AR-3a High-Frequency Spectral Balance

[tysontom]

In a recent thread, “Ar-3a Tweeter – Carl’s and Roy’s Observations – Potential solution??” Jerry “onplane” comments on the AR-3a.

“the other two drivers suffered from the same syndrome, this might not be a terrible problem. The mid and woofer, however, don't seem to have suffered this same fate. This leaves the high frequencies coming from the Ar-3a somewhat "muted".

In that same thread genek mentioned: So AR speakers have no settings that "boost" mid or treble; "full" is what you use with an amp and recordings that don't have "excessive" highs …

Genek is correct! There is no “boost” setting and for a very good reason. Back when the AR-3a’s were designed the primary source of music was vinyl records and the vast majority of playback devices had ceramic phono cartridges, which were woefully inadequate in accurately reproducing high frequencies. So in order to sell their recordings, the record companies artificially “enhanced” high frequencies to compensate for this deficiency.”

No, this is an incorrect assumption, and it has nothing to do with the balance of high frequencies in AR speakers. When the AR-3a was introduced in 1967, ceramic-phono cartridges were practically non-existent and were only used in the cheapest one-piece music-box sets, and rarely in high-fidelity sound systems. Nearly every quality system had either a Shure (or Stanton, etc.) magnetic cartridge, and some even used moving-coil-type cartridges. Ceramic cartridges became nearly obsolete nearly ten years earlier by the time quality stereo became a reality. The real change to improved sound from the old ceramic versions was the introduction of GE’s variable reluctance cartridge; then along came Shure Brothers and others with their superior moving-magnet cartridges.

The engineering philosophy of AR engineers was without exception to design speakers with the highest-possible sonic accuracy. Remember AR’s famous words in one of the brochures: “The accurate reproduction of sound is one of man’s most benevolent gifts to himself.” AR felt that accuracy required flat acoustic-power response in the listening environment that replicated the sound field one would experience in a concert hall or auditorium, etc., and to get this degree of flat acoustic power, the speaker would have to exhibit extremely wide dispersion at the highest frequencies. Unfortunately, there was a bit of a trade-off between efficiency and wide dispersion in the design of tweeters capable of providing the extended power response required. Therefore, at the time, the ¾-inch AR-3a hard-dome tweeter had somewhat lower sensitivity than either the midrange or woofer, and this resulted in what many people felt was a sense of “reticence” or a “laid-back character.” AR considered padding down the woofer and midrange to match the tweeter’s sensitivity, but this was an unattractive proposition considering the already low efficiency of the AR-3a.

What AR found was really happening, however, was that to obtain sound in the living room that approaches the original spectral balance of live sound in a concert-hall seat, it is actually necessary to turn *down* the level controls of both the mid-range and tweeter units in an AR-3a! The reason is simply this: microphones for both live broadcasts and recording sessions are invariably set up in the near field of the sound source, while concert-goers are in the reverberant field. In concert halls the reverberant field has a far more drastic high-frequency roll off (relative to the bass) than is true of living rooms. This was clearly shown in measurements made by Bolt, Beranek & Newman in their (unpublished but reported) tests of nine concert halls, and these halls were measured when empty! Thus what is put into the speaker systems is a near-field spectrum, usually made even “hotter” because the instruments are aimed at the microphones during the recording process. To produce the *same* spectral balance at the ears of listeners in both concert halls and living rooms, the reproducing system must compensate for the difference between high-frequency roll offs in concert hall and living rooms. For those circumstances requiring truly flat acoustic energy input to a room, such as the playback of live recordings made at home, etc., then the treble tone control might have to be turned to maximum or the use of additional tone control input.

--Tom Tyson

[Steve F]

The whole subject of the original ARs’ spectral balance is fascinating. Aside from their trend-setting bass response, the aspect of their performance that stirs the greatest discussion is their spectral balance: the entire “Not enough highs” issue.

The reason for their seemingly laid-back high end is well documented, and succinctly re-stated by Tom in the above post. AR strove for accuracy, wide dispersion, smooth power response in the reverberant field, as well as low distortion. In order to achieve their dispersion goals, it was necessary to invent a miniature tweeter— a ¾” dome—to deliver the widest possible dispersion.

Unfortunately, a ¾” dome driver has only a ¾” voice coil, and since ferro-fluid cooling wouldn’t be introduced to the speaker industry until 1976 (another AR invention), the original incarnation of AR’s ¾” dome tweeter had very restricted power-handling and low efficiency. This low power handling meant that the voltage drive through the crossover had to be restricted (“padded down”), or else the tweeter would be too prone to thermal failure (“burn out”).

