Search this Blog




Monday, December 28, 2009

Arthur Salvatore's wisdom

From superb Arthur Salvatore's site, something I unconditionally subscribe "in toto":

The first and most important priority being that the component must be able to:

Accurately reproduce low-level musical information.

The ultimate effect of this capability is that the component will sound "natural", "musical", "complex", "expressive", "intelligible", "Unpredictable", "alive" and, most of all, "complete", in contrast to sounding "mechanical", "simple", "dead" and "electronic".

This is the rarest and most elusive quality in the quest for accurate musical reproduction. Most importantly, it is my personal experience that it allows and even compels the listener to become more involved with the music and forget about the system.

The second most important sonic priority is:

This is the gut feeling and sense that there is something actually "present" and/or "alive"; As someone else has already written: "There is a 'there' there."

Whenever low-level information is combined with immediacy, the overall effect will be a primal and sustaining sense of "reality". This should be every audiophile's ultimate goal. Why is "completeness" even more important than "immediacy"?

Ultimately, it is still more "involving", in the long-term, to experience music that is complete, complex, natural, unpredictable and with "life", even when listening to it through a veil, rather than the alternative of listening to something that is "right there", when "what is there" is "dead", monotonous and "incomplete". What is the point of that?

The Likely Connection between these two Highest Priorities

I have found that, generally speaking, as a component's ability to allow more low-level information to be heard is enhanced, so is its sense of immediacy. This correlation is logical, since immediacy is compromised by a high noise level. This is also another reason why reproduction of low-level information is the higher priority of the two, and highest overall as well. This connection (or relationship) is not "Absolute", considering the prime example of the Martin Logan CLS, which is both super-immediate and also somewhat dry sounding.

Other audio parameters, such as "speed", "precision" and "cleanness" are somewhat less important and also much easier to attain. The reproduction of the frequency extremes, especially bass, and the recreation of a "soundstage" are also less important to me in reproducing music.

My lowest priority is the ability of a system to play "extremely loud", which I define as more than 105dB. Though in the final analysis; everything has some importance.

My choice of priorities is not just simple self-indulgence on my part. I’ve noticed numerous audiophiles reacting in a similar manner when hearing improvements in these areas. I've been around long enough in this passion to observe many audiophiles "evolve" (defined by me as an irreversible change in direction) over time in their audio priorities.

There have been some common and predictable trends:

1. The first step above "pure junk" is for more "bass and power"; with most people never "growing" any further.

2. Next comes a taste for superior midrange and high frequencies, but without losing the "bass and power". It is here that "Audiophiles are born". However, most audiophiles stop evolving at this point; being reluctant to take their main focus from "bass and power", and consequently only search for further enhancements. The most expensive and complex components are those "enhancements."

Still, the basics and fundamentals of music reproduction have now been accomplished at this stage, which means the system can now be accurately described as "High-Fidelity". This is the single most important milestone on "the journey".

3. The next step is much more difficult; replacing the past focus on "bass and power", and/or "convenience", for midrange naturalness and low-level information. This "area" is where the vast majority of musical information resides, and it's also where analog software and tube electronics excel.

It is not a mysterious coincidence that those audiophiles who end up preferring tube electronics very rarely go back to solid-state. It is also at this stage that audiophiles will make a final preference for analog over digital.

4. Finally, some of those left may decide to go to radical and extreme lengths to maximize the retrieval of low-level information and minimize the system's inherent, unnatural qualities.

This objective can only be achieved by evolving to a "minimalist" philosophy, along with the resulting components and systems. This is a long and extremely difficult process, with the added hazard that even just one "mistake" will have disastrous results to the final sound quality.

Important- For most audiophiles (and readers), my personal priorities will not "match" their priorities. Accordingly, they may prefer the reference components in the "lower" classes, or components not even in any category, to those I have placed in the highest class.

Further- I realize that the above "evolution of priorities" is overly simplistic, so I might write about this in more depth at a later date.

The only audiophiles that do not evolve like the rest of us are:

1. Most audio 'reviewers', who claim to like everything equally (at least in public), never evolve, and rarely, if ever, declare a decisive preference for anything.

2. A number of audiophile "scientists"* who don't believe that there are any real sonic differences (let alone improvements) in components, except speakers, to evolve to and/or with.

*These are the only "scientists" I know, besides anti-evolution Fundamentalists, who don't like to be surprised and also totally lack curiosity.


Does this particular choice of sonic priorities favor some types of music over others?

