Wednesday, March 30, 2016

Bad Judges or Bad Reference - Part II

I feel I need to further explain why is it that the Spring Break Nationals cars I heard were so bad?

composition image in four parts. First, micro musicians in front of larger microphones on a white stage. Second, micro musicians on top of a car dash serving as their stage. Third, real sized musicians on top of a car dash serving as their stage. Fourth, real sized musicians in front of microphones on a white stage.
Which is the real stage and which has the midget-sized musicians?
While there are a few other reasons, I will take a look at one property of sound reproduction that comes close to my heart; one that makes a big difference in the realism heard from the system.
In a nutshell, I am talking about designing a system with a very good step response. Read my previous posts and you will inevitably come across many mentions of the importance I place on this acoustic measurement. Good systems have a great step response and mediocre ones don't. It is a fact that remains independent of whatever else is done in the frequency/amplitude domain.
But why take my word for it? Why not look at what others comment, when unprompted.
Image of Quad ESL 989 electrostatic speakers inside a living room with many pink accents
Quad ESL 989
The one place where you can reliably get both, a good sound review and good measurements is Stereophile magazine. While it could be argued that The Absolute Sound or other international publications are superior at gauging realistic reproduction, none of them do measurements at the level of Stereophile. To me, the art must be matched by the science. Furthermore, the fact that Stereophile has many years of historic record keeping means that we can come up with some sort of statistical mean relevant to any kind of measurement.
Let's take two speakers from well regarded makers. Neither pair is cheap by any stretch of the imagination. Yet, one resides within the mid-fi world while the other sets foot inside the real high end.
I am referring to the NHT 3.3 and the Quad ESL 989 respectively. Yes for this comparison, I selected speakers that had been around a long time with the hope that most people would have heard of them.
The pair of new NHT 3.3's sold for about $4,000 while a pair of the Quad ESL 989's commanded a price tag of around $8,000. Such a small difference in price, yet worlds apart in performance.
NHT is generally a solid mid-fi speaker maker. Their products have seen strong following and plenty of industry accolades over the years. Still, theirs was never an attempt at the absolute state of the art, no matter what their marketing materials would claim. The fact remained that their path of choice was always one that kept them away from using the best components or having great enclosures. I remember a time when I did my usual ear-to-the-side-of-the-box test and was horrified by the loud ringing the speaker cabinets made. No high end manufacturer would ever allow for such large amount of energy to be dissipated in a poor and unpredictable way within the room. I also recall their engineers acting like I had just farted after I cringed from the disappointing experience.
image of black NHT 3.3 speakers inside a room with many wood accents.
NHT 3.3
But the real test, as always, is the listening session; an event that would further disappoint. While the speakers were bright and dry, what really told me the story about their heritage was how compressed the sound's physical dimensions were. The height of all instruments was ridiculously small. Depth was none and width was definitely inside the speakers; a fact that seemed difficult to conceive after taking in consideration the speaker's narrow cabinet profile, the asymmetric mid and tweeter placement, and their use of a felt strip next to the drivers. In other words, the engineers must have worked extra hard to make a speaker sound that bad considering how many other good things were going for it.
But enough of my opinion. Stereophile's Thomas J. Norton reviewed a par of NHT 3.3's for the magazine in 2005, after the brand had been acquired by Rockford Fosgate but before it was divested in October of that year. Believe me, it's a long story.
While the speaker received a good "B" rating by the magazine, Norton found it necessary to mention the following:

My rendition of a horrible Step Response curve on dark gray background of NHT 3.3 Speakers as publushed by Stereophile Magazine
"The 3.3s' soundstage was tight and focused, accommodating the program material both in the precision of its lateral detailing and in its depth. On some recordings there was noticeably less depth and pinpoint imaging than with soundstage champs like the WATTs/Puppies. But on others—the organ on Dorian's Pictures at an Exhibition, the synthesizer, vocalist, and chorus on Enya's Watermark (Geffen 242332), and the instrumental and vocal interplay on Eric Bibb's and Cindee Peters' Opus 3 compilation (CD 7706/03) jump off the pages of my listening notes—the depth rendition and lateral focus were all I could have hoped for. The only drawback I noted was a limiting of the soundstage to the space between the loudspeakers."



