The DSR system, the technological heart of the ESB 7 Speaker Series, was a revolutionary insight. Entirely developed in the ESB research center, the system has allowed us to create, for the first time, in great detail the image of the stage and its actors, playing a musical program. Just like live.
Renato Giussani began designing speakers for ESB in 1979.
A mechanical engineer, with a passion for loudspeakers from an early age, Renato was also a writer for, and one of the founders of Audio Review magazine. He left ESB in 1984 to found Giussani Research and that company continued speaker development work with ESB. Sadly, Renato passed away on July 25, 2014, but the company lives on under the leadership of his son Marco Giussani.
Renato Giussani led the research team that developed both the 7 Series speaker system and the DSR concept.
The Innovation of the DSR is an achievement of ESB that as yet has not applied by other manufacturers, despite the obvious ability to create a wider and more stable sound stage than any other acoustic systems. The explanation of the failure to imitate (often almost obvious for other technologies) lies in the fact that construction of the speaker is a bit complex (the 7/05 is made up of 90 different pieces of wood) as is calculation of the angular dispersion (lobe) of the different speakers frequencies on which they have to work.
The D.S.R. System
(Article published in the ESB catalog, shown above)
The concept of "Acoustic Spectrum Distribution" of audio frequencies is applied in the DSR in two distinct modes that we can, for simplicity, call Horizontal Distribution and Vertical Distribution.
THE HORIZONTAL DSR
In any stereophonic system the maximum amplitude of the acoustic scene reconstructed is coincident with the distance that separates the speakers. In any listening position the viewer perceives an overall signal consisting of the sum of the value of the direct field (the sound arriving at him directly from the speaker) and the reverberant field (the sound arriving at him from reflections) that form the sound environment. In the pattern of acoustic transmission of the musical signal in the normal home environment, the sound pressure level characteristic of the reverberant field is clearly predominant in the low frequencies while at the high frequencies, the direct sound is predominant.
In another section of the site, Doug Sax will say what is excerpted written below to explain the DSR, not in technical terms but in the words of an audiophile (see full review here):
"The result was stunning, way beyond accurate; a seamless, effortless recreation of the instruments, the room, the musician’s interplay and emotion - I have never heard a speaker re-create the spatiality of these Italian wonders. The ESB 7/06, a restorative experience"
1985 ESB Catalog
When in 1975 the ESB company introduced the concept of vertical alignment with the close arrangement of its midrange and soft dome tweeter, the concept of a united Middle-High was taken as a reference by manufacturers around the world.
THE PROBLEM
If the listener is equidistant from the two speakers and these emit equal signals, the virtual source of what you will hear will be precise and centrally positioned between the two real speakers. But when the listener cannot be in the exact axis of the symmetry of the system then the shift from the ideal location will cause an increase if sound pressure of the direct signal from the near speaker and a decrease from the far speaker. The important thing to note in terms of perceived overall field: Because of the existence of the reverberant field, and the restrictions placed on the listening area in the normal home environment, there will be changes in the levels, and perceived position, of many frequencies, mostly above 1,000/2,000 Hz. Of the level changes that eventually come into play at lower frequencies, they can be of either sign depending on the development of the field of early reflections and standing waves that will be established in that particular environment and, as such, are not to be taken into account for listening position.
So, the effect of listening to an asymmetric position, with speakers traditionally oriented whether standard, unidirectional, or wide dispersion, will' always affected by two types of distortions:
1) Perspective; following the localization of virtual sources slipping towards the closer speaker. This occurs for all sources except for those generated by signals "only left" or "just right", for which the apparent width of the acoustic scene does not vary, but is deformed by compressing one side and expanding the other.
2) Timbre; resulting in the decrease in the level of high frequencies perceived by the far speaker far and the increase of those received from the near speaker.