So there was the 3a, with great bass, very smooth FR through the bass-to-mids, very wide dispersion at the highest frequencies, but a somewhat downward-sloping response towards the high end.

AR’s own data sheets showed that in order for the 3a to achieve flat response it was necessary to advance the Mid-High controls to ‘MAX’ and also to turn the treble control on the AR amp up to the 1:00 or 2:00 o’clock position.

Yet AR maintained that the most realistic sound was most often obtained with the controls at their ‘Norm’ positions, not ‘Max,’ as the Norm positions most closely replicated the ‘concert hall slope’ of the high frequencies as heard by listeners about 1/3 of the way back in the hall.

Critics of AR speakers are quick to point out two things:

1. The spectral balance heard by listeners in a symphonic concert hall was entirely different than that heard by a listener sitting 15 feet from the stage listening to a jazz quintet at the Village Vanguard in NYC or The Jazz Workshop in Boston. In those venues, the highs were much more prominent than 1/3 of the way back in Symphony Hall or Carnegie Hall. AR’s assumption that the “concert hall slope” was the correct and only target for acoustic replication in the home was open to question, to say the least.

2. When AR was able to drive their tweeters harder with the advent (no pun intended) of ferro-fluid in the 1975-6 intro of the ADD line, they did so, and thus completely changed their speakers’ sonic balance. A sharp prosecuting attorney would ask, rhetorically, “If the 3a’s spectral balance was correct, why did the 11—essentially the identical speaker—have such a stronger high end?

In several places on the 1971-1972 technical data sheets (the ones that show the responses of the individual drivers), AR shows the response of the woofer, midrange, and tweeter on the same graph. The AR-5, for example, has this graph. It clearly shows that the on-axis response of each succeeding driver is lower in level than the one before it. The woofer is highest; the mid is 2-3 dB down from that; the tweeter is 2-3 dB down from the mid. So if you read AR’s own lit very closely, and look “between the lines”, so to speak, you’ll find that they ‘admitted’ that their Classic speakers had an attenuated high frequency response.

It’s obvious that the Classic ARs sacrificed a ‘flat’ on-axis HF response in favor of wide dispersion and smooth (albeit somewhat downward-sloping) far-field power response. In terms of listenability and freedom of listener placement, these were very good design trade-offs. In terms of retail success and advantage in quick showroom A-B demo comparisons, they were disastrous design decisions.

Steve F.

[Carlspeak]

Thank you Tom and Steve for pointing out the original design goals of AR engineers in achieving the spectral balance they sought to create with the great innovations in tweeter and speaker designs.

It's also interesting to note in Steve's two listed critiques of AR speakers that item one re-affirms what I have come to understand that you can't design a single speaker to be all things to all people and for that matter, all recording venues. The tens of thousands of brands and variations in speaker design that have come into being since then also supports that notion.

The thread link below is to my AR3a pot setting survey. It's here to remind all where we are now with pot settings vs what evidently was the original intent. Either current owners aren't satisfied with the original 'laid back' highs or, as I postulated earlier, the tweeters have indeed deterioriated somewhat in performance over time.

http://www.classicspeakerpages.net/IP.Boar...8&hl=survey

[onplane]

Either current owners aren't satisfied with the original 'laid back' highs or, as I postulated earlier, the tweeters have indeed deterioriated somewhat in performance over time.

Carl, I don't think the issue is "current owners aren't satisfied with the original 'laid back' highs". I believe the issue is modern recordings with digital playback devices no longer emphasize the high frequencies. In short the “game has changed”.

In all fairness, Carl, you and Roy have speculated on the fact that “the tweeters have indeed deteriorated somewhat in performance over time” and I’m in full agreement. This is a real possibility.

Tom, thanks for your well written memo. I respectfully disagree with your comment … “When the AR-3a was introduced in 1967, ceramic-phono cartridges were practically non-existent.” Now, in all fairness, this may be true for listeners of classical music, but who do you think bought the vast majority of the records in the 60’s? Remember this was the time the Beatles dominated popular music. Classical music never, ever dominated record sales.

Vast majority of the record companies’ customers had those cheap “portable” payback sets and ceramic cartridges dominated those by far. Even consoles had ceramic back then.

Now, the official AR party line differed from both of us! According to AR, record companies enhanced the high frequencies because the speakers at that time had deficient energy output in the treble range.

So there are 3 potential reasons why record companies enhanced high frequencies. In the grand scheme of things, probably all three are true to a degree. Further, it doesn’t make any real difference as long as we all understand:

1. this high frequency emphasis existed back in the 60’s

2. Records were indeed the primary source of music for AR’s customers

3. AR designed their speakers to compensate for “their world”

Well the world has changed. Fast forward 30 years and now we need those highs and Steve mentions how AR responded to the changing world when they introduced the ADD line.

From my perspective, this clearly demonstrates that AR was watching the market, understood the changes and made adjustments at the appropriate time.

Regards,

Jerry

[Diamonds&Rust]

Yet AR maintained that the most realistic sound was most often obtained with the controls at their ‘Norm’ positions, not ‘Max,’ as the Norm positions most closely replicated the ‘concert hall slope’ of the high frequencies as heard by listeners about 1/3 of the way back in the hall.

Critics of AR speakers are quick to point out two things:

1. The spectral balance heard by listeners in a symphonic concert hall was entirely different than that heard by a listener sitting 15 feet from the stage listening to a jazz quintet at the Village Vanguard in NYC or The Jazz Workshop in Boston. In those venues, the highs were much more prominent than 1/3 of the way back in Symphony Hall or Carnegie Hall. AR’s assumption that the “concert hall slope” was the correct and only target for acoustic replication in the home was open to question, to say the least.

2. When AR was able to drive their tweeters harder with the advent (no pun intended) of ferro-fluid in the 1975-6 intro of the ADD line, they did so, and thus completely changed their speakers’ sonic balance. A sharp prosecuting attorney would ask, rhetorically, “If the 3a’s spectral balance was correct, why did the 11—essentially the identical speaker—have such a stronger high end?

Steve (and Tom),

I thought you guys in particular might get a kick out of this:

http://www.thefreelibrary.com/MEHTA'S+...NCH-a0112146906

It seems that the spectral balance of live performances can now be considered an audiophile problem to be solved by the conductor.

My favorite recording of Beethoven's 9th exhibits every "problem" this reviewer cites, and it is those problems which (to a large extent) account for its being my favorite. I think Zubin Mehta does the 9th a favor by keeping it a sonic amalgam rather than a competition between individual "gleaming" parts.

Apparently the new hall's "brightness" is considered an innovation to be emphasized - period.

I find myself wondering if the new preference for increasing brilliance isn't a reaction to our average exposure to "flat" frequency response; accuracy in the maths being the ultimate expression of accuracy at the listening position.

It seems that now a symphony must be spectrally flat to be "correct."

On some CD of mine, talking to the soundman the performer makes a request like, "...can I get everything louder than everything else?" Is that the goal we are chasing today? Does the average listener prefer it this way, or is he / she just accustomed to it?

Bret

[genek]

Does the average listener prefer it this way, or is he / she just accustomed to it?

I think it's a chicken and egg thing, the average listener is accustomed to it and prefers it because of being accustomed to it. The ear/brain will adjust to lots of things, filtering out noise and other distractions over time, but "difference" is always noticed and seldom liked.

Anytime I visit my local audio dealers ("home theater" establishments, because there are no longer any traditional audio salons around here), I find myself tinkering with the tone controls on anything I listen to and turning the treble down about 30% from default. The "new normal" is that much brighter than it used to be.

[genek]

When AR was able to drive their tweeters harder with the advent (no pun intended) of ferro-fluid in the 1975-6 intro of the ADD line, they did so, and thus completely changed their speakers’ sonic balance. A sharp prosecuting attorney would ask, rhetorically, “If the 3a’s spectral balance was correct, why did the 11—essentially the identical speaker—have such a stronger high end?

It’s obvious that the Classic ARs sacrificed a ‘flat’ on-axis HF response in favor of wide dispersion and smooth (albeit somewhat downward-sloping) far-field power response. In terms of listenability and freedom of listener placement, these were very good design trade-offs. In terms of retail success and advantage in quick showroom A-B demo comparisons, they were disastrous design decisions.

I disagree on "sacrificing." Classic AR speakers still represent (to my ear, anyway) the most "natural and least "colored" sound I've ever heard come out of a speaker.

The reasons why AR spectral balance changed after '75 were most likely (a) Edgar Villchur left the company and was no longer setting the "guiding philosophy" for AR, and ( AR's new corporate owners' "guiding philosophy" was "sell the consumer whatever the consumer wants" and "do I ever go to concerts? Yeah, I never miss Black Sabbath whenever they tour here." I wonder if anyone at Teledyne, Audiovox or any of the other companies that subsequently owned AR ever hear any music that didn't originate from an amplifier.