While the speaker played tones correctly, Norton made reference to the same kind of midget-image effect I described when talking about the Spring Break Nationals cars in my previous post. I mentioned that these cars tend to play ping-pong between the right and the left in such a way where the sound never goes past a very sharp boundary. It is also what makes some judges positively comment on how focused images are. Believe me, if instruments were midget sized, all the concentration of energy would always make them seem focused. But this effect is a distortion rather than a positive.
No, these weren't the only problems with the, so called, great cars. So hold back on your desire to jump to conclusions on the unfairness of my disappointment with cars that did some of what a mid-fi speaker do. Believe me, these cars don't come close to even the NHT's. But it is the flaw of compressed physical dimensions that I want to call attention to.
The distortion is so damaging to the reproduction that even Norton saw the need to compare the NHT's poor performance to that from the Wilson Audio's. I guess that great minds think alike.
And what's the reason for such dimensional distortion? Well, a poorly designed crossover network; one where part-count and cost are more important than performance. Take a look above at the impulse response as published by Stereophile in the same article. Each swing up or down is literally destroying the dimensional accuracy of the sound across progressively longer wavelengths.
Now, for the good news. Let's focus our attention on the Quads. The brand has been around for a long time and it is well known for making speakers of all sizes of an exceptionally predictable high quality. In my opinion, they simply never disappoint; even when considering that many electrostatics are notoriously difficult to set up.
In 2002, Larry Greenhill said the following when reviewing the Quads:


My rendition of an almost perfect Step Response curve on dark gray background of Quad ESL 989 Speakers as publushed by Stereophile Magazine
"The ESL989s' imaging was topnotch, conveying a seamless, wall-to-wall soundstage that did not seem to emanate from the speakers themselves. They captured the soundstage depth and width of "Naris," from Patricia Barber's Blue Café (CD, Premonition/Blue Note 5 21810 2). Percussion was open, airy, fast, and transparent. José Carreras in Misa Criolla was startling, his soft tenor in the foreground, a large, muted drum playing deep and to the left of center stage, backed up by the large, distant chorus. The perspective was spacious and eerie, suggesting the desolation and emptiness of a high South American plateau. Suzanne Vega seemed to materialize between the two ESLs, close enough to touch, as she sang "Tom's Diner," from Solitude Standing (CD, A&M CD 5136)."



Interestingly, there was one car that created a sound that could have been described by Mr. Greenhill's comments. But do you think that the show judges were able to recognize it's exceptional nature? No (in case you were looking for an actual response).
There is something pervasive about the soundstage of speaker systems with a triangular-looking step response. They are all big, rounded and well defined; whether from back to front, side to side or top to bottom. They are as close as space reproduction comes to being real. For God;s sake, just close your eyes and pay attention the the most basic of sounds. Don't be surprised if you don't find midget images anywhere.
Illustration of an almost perfect step response over a dark gray backgroundFor your reference, I am also sharing a model-based graph that illustrates the theoretically perfect step response.
So, could it be that the car that I am making reference to as having "conveying a seamless, wall-to-wall soundstage that did not seem to emanate from the speakers themselves" also has a triangular step response? Yes. This is something that any sophisticated professional listener should be able to easily recognize. Clearly there is no hope for the show's judges.
If you have any doubt, it takes an exponentially better engineer to design a system that does a good step response on top of doing everything else right. So, awarding a point or two of separation between the cars that do it and those that don't is simple wrong. A point or two are linear in nature while the performance difference is, again, exponential. This explains the exponential difference in price between the home speakers that achieve it and those that don't.
I remain terribly disappointed that Spring Break sound judges:

  • were not able to identify vehicles with compressed dimensional reproduction,
  • didn't know the exponentially higher value of achieving uncompressed dimensional performance,
  • and were not able to recognize the superior and more natural sound that comes from a well designed system. 

No comments:

Post a Comment