Yes. These priorities favor the reproduction of Acoustical Music, which is the most subtle and difficult to reproduce, meaning Classical, Jazz and Folk/Ethnic.

I have found that music which is primarily Electronic, studio oriented and/or requiring a continuously loud volume to come across, is far easier to reproduce. There are countless components from the past, and present, which will make electronic music lovers very satisfied. Thus, it should be obvious that for those audiophiles who have a different ranking of priorities than described above, this entire list may be virtually useless to them, at least for the present.



Some readers may now want to have a better understanding of what "low-level information" is and why I feel it so important to music reproduction. Fortunately, there is a simple test and demonstration that anyone can perform on any system:

Just play an excellent acoustical recording, either CD or LP, where the music has both very soft and very loud passages, at a natural volume level.
Then ask yourself this question:

Does the system still sound "just as good" at all volume levels? If it does, that system passes the test!


Any decent system can resolve the pre and post echoes that are audible on some records; mainly those that were recorded in the early days of stereo and/or those with dramatic, dynamic swings directly following relatively soft sections. (One Famous Example-The beginning of the 4th movement of Scheherazade-Reiner/RCA.)

The test involving those "echoes" is extremely simple: the more obvious the echo(es), meaning the more difficult it is to ignore and the more detailed it is, the better that component's (or system's) ability to (generally) reproduce low-level information. While this test is somewhat simplistic, because music is not directly involved, it is an easy first step as a listening exercise and for understanding the general concept.

The Basic Rule is...

The softer in volume a system can play, while retaining ALL of its sonic strengths, the Better that system is in retaining low-level information.

This same test can also be used when auditioning a different component within the same system.

The challenge for this competing component is simple:

A. Does the new component allow the system to sound just as good, or even better, at a lower volume? That is the goal. Or...

B. Does the sound instead start to deteriorate at the same, or even a higher, volume? That is what you don't want.

Always keep in mind:
The relative absence of low-level information is a very serious problem at all volume levels, but it is most easily noticed, and it is most degrading of the music, at softer listening levels.

This same principle applies when the volume is lowered (moderately) with the volume control. The Rule is:

The More you can lower the volume, before the sound of the system begins to "deteriorate", the less of a problem that system has.

However, one must be very careful at this point, because most listening rooms have a very precise, optimized, volume range. Above that range, the music becomes raucous and distorted, while below that range, the music becomes too laid back and starts to sound "dead". So the change in volume must not be large enough to trigger, in either manner, the room's own problems, or the results may be misleading.

What if the system does start to sound increasingly "dead", "dry" and "veiled" as the volume goes down naturally, or with the moderate use of the volume control?

If this occurs, there is a problem somewhere in the system.

ALWAYS keep this RULE in mind:

There Must be a (Serious) problem when any system (or component) has to play "louder than life" to sound "natural" and "alive".

That is the unmistakable sign that musical information has been lost somewhere, and an unnaturally high volume is then being used as a "compensating device" by the listener, usually without even knowing it. This is the primary problem with most of the high-end systems I have heard over the years, and at all price levels.

Analogy- This is the audio equivalent of eating more "junk food" because it lacks basic nutrition.

Or what exactly should I be listening for?

To "steal" some previous thoughts (and words) from My Audio Philosophy:

Low-level musical information encompasses the widest possible array of musical sounds;

1. The harmonics that identify instruments and enables them to sound natural or "musical";
2. The decay of the individual notes and their harmonics;
3. The subtle, instantaneous shifts of dynamics and their intensity and emphasis (also known as micro-dynamics and dynamic shading) enabling musical "expression" to be sensed, heard and felt;
4. The sense of ambience and space, allowing the listener to both hear and be "there";
5. The complexity and separation, or absence of homogenization, of all of the above, reducing "boredom" and "listener fatigue";
6. and the sense of both continuity and a continual and consistent presence, which has also been described by others as "continuousness".

It is also indispensable that all this musical information be retrieved accurately;

Both in relative level and in phase.

This allows the music to sound "natural" and appear "intelligible". This is especially relevant with speakers, which have the most problems of any component with the accurate reproduction of both musical timbres (relative level) and with relative timing (phase).

This is this musical information that, more than anything else, allows the listener to believe that the music he/she is hearing, and experiencing, is a unique and human event, rather than one that is electronic, mechanical and ultimately contrived, like if "the pod people" from the film "Invasion of the Body Snatchers" had taken over the world.

Low-level information turns "notes", "sounds" and even "noise" into MUSIC.