THE SOLUTION
In a classic experiment Stevens and Newman (1) show that to locate the sound sources in the space of our sound field, the auditory system utilizes the information of both time and intensity. Or, in the presence of two identical sound sources working with the position of apparent center the listener will seem to be more close to one signal if it is stronger than the other, or if it arrives before the other. But the experiment also showed that for frequencies below 1,500 Hz, the arrival time was the overriding factor for localizing the position and for frequencies above 3,000Hz the differences in intensity were more dominant. Let us recall what was said on the fields of direct and reverberant origins. Whereas the acoustic field at low frequencies often suffer from irregularities relevant to reflections and standing waves. While the reverberant field and standing waves are largely a function of the listening environment, the stereo speaker system designed to provide correct variations of intensity in the mid and high frequencies can do much to maintain correct spatial information.
This new approach to DSR (horizontal), to solve the problem of deformation of timbre and perspective for listening positions not equidistant from the two speakers, consists of the following two propositions:
1) Under normal domestic conditions, localization of virtual sources in a correct soundstage depends mainly on differences in intensity 'between the two left and right channels at frequencies above 1,000 / 2,000 Hz;
2) The system must compensate for perspective distortion with a speech function of frequency such as to obtain also a correct timbre over the entire listening area provided.
As has been demonstrated above, the invariance of the localization of virtual sources, can be achieved, even in the presence of lateral movement of the listener, by orienting the axis of maximum emission level of the mid-high frequencies of each speaker to the other extreme of the possible listening positions. This situation and better illustrated at the end, in which there are schematically represented the speakers, the virtual source "V", the listening positions <1> and <1 '> and carriers, the length of which indicates the acoustic pressure in the two positions caused by the direct radiation of each diffuser. The result from research is that the change in level of the medium-high frequencies caused by the displacement from the listening position <1> to the <1'> is compensated by a level variation of opposite sign, as a function of the angle of emission. The orientation of the emission lobes shown gets an effect of the desired type and, by appropriately dimensioning the various parameters, this effect can be made exactly opposite to that caused by a lateral movement, extending the expected listening distance.
The traditional high system dynamics of ESB, always superior, goes perfectly with the new listening needs created by the digital revolution. The 320 mm woofer is one of the best of its type and ensures low distortion. The DSR system distributes the audio spectrum horizontally in an unconventional way by changing the angle of emission. With this technique, the perspective and the timbre of the acoustic scene remain unchanged for any listening position within a wider area.
In the presence of electrical signals having the same spectrum at the input terminals of the two speakers, the listener will then perceive from "L" and "R" a total, sum of the respective direct fields with that reverberated. The total will be the same for any reasonable position listening; i.e. for each, the desired invariance perspective of the acoustic scene is the invariance timbre of each virtual source it contains. The diffusers 7/06 correspond exactly to the demands of orientation (34 degrees) for a listening distance equal to 1.5 times that separating diffusers (central listening in front of the panel of the woofer inclinator of 18 degrees). This condition is also provided by Kates (2) in its table 1 for Y / D1 = 3, and the theoretical solution planned to "high frequency" contemplates a width -3dB lobe of dispersion of 90 degrees. In the case of 7/06 the hypothesis of the formation of a reverberant field having a behavior as a function of frequency depends on the acoustic characteristics of a typical domestic environment and the result of the invariance of timbre over the entire listening area resulted in a lobe dispersion of amplitude varying between 110 to 2,000 Hz degrees and 60 degrees to 12.5 KHz. with a value of 90 degrees to 4,000 Hz. The final result is that, compared to the conventional proposal, these speakers, in addition to having the axis of maximum level suitably oriented, are characterized by a dispersion suitably limited and decreasing continuously with an increase of the frequency, according to a predetermined pattern.
From the choice to distribute the audio spectrum horizontally as a function of the angle of emission we receive the following advantages:
1) The possibility to properly resolve the horizontal structure of the acoustic scene in the various virtual sources from any primary
listening position.
2) Perception of sonic information from each virtual source that is correct at any listening position within the field.
THE VERTICAL DSR
The speakers have a height 7/06 as compared with the other two dimensions, and speakers are located at considerable distances from each other. For example between the center of the woofer and that of the mid-low of 7/06 there are 63 cm.. his is fairly normal for a conventional design. However, the logic of "maximizing" dispersion angle free from alterations leads to choose to arrange the component vertically on the panel of the diffuser at a minimum distance between them, and possibly less than 1/2 of the wavelength of the crossover frequency. In the DSR system such distance is chosen rather than values near a whole wavelength at the crossover frequency (woofer with low average).
