A great deal of energy is being expended to improve the quality of digital audio, and therein lies one of the most vexing intellectual paradoxes of the audio arts. Implicitly this means we are seeking higher musical quality. Yet, how do we objectively judge and quantify such an advance? Can conventional distortion measurements aid us in this task? Can we gain insight into the technical impediments of digital audio with conventional test equipment? How can the audio arts advance without analysis tools of expanded aural capabilities?


As I have asserted so many times over the years, in my writing, the fundamental intellectual paradox of audio engineering is that the essential work that audio circuits perform, which is the encoding, decoding and the amplification of a music signal has never been measured. Let me say that again: the work of audio circuits has never been measured. The metaphor I use is..... a test signal is to music as my tear is to Niagara falls. A steady state test signal or even a multi-dimensional pulse test, at best, is a very distant relative of music. The obvious is true: we do not yet know how to measure the quality of a musical signal, and may never be able to make such a qualitative judgment even if we had higher resolution test equipment because this task is about measuring beauty; an indescribable culturally determined emotion.

What makes audio engineering such a compelling and unique art is that the "stuff of audio" is a completely dematerial substance which has a profound effect on our emotions and body. The experience of music is so complex and multi-dimensional that even the simplest experience of music defies quantification. (Note: rereading Harry Olson’s 1952 book, Music and Engineering, especially Chapter Seven: Properties of Music, will stimulate your imagination)

What is the right strategy for audio engineer and designer when he has dedicated himself to improving the quality of an audio device when even his most sophisticated test equipment is a primitive metaphoric tool for analyzing a circuit’s competence at processing music? Or to put it another way, wouldn’t it be nice to live in a world where .5% total harmonic distortion was a guarantee of higher musical quality than 1% total harmonic distortion, but alas, it isn’t and never will be.


The instinctive answer is that a skilled professional, who is a gifted listener, can benefit by using the world’s most sophisticated test instrument; "the musical mind". There can be no refutation that there is no test equipment that equals our skill at making very subtle discriminating judgments about musical quality. If that weren’t true our one hundred thousand year tradition of making music and musical instruments, which existed long before there was test instruments, would not exist.

We suggest that gifted listeners, i,e, musicians are responsibile for the current drive to improve the quality of digital audio. Clearly there is a discrepancy between the original quality claims for digital audio and what is being experienced.

The strategy for the gifted listener should be obvious: not connect the device under test to a refined audio system and with your "gifted ears" judge its advanced competency at processing a musical signal? Even with the best of intentions this is a daunting task because how can one separate the evaluation of single element from a system where all elements are interacting?

I offer this quote from one of my articles recently published in Positive Feedback Magazine to illustrate the problem of evaluating a single device:

      "All audio systems are chaos systems with each component interacting with every other component in ways that are infinitely complex and constantly changing....resulting in wild resonating interacting electro-mechanical gizmos. You and your room are also elements of this chaos system, you, being the most dominant chaotic element. There is no such thing as a passive component (meaning especially you), and according to Heisenberg, the observer alters the event being observed, and given the well known bio-rhythms of humans and cycles of our glandular system, how can we measure quality in an audio system....especially when a test signal has about as much correlation to music, as I do to Dennis Rodman? In short , most audio systems produce, I believe, at least ten percent gross distortion while playing music.........A simple confirmation of this "normal" gross level of distortion can be gained by first listening to the best Stax electrostatic headphones, and then listening to your system.

Those who have worked in the recording industry over the last two decades are asking...(1) Are the new digital formats significantly better, and will they ever attain the natural harmonic and dynamic quality of ancient analog recording? Or, to ask this question another way.... Why hasn’t the fundamental musical qualities of digital audio, after a decade of refinement, clearly surpassed analog recording that was employed twenty years ago? (2) Even with the improvement in the resolution and quality of new DSP technology is there some other aspect of audio circuitry that is inhibiting the realization of digital’s full potential?

To answer these question an analysis tool of unprecedented resolving ability is needed. To put this another way a "benchmark", or standard needed to be created for judging the qualitative advances in digital signal processing, which could be used by professionals as a tool for judging the true musical advance of their digital developments.


My partner Mark Conese and I felt that the single greatest impediment to advancing the recording arts was the universal "low level resolution" of commercial monitor systems. "How can you take a quality picture when there is thick grease on your lens?", Is the question we ask. So we decided to respond to this audio engineering storm with a bold attempt to create EAMMS©, and Expanded Aural Matrix Monitoring System; an edge of the art monitor system that would have such a high level of resolution that it would be able to reveal, without mercy, the mettle of digital audio. We took on the challenge of creating a monitoring system that would make no concession to commercial or practical reality. We were giving ourselves the task of creating an acoustic analogy of a high resolution astronomic telescope that could resolve the finest quantum of musical information.