Finally, if I was to make an analogy with fine food and drink, I would describe low-level musical information as the equivalent of the "Aftertaste". Its very existence and character separate the different qualities of food and drink from each other. This is what also happens with music when its reproduction is "complete".



The term I formerly used, "noise-floor", has been in use in the audio world for a number of years now, and while I do recognize the vital importance of the underlying concept, I also felt that the choice of words ("noise-floor") was very unfortunate, because it has proved to be a confusing term for many audiophiles, even veterans.

A much better term, in my opinion, is: "Sound-Floor"

I feel this way because:

The expression "noise-floor" has Nothing to do with traditional "noise".

"Traditional" meaning the measurable noises (hum, thermal hiss, mechanical buzzes etc.) that emanate from all active, electronic components; such as amplifiers, preamplifiers, motors and even CD players. While loudspeakers, which are a passive component, have no "traditional noise".

In contrast, all audio components, passive AND active, including loudspeakers, have a "sound-floor". And that is not all...

So does the actual software; records, tapes and CDs, and in this instance, I am, once again, not referring to their background "hiss".

For a better or different perspective, the reader must realize that...

The key word in this expression is not "sound" or "noise", but "Floor". The word "floor", in this instance, is an indication of the "lower limit" of a particular capability of that component.

The "sound floor" (or "noise floor") can be best described as:

The "lower limit" of an audio component's capability to reproduce (or pass) softer and softer sounds.

Put in another manner, the "sound-floor" can be described as:

The softest sound that can be heard or sensed through that component (or system).

(And that by definition means)

Any sound that is lower (or softer) than the "sound floor" Must be Inaudible.
Analogy- It is the audio component's (or system's) direct equivalent of the listener's ability to sense or hear "soft sounds".

A component with a high "sound-floor" will obscure and mask a large amount of audible sound (music), while a component with a very low "sound-floor" will reveal virtually everything about the sound (music). Unfortunately, "the weakest-link-in-the-chain" rule applies in this case. This means that if even one component has a high sound-floor, so must the entire system.

This is the reason why systems that have a high sound-floor will be played at a higher volume, usually without conscious awareness, in an attempt to hear what is missing.

Why do certain components have a higher or lower "sound-floor"? That is not entirely known. What is known is that components that use overly complex circuits and layouts, longer signal lengths and poor quality passive components (wire, resistors, inductors, capacitors, speaker drivers and vibrating cabinets etc.) generally have a higher "sound floor" than those components which avoid their use. Poor execution will also compromise the "sound floor".

Also, everything being equal, tube preamplifiers and power amplifiers will almost always have a lower "sound floor" than their transistor equivalents. Ironically, this is true even though their actual "noise" (hiss, hum) will usually be measurably higher. Speculation about this phenomena has focused on the greater simplicity of most tube circuits (especially single-ended-triode designs) and the fact that the actual amplification occurs in a vacuum, not silicon or some other material.

So to summarize, there are four requirements in order for a component to have an exceptionally low "sound floor":
1. A simple, though highly competent, design
2. The use of the best quality parts, both active and passive, within the component
3. The finest execution of the above, both in build quality and in close attention to (small) details
4. The shortest signal length(s) possible

If any of the four requirements are compromised, the "sound floor" will rise accordingly, and the recorded sounds (and the music) will be permanently lost. Sadly, only a few rare and outstanding components meet all the requirements. Searching for them, and hearing them, one way or the other, should be high on the list of a serious audiophile's priorities.

There is also a relationship between a system's sound-floor and listening FATIGUE.

When a system has a high sound-floor, meaning more of the musical information is missing, the listener will then (automatically) attempt to fill in "the missing parts" with his brain.

This continual effort, usually unconscious, will eventually cause "listening fatigue". The existence, and even the degree, of the fatigue is dependent on the previous experiences, and expectations, of the listener.

For example, digital recordings and sources are known to have a higher sound-floor than good analog. This is the reason why some listeners, who are used to analog, may experience fatigue with digital, despite digital's other sonic advantages over analog.

While other listeners, who are used to primarily digital recordings, do not appear to suffer the same fatigue.

Finally- The first, unavoidable and ultimate rule in high-end audio is:

You can NEVER have it all, so everyone must eventually make choices, and that means choosing priorities.

Anyone who says otherwise is either very ignorant or lying."

A superb essay, indeed... many thanks to Arthur Salvatore for the above as I consider it as an important, better, a major achievement in defining the relationship between audio & music.


Thanks also to Denis Guzzo, photographer extraordinaire, for the magnifying lens portrait...

No comments: