Great Misconceptions
by Steve Nugent

Steve Nugent is President of Empirical Audio, a holder of numerous patents in electronics, and a gifted cable designer/modifier of fine audio equipment. His website will be found at www.empiricalaudio.com. Positive Feedback Online is a forum for the views and ideas of all who write well and have something constructive to say about the possibilities of the audio arts; designers, manufacturers and distributors are quite welcome here, provided their articles are non-partisan in nature, and avoid the self-promotional. Interested parties in the industry are invited to submit candidate articles/letters to [email protected] for consideration.

During my recent exhibition at T.H.E. Show 2004 I noticed some behaviors that I would like to comment on. My company is a regular exhibitor at CES, usually at T.H.E. Show. We always have an active exhibit, which means that we play music through one or more systems. This year we featured a system comprised of 100% modified (hot-rodded) components, since we are in the modification business as well as the cable business. 

As the show progressed, I noticed that attendees had certain common reactions to the quality of the music playback in each exhibit. The attendee generally attributed good or bad sound quality to the loudspeakers being used. As you can imagine, when an exhibitor was not specifically demonstrating loudspeakers this tendency was frustrating. I was one such frustrated person. I'll be the first to admit that loudspeakers are one of the more difficult things to design well, as well as being the most visible component. However, having a component modification/cable business has taught me several things:  

1. Improvements can be made to virtually ALL electronic components (some having an even more pronounced impact than changing speakers), mostly due to deficiencies in power delivery;

2. many components, particularly amplifiers and speakers can be synergistic or non-synergistic due to electrical interactions, and;

3. even very old components and loudspeakers which were solid designs in their time can be made to sound incredibly good, even by today's standards. A case in point is the modified 16-year old KEF 104/2's that we demonstrated at T.H.E. Show this year (a 20 year-old design). Dealers that carry the KEF line and heard them in our suite commented that the modified 104/2's sounded even better than the new KEF reference designs. We have no plans to offer this mod to our customers; we just liked the sound of these speakers unmodded, and even better after the mods.  

The message that we had difficulty delivering to many attendees was that it was EVERYTHING in our system that enabled our sound, not just the speakers. The entire system had to be modified in order to achieve the sound that we demonstrated there. 

I conclude from this common "speaker attribution effect" that the CES/T.H.E. Show attendees are evidently convinced that the electronics must not make any difference, or, at most, but a small difference. Only the speaker makes a huge difference. The fact is that any component, from the transport to the cables in the system can make a huge difference. Many audiophiles who have modified their components have experienced this. After all, audio reproduction is a system comprised of electronics and connections, as well as speakers. And the electronics are anything but perfect. The electronics or cables can muck-up the sound quite easily, and make an otherwise good speaker sound bad. 

Good electronics design makes for good sound, and I'm not just talking about what chips are used in this or that DAC or preamp, or even the circuit design. This is the other attribution error that audiophiles often make: that a component will sound good if a particular chip is incorporated in that design. Parts selection is certainly important; however layout, power delivery and circuit topology (where things are located and how they are wired together) are just as critical. In fact, I have personally dissected many designs that incorporate superb electronic devices, only to find them wired together poorly using circuit-boards, or poor wiring harnesses, or circuit boards laid out with poor power delivery topology—all resulting in disappointing sound.  

The design flaws mentioned above aggravate the transient power delivery to the circuits. Transient power delivery is the ability of the power distribution system in the component to provide short-term high-current bursts when the devices demand them. By devices, I mean transistors, op-amps and other integrated electronics. Transient power delivery deficiency can take an otherwise great circuit design and shoot it in the foot. When these short-term current demands are not met, voltage sags occur on the integrated circuit dies, in the packaging, or on the circuit boards. These voltage sags prevent the music from being reproduced accurately by the devices. The effect can range from a compressed or un-dynamic sound to a "dark" sound. Once the things that limit the transient response are eliminated in the design, even poor sounding components can sound extraordinarily good. Case in point was the modified Adcom GFA-585 that we demonstrated at T.H.E. Show this year. Most folks commented that it did not sound anything like an Adcom, and one person even commented that it was the BEST amp that they had ever heard.  

