and a million or so, max. After decades at Fairlight I should have known better! Ten years later, nearly a dozen engineers including a PhD of mathematics, one Professor of electrical engineering and four BE Honours graduates, contributed to nearly half
a million lines of software code, otherwise known as DEQX-HD™… but of course some audiophiles mainly want to know what type of capacitors we use!

Kim Ryrie

CEO – DEQX Pty Ltd, Sydney.