The considerations that are the basis of this choice take into account the resolution of the springs by our auditory system as a function of the angle of reception and vertical frequency (Rodgers, 3).
The actual acoustic sources are placed in a space of three dimensions and they themselves have three dimensions. Our auditory system is able to distinguish the various signals it receives from different directions, both horizontally and vertically, and thanks to the different location in space, it can select the better signal to which it wants to "pay attention", separating out ignoring others simultaneously present (eg. like when you talk to a person in the confusion of a crowded room, "Cocktail Party Effect").
With an artificial acoustic source (the speaker), which issues all signals from a single point, this operation on the vector intensity is not an issue. The sound simply is. But, distributing the areas of emission on the vertical dimension of the speaker (not as disturbing as the stereo effect horizontal) so that different signals correspond to different areas of emission requires the auditory system to analyze the various signals using differences if the spectrum and the reception angle to find the sum result. There is no doubt that this listening situation is more realistic than that in which the three dimensions of the real world are reduced at the center of a "pulsating sphere". A distance of the transducers higher than that chosen (by contradiction of some meters) would lead to the difficulty where-in the auditory system must consider each speaker as the acoustic source coherent; as saying that various portions of the spectrum emitted by elements would appear as completely separate. This would prevent the possibility to reconstruct the feeling of unique signal originating from a single extended source, uniquely positioned in the space. This condition must be necessarily respected independently for the issuance of the spectrum of each virtual sound source, regardless of the maximum vertical size and the portion that will be subjectively attributed.
From the choice of distributing in the vertical direction the audio spectrum, as a function of the angle of reception, we receive the following advantages:
1) Possibility to solve the overall programs in the various elementary signals;
2) Contribution to the soundstage of a realistic virtual vertical dimension;
3) size sensing of the zones of emission as a function of the emitted spectrum congruent with the actual situation.
From the choice of distributing the audio spectrum in both horizontal and vertical, finally it derives the advantage of giving the acoustic scene a three dimensional space in which the physical presence of the speakers is less perceptible
The DSR system plan to deploy vertically the audio spectrum according to the angle of reception.
The result is a realistic soundstage with a vertical dimension within which it is more easy to distinguish the different music sources in their correct locations.
Richard Heyser and the ESB 7/06
A test published in the American journal Audio, through a different methodology of measurement, explains more about ESB's DSR project. See the response curve (incorrect) detected by Heyser at 1 mt. of distance and the curve (correct) detected by Giussani at 2.5 mt. of distance. Everything from the website of Renato Giussani, working at that time as ESB chief engineer, and that was the discussion between Heyser and Giussani as described below.
When, in the 1980s , Dr. Richard Heyser (you can learn about Richard C. Heyser in the short box at the bottom of this page) was commissioned to carry out, on behalf of the American high fidelity magazine AUDIO, did the measurements of speakers and wrote all (more or less as I did for Audio Review...), the technical test of ESB loudspeakers 7/06 (published in the November 1983 AUDIO Magazine), were not good. But he did not bother that much of the possible damage the review could do to a small Italian firm and its young designer, who then was me.
He could not understand why the speakers that did not measure well, sounded so much better than other speakers. But the problem was not the speakers. The problem was the measurement standard of the day. They took the measurements using regular (at that time) AUDIO standards without taking into account my recommendations to make the measurement corresponding to that in an anechoic chamber at a distance of at least 2.5 meters, not 1 meter as was standard. As the cabinet is 140 cm high and speakers as a result of the project DSR, are very distant from each other, and a share of the tweeter that would take account of the right height of the ears of a seated listener.
The result is that the measure which was produced and published on AUDIO was very similar to the one on the top, referring to the response at 1 meter distance on the axis of the tweeter of a system (The Audio Speaker 1.3) which has a distance woofer -midrange similar to that in the 7/06 separates the woofer from mid-bass.
Despite the glaring hole in response intersection woofer / mid-bass (caused by the large time lag between the emissions of both components measured by the abnormal position) and the imbalance level woofer / mid-high (which, however, in its replies in environment, rightly, not shown) the overall test was quite flattering, which left me more than a little puzzled... Shortly after the publication of test AUDIO (remember that I am referring to the original magazine, the USA), I had the opportunity to meet Heyser at the Consumer Electronics Show, and I pointed out the error and inconsistency, and I asked him a correction. Among other things, I told him that my measurements had never found similar holes except at close range, while at the normal listening distance all was always good.
The graph here on the top is the one that could have published AUDIO if he deigned to make the measure (albeit in free field) to 1.1 meters from the share of the base case and 2.5 meters from the front panel. As shown, the hole on the lower and middle there, as was to be expected from a crate that, on balance, that sounded good and had understood he said. Although, inexplicably, she did not understand if the exceptional stability of the stereo perspective he had heard was due to the DSR or not. And what else? As it was the only characteristic of 7/06 which could reasonably be traced such a marked difference in this respect, than all conventional systems.
PS: The answers are published simulations of frequency responses, calculated with Cross-PC 4.0, a system for The Audio Speaker 1.3 (at 1 m distance on the axis of the tweeter GRAF.1 and 2.5 m away, 110 cm from the portion of the base case GRAF.2), without introducing the calculus of variations introduced by the presence of the floor.
And here are the conclusions of the test published in Issue. November 1983 AUDIO Magazine (courtesy of Michael Sastra of Audio Classics. Thanks Michael!) and the offending graph
Note that despite any measurement questions, the sound quality and the size of the "sweet spot" listening area were extraordinary
"Piano and human voice, the two most difficult sources, are accurately reproduced, with no change in sound balance over the important frequency range which covers the octave below Middle C through the octave above Middle C. Spectral response in the frequencies above this range suffer from small irregularities, but, on the whole, these frequencies are reproduced well. I tried various combinations of the three equalizer settings, and I believe that the best overall spectral balance is obtained with the reference positions identified by ESB as “Normal.” This system can handle very high sound pressure levels with ease, but I sensed a tendency for certain program material to get “blasty” at high levels; female vocal and horn seemed to be the most bothered by this problem in my opinion. Subsequent technical measurements revealed this to be the equivalent of a “wolftone” about two octaves above Middle C. (This is discussed under “Measurements.”) Stereo imaging is excellent, and the stereo illusion remains intact at all reproduction levels-soft to extremely loud. I cannot say whether ESB’s “Distributed Spectrum” concept was the cause, but the stereo image is solid and not materially influenced by where one sits. I was favorably impressed by the sound of the system when I first heard it at the January 1983 CES, and remained impressed when-given the opportunity to run it through its paces. lf one places it on a firm platform, so it will not readily topple, and throws the grille away, the sound is excellent."
Richard C. Heyser
One-meter on-axis sound pressure level for a constant corresponding to 1 average watt into 4 ohms.
In addition, the response across the range up to 1000 Hz visible on this graph probably suffers the effects of some reflections not well sorted by fenestration temporal choice for the measurement, which may have also influenced the detected level, accentuating the trend uphill curve already caused by the measurement of the response to be too close, which put the microphone a lot closer to the tweeter and mid-high- and mid-bass and bass.
Moral of the story: Somertimes you need to stop measuring and just LISTEN
107th Convention - Jacob K. Javits Center, New York. New York, USA
1999 Sept. 24-27
TECHNICAL COUNCIL “OPEN HOUSE” RECEPTION AND RICHARD C. HEYSER MEMORIAL LECTURE
Saturday, September 25 5:00 PM
This year’s Technical Council Open House Reception introduced a new feature, The Richard C. Heyser Memorial Lecture, which was established May 1999 by the Audio Engineering Society Technical Council and Board of Governors to honor the memory of Richard Heyser. Mr.Heyser was a scientist at the Jet Propulsion Laboratory of the California Institute of Technology, an inventor who was awarded nine patents in the field of audio and communication techniques, including time-delay spectrometry. He also was widely known for his patience and ability to clearly present and communicate new and complex technical ideas. Mr. Heyser generously aided the Society not only through his technical contributions, but also through his service to its growth and organizational development as an AES Governor and the Silver Medal recipient. He died in March 1987 shortly before he was able to assume office as the President of the Society.