The specifications for a typical component generally include frequency response, distortions, signal to noise ratio and the like. The difficulty is that none of these specifications address the components' ability to deliver transient power. All of these specs are measured "steady-state," with an unchanging input stimulus. In order to measure transient response, one must apply a burst waveform, such as a "rim-shot" in a piece of music, a voltage step, or an impulse, and then develop some standard quantitative metric for accuracy in recreating this transient waveform. The closest metric that we currently have to a transient measurement is the qualitative 1kHz and 10kHz square-wave plots such as those published in Stereophile equipment reviews. 

It is good transient power delivery that results in a live, dynamic sound from any component. We have found that it is the transient dynamic response that brings the music to life and makes it not only un-fatiguing, but a joy to listen to. The difficulty is that very few circuit designers (even the world-renowned ones) are well-versed in the other disciplines, including power delivery, grounding and shielding, transmission-lines, electromagnetic fields and cable design (metallurgy and optimum parametrics for each environment). Many of these designers have great reputations as well, and it would be a slap in the face for some of them to seek help in these areas. As a result, most stock components tend to be compressed sounding and often create their own sibilance. It is a shame, but on the other hand gives modders a large aftermarket for improving components. 

The exhibitors are also at fault here. It seems from the attendees' feedback and my own observations that the exhibitors are hard-pressed to assemble good-sounding systems. And it's not just the room acoustics, trust me. The feedback from attendees (not press) on the web about the 2004 CES show and T.H.E. Show had this as a recurring theme. This does not bode well for fine audio as a business; if the exhibitors cannot assemble a good-sounding system, how can the consumer be expected to do so?  

In my opinion, many of the exhibits that did not sound good had poor sound quality due to a mismatching of components, such as tube amplifiers driving low-impedance speakers, or low-efficiency speakers being driven by tube amplifiers with relatively high output impedance (low damping factors). Those few exhibitors that did pay attention to synergy sounded much better than the rest. I will not mention names, but most of the poor sounding systems consisted of tubed amplification driving large low-impedance woofers which did not control them very well. Again, transient response is the culprit. The power required by the woofers cannot be supplied when they need it due to the power delivery topology in the tube amplifiers, or a mismatch of the output impedance and the speaker impedance. 

In addition, many of the front-ends were not good sounding and created an unfocused image, which in turn reflected badly on the speakers because of the attendees' "speaker attribution effect." Many exhibitors cannot afford extremely expensive front-ends to show off their speakers, so they often end-up with mediocre sound. What many have not discovered yet is that modded components can often address this deficiency, and at reasonable cost. In fact, the cost can be so low that the exhibitor can have two sets of electronics, with one for backup in case something fails at show time. (I'm not aware of any other exhibitor that brings duplicate electronics to the show.) Also, many of the exhibitors do not seem to be cable believers, so they use inexpensive cable, with mediocre sound resulting.  

I wish I knew how to educate the exhibitors so that they could deliver better sound at the shows, but I do not. The best I can do now is to write this article pointing out some of the roadblocks to good-sounding systems, and some options. I would also hope that some attendees that read this will not automatically attribute good sound at their next audio show just to a "good-sounding" speaker, but will understand that this is a hallmark of a good-sounding system. Also, some component designers should pay more attention to design that properly addresses transient power response, and perhaps leave their egos outside long enough to get help in the areas where their expertise is weak. The improvements that they would achieve in sound quality would be well worth the inconvenience. 

In conclusion, I hope I have made both exhibitors and attendees aware of the deficiencies that exist in most electronic components, and that audio systems are just that: systems comprised of electronics and cabling, as well as speakers. In order to achieve a good sound, one must plan the system carefully and take into account the characteristics of every component and cable.