For most this reference system will seem "abnormal" in that it employs an all tube, mostly triode signal path, has no inductors in the signal path, the speakers are directly connected to the plates of the output tubes, and is optimized for a custom dual concentric horn monitor speaker with Alnico type magnets. This totally radical yet logical approach was chosen because it created absolutely the simplest, most direct signal path with the greatest potential for natural timbre, aural matrix integrity and authentic dynamic range.

Not once in the design process did we ask ourselves "Is this practical?", instead we only asked... "What is the ultimate expression of musical authenticity". You will note later on that we even had to have a new type of silver foil interconnect, speaker and microphone cables designed to our specifications.

Our goal was to create a system capable of enormous dynamic range and natural tonal integrity so we could determine the true nature of digital’s dilemma. Using EAMMS© we have observed that, in spite of digital audio’s potential for expanded dynamic range, it has not yet achieved its promise. And most importantly, tone, the single most important aspect of music, seems to be significantly distorted by digital audio processing, and we suspect that the villain here is jitter, circuit layout, power supply design and a host of other design considerations which must be addressed to realize the full potential of the new digital recording formats.


THE ACOUSTIC SPACE: This is the most important element of any reference audio system because room acoustics are the foundation of all judgments of quality. As you know the room acoustic have a dramatic effect on the operation of all audio devices because they alter the electrical performance of all speakers and amplifiers. We have a 35x55 feet state of the art room that is used for recording and for housing and demonstrating EAMMS©. This is essential for examining the dynamics and acoustic wavefront of music. Our Steinway concert grand is also in the room, and this room has been used to record some truly fine music (see list enclosed).

SIGNAL INTEGRITY: The golden rule of audio engineering is eternal and it is to maintain signal integrity. This means that the signal should be maintained in its original form as much as possible. A musical signal passing through a capacitor, an inductor, a phase splitter, and IC, transistor, tube, or wire is altered. Our goal was to create a system that did the least damage to the analog musical signal from its source to the terminal at the transducer. We believe that inductance in any form is the enemy of high resolution.

REGULATED POWER SUPPLIES: Our experience reflects the ancient wisdom that a state of the art audio device has regulated power supplies to eliminate the distortion of all unregulated supplies. Years ago we reached the conclusion that "normal" solid state rectification produced a spectra of odd order very high order harmonics that added a harmonic glare in the audio range. Our preference is for active regulation using tubes as pass devices. Where current demands are beyond the limits of tubes regulators with use solid state regulation. Uniquely, all of the amplifiers used in our reference amplifier which we call TRIODZILLA © has all of its voltage rails (for both the input and outputs stages) independently regulated. In our testing it is apparent that unregulated power supplies create a mask and fog in the musical signal which prevent a high degree of resolution. A corollary of this is to be very careful in using electrolytic capacitors which also create a glaze on the music signal, which is why we use ancient oil capacitors which do not ring like electrolytic capacitors.

DIRECT COUPLING: Our goal has always been to eliminate all capacitors, transformers, and inductors in the signal path. While direct coupling is more difficult to achieve there can be no doubt that it eliminates all of the capacitance and inductance that significantly lowers the resolution power of any signal path. This philosophy is expressed in the elimination of all passive speaker crossovers so that there are no capacitors and inductors between the amplifier output stage and the speaker transducers. This obviously means active filtration at the inputs of the amplifiers.

TRIODZILLA © OTL AMPLIFIERS: Because of its size(24 feet), weight (almost 2,000 pounds) and enormous appetite for electricity (don’t get too close) we have affectionately labeled this amplifier TRIODZILLA (see separate article). Perhaps the most esoteric of all tube amplifiers, OTL amplifiers do not use or need output transformers and this permits us to connect the transducer directly the output tube. OTL amplifiers have impressive performance in terms of bandwidth, rise time, natural harmonics and phase shift, but require connection to high impedance/high efficiency speakers. We use custom made balanced differential OTL amplifiers which means each of the four amplifiers requires their own dedicated independent regulated power supplies (four per amp). It should be noted that Dr. Harvey "Gizmo" Rosenberg was the first to commercially perfect OTL amplifiers in the early 1980s and is recognized as one of the world’s leading authorities on OTL amps. This is the first time an OTL amplifier has ever been used in a professional environment, and it is safe to say that those experiencing its unique qualities for the first time will be amazed. (Note: we do not use commercial solid state amplifiers or transformer coupled tube amplifiers because they either do not have a high level of resolution or create a large quantum of unnatural odd order/high order harmonics)

EAMMS©: DUAL CONCENTRIC HORN LOUDSPEAKERS/ALNICO MAGNETS: It is well known that there is only one way to reproduce the authentic dynamics of real music and that is with a horn loudspeaker. Horn monitors have never been used in a professional recording studio environment because of their size, weight, lack of refinement and need for a very large acoustic space. In addition horn loudspeakers must be "tuned" to their amplifiers, which is a complex, time consuming process, and a skill few possess. We started with a Tannoy Westminster Royals horns (sold only in Japan) which has a fourteen foot folded horn and completely disassembled the speaker and rebuilt it to our specification. Each monitor now weighs close to eight hundred pounds. This two way speaker uses a dual concentric 15 inch Alcomax driver, and is completely different type of driver than the one used in the Tannoy pro line that most are familiar with. Both the tweeter and woofer are directly connected to their own dedicated OTL tube amplifier. Our experimentation has confirmed that this is the only type of monitor that approaches the aural integrity of live music. All of our experiments with multi-arrayed loudspeakers has caused us to reject this type of speaker because they can’t accurately produce a whole-istic musical wavefront, which means unrestricted dynamics with natural harmonics. We have never discovered a multidriver professional loudspeaker that did not discombobulate the aural matrix and therefore mask anomolies in the audio electronics chain.

TUBE CROSSOVER/TUBE REGULATED POWER SUPPLIES: The contouring of the speakers frequency response is accomplished with a custom designed two way tube crossover with its own tube regulated power supply.

SILVER FOIL INTERCONNECTS, SPEAKER AND MIC CABLES: After years of research our experiments indicated that the most lucid aural presentation was accomplished by using our custom designed silver foil cables which have extremely low inductance(See article: How Much Less is How Much More?). These cables are manufactured to our specification by Goertz Inc in Bridgeport Ct. We do not use copper cabling because of its relatively harsh harmonic glare. We also only use silver foil microphone cables for all of our direct to two track recording.

AMPEXZILLA: This is our name for our highly modified Ampex two track 1/2 inch tube tape recorder, which has the benefit of new circuitry, tube regulated power supplies and a modern Ampex transport. Our experiments indicates that this machine is the best standard for tonal accuracy and dynamic range. In our experiments recording our Steinway and other musicians we feed microphones directly into the recorder. The reason this "hotrodded" ancient recorder is such an important element in our test site is that we have not yet found a digital tape recorder able to equal its dynamic range and natural tonal presentation. This creates an opportunity for the professional to determine what impediments in circuitry is prevent digital audio from reaching its dynamic potential, which is claimed to be superior to analog.

CONCERT GRAND STEINWAY: We just invested $10,000 in updating our 1943 Steinway Concert Grand to a very high "voice" standard. We record this piano for the ultimate insight into the quality of any device under test. We believe that the concert grand piano is an acid test of natural dynamics and timbrel accuracy. This is absolutely the best instrument for testing digital processors and recorders because "the dynamic envelope" is full revealed.


We recently did a test recording of one of the most advanced new digital formats for the company developing this format and discovered that it was not an advance. In comparison to the previous generation of technology the aural matrix was more compressed and blurred. The test engineer was amazed because the new device, according to his tests, was superior.

In testing and comparing different types of tube microphones we experimented with differents types of power supplies and observed how the tonal and dynamic character of these microphones were dramatically altered by the power supply used.

In comparing different brands of XLR microphone cables we noted a great deal of aural compression compared to our radical silver foil cables.


All of these electronics have been replaced by two, ten pound, Berning 811ZOTL SET amplifiers. We are getting better sound from twenty pounds of David’s new circuit than we got from 2,000 pounds of Triodzilla. The paradox never work your butt off to reach the top of the mountain and as soon as you get there, a new peak appears. I couldn’t appreciate David’s accomplishment unless I had created Triodzilla first.



Back to The Triode Guild©               Back to Meta-Gizmo   

e-mail Dr. Harvey "Gizmo" Rosenberg:

Copyright© 2005™ and Dr. Harvey "Gizmo" Rosenberg      All rights reserved.
All the material contained within the above articles may not be reproduced without his express permission.