[Diamonds&Rust]

I disagree on "sacrificing." Classic AR speakers still represent (to my ear, anyway) the most "natural and least "colored" sound I've ever heard come out of a speaker.

The reasons why AR spectral balance changed after '75 were most likely (a) Edgar Villchur left the company and was no longer setting the "guiding philosophy" for AR, and ( AR's new corporate owners' "guiding philosophy" was "sell the consumer whatever the consumer wants" and "do I ever go to concerts? Yeah, I never miss Black Sabbath whenever they tour here." I wonder if anyone at Teledyne, Audiovox or any of the other companies that subsequently owned AR ever hear any music that didn't originate from an amplifier.

Foul ! <throws the yellow flag, holds-up the red card, etc>

As Roy was saying the other day: There's almost no opportunity to hear unreinforced acoustic instruments playing even when you go looking for it.

The reason the spectral balance changed in 1975 was because it needed to be changed and it finally could be. I *promise* you, the AR-11 sounded dead, muffled, and downright dull A/Bed against anything else in the shop (in its price range). You are familiar with the "next room" test, no?

Remember when Loggins and Messina's record "Best of Friends" was released (about 1976)? There are a lot of acoustic instruments on that record and it was a very popular "demo". From the next room, the mandolin sounded like there was a mandolin the the "next room" through a pair of Koss 1030s. Through any pair of ARs it sounded like there was a mandolin recording playing next door.

While I agree with you that the AR classics were extremely pleasant to listen to if you were listening to "Der Rosencavalier Waltz" and imagining yourself in row 25 of any reasonable concert hall, they really are not "spectrally balanced" for an accurate reproduction of closely auditioned voice, guitar, and tuba like a Leon Redbone recording or even barbershop quartet recordings.

I know what you mean, but it does depend on context, just like Tom said. The Boston Pops? Get a 3a. Mitch Miller? Get a 3a. The Limelighters? An 11 is a better choice.

In my opinion and to my idea of what "live" sounds like, of course.

Bret

[genek]

While I agree with you that the AR classics were extremely pleasant to listen to if you were listening to "Der Rosencavalier Waltz" and imagining yourself in row 25 of any reasonable concert hall, they really are not "spectrally balanced" for an accurate reproduction of closely auditioned voice, guitar, and tuba like a Leon Redbone recording or even barbershop quartet recordings.

This may be true if you're one of those high end purists who thinks tone controls are the work of the devil. Otherwise, all that is needed is about a 30 degree turn of a knob one way or the other.

[Diamonds&Rust]

This may be true if you're one of those high end purists who thinks tone controls are the work of the devil.

Wait... you say "high end purist" like it's a bad thing.

Tone controls the work of the devil? Well, not all tone controls, just the ones without turnover switches.

I'm not saying that you aren't entitled to your opinion or that mine has any more validity. I just find it difficult to believe that the ADD guys came into AR with the idea of screwing-up a perfectly uncolored and spectrally balanced speaker system in order to sell more. I thought that's what you accused them of. They could *always* have added a speaker, or even a full line of screw-ups. Instead, they went to work trying to get more high-frequency power to evenly match the woofer's power - like the LST. That is to say, they were trying to make the system truly flat instead of merely "uncolored."

Whether they were pandering, or updating, I suppose is a matter of how you look at it. They certainly didn't add the "punchiness" that was sought-after at the time, or drop the prices. They didn't make them small. They didn't make them any prettier with the grills off (maybe they did... it's *hard* to find something as bunz-ugly as a "Classic" AR with its grill off).

If they meant to make the 11 screechy to compete with the up-and-coming designs, they certainly failed miserably, and it wouldn't have taken much to screech right along with the worst of them.

But "getting real" for a second - The tone control haters aren't likely to pick a classic 3a as a high-end speaker. Surely, as you say, high-end speakers are "uncolored."

Paradox?

Bret

PS - I preferred listening to the 3a over the 11, also. I can't say that for any other ADD / pre-ADD model comparison.

It's entirely possible to want to hear Leon Redbone, Black Sabbath, and Leontyne Price the same *day.* Flatt and Scruggs, too. Okay, I confess, I'm straining credulity to make a point. I'm really with Victor Borge on the whole soprano issue... It is not possible for me to want to hear Leontyne Price. Ever. Not even singing a die-aria.

When she starts singing I'm the one that goes "Aieeeeda!"

[genek]

I just find it difficult to believe that the ADD guys came into AR with the idea of screwing-up a perfectly uncolored and spectrally balanced speaker system in order to sell more. I thought that's what you accused them of.

Not quite. I'd say they came in with the idea of selling more and/or spending less to manufacture in order to make more per sale, and the loss of the uncolored and spectrally balanced speaker system was just the result.

As for "high end purist" being a bad thing, it is if you're so bent on the snob appeal of "pure" "straight wire with gain" that you end up rejecting an enormous amount of records because their sound is "off" in ways that could be corrected with a 30 degree turn of a knob. Don't know how many of those types are still around, but back in the 70's they seemed to populate the hifi scene a lot.

[Diamonds&Rust]

Not quite. I'd say they came in with the idea of selling more and/or spending less to manufacture in order to make more per sale, and the loss of the uncolored and spectrally balanced speaker system was just the result.

I don't want to argue, and I don't think we really are, but it's the ADD guys who spent money on the development of the new, higher-output tweeters, eventually adding ferrofluid. It's also the ADD guys who brought in the expensive "Compulytic" capacitors in the 10pi and 11. And the cabinets had to be more expensive to manufacture if you stop and think about the cuts and the big pieces of walnut along the bottom. They tried to fix the potentiometer problem and the diffraction issues even dumping the grill.

In fact, I cannot see even one place where there would have been a cost savings to manufacture an 11 over a 3a. Quite the opposite. Maybe the aluminum badge was cheaper than brass?

Even if our opinion is that they didn't really improve anything, I can see no reason to believe their motives weren't pure. Of course they wanted to sell more speakers, but making a better speaker would have been the way to do that.

The AR-3 was the best, least colored, speaker AR ever made. And then came the 3a. I suspect the 11 was another attempt to improve things, that's all.

Now I'll shut-up, let you tell me where you see any cost savings, and I'll ponder your points in silence.

Bret

PS - You make a good point about recordings, whether from the 70s or today. It's annoying that there isn't an absolute standard so we don't need to re-equalize our systems for every record we play.

[genek]

Now I'll shut-up, let you tell me where you see any cost savings, and I'll ponder your points in silence.

PS - You make a good point about recordings, whether from the 70s or today. It's annoying that there isn't an absolute standard so we don't need to re-equalize our systems for every record we play.

I'm sure it was all perception on my part when I first saw and heard the post-classic models. Oiled walnut replaced by vinyl, natural fabric replaced by synthetic knit stuff, accompanied by sound I didn't like. In other words, everything that had until then distinguished AR speakers from "everyone else's" was gone, and the speakers (to my eyes and ears) looked and sounded the same as "everyone else's."

Of course, I'm still listening to the same speakers I had back in 1975 and haven't found myself forced to actually choose new speakers, so it's entirely possible that sometime during the 25 years that followed post-Villchur AR may have turned out something I would have liked had I had the necessary motivation to listen to every model they ever produced...

[2Turbos]

I'm sure it was all perception on my part when I first saw and heard the post-classic models. Oiled walnut replaced by vinyl, natural fabric replaced by synthetic knit stuff, accompanied by sound I didn't like. In other words, everything that had until then distinguished AR speakers from "everyone else's" was gone, and the speakers (to my eyes and ears) looked and sounded the same as "everyone else's."

Of course, I'm still listening to the same speakers I had back in 1975 and haven't found myself forced to actually choose new speakers, so it's entirely possible that sometime during the 25 years that followed post-Villchur AR may have turned out something I would have liked had I had the necessary motivation to listen to every model they ever produced...

I may be like some of us on this forum who own other speakers along with the AR-3a.

It seems to me most good speakers are able to reproduce music to a very satisfying degree.

If the speaker is able to do this, the music should be the most present thing, not the speaker.

Yes, there are speakers that do get in the way of the music for sure. The AR-3a is NOT one of them.

Use the level controls to adjust the speaker to the room, preference and listening position. After that,

listen to the music and forget about the speaker!

[genek]

I may be like some of us on this forum who own other speakers along with the AR-3a.

It seems to me most good speakers are able to reproduce music to a very satisfying degree.

If the speaker is able to do this, the music should be the most present thing, not the speaker.

Yes, there are speakers that do get in the way of the music for sure. The AR-3a is NOT one of them.

Use the level controls to adjust the speaker to the room, preference and listening position. After that,

listen to the music and forget about the speaker!

This is the way I've set up my classic AR's every since they were new.

Anytime I've moved or changed something else within the system, out

come 3 or 4 old standby recordings, classical, jazz, rock. I adjust the

speaker trimmers for a happy medium that enables me to make

listening adjustments to my taste with the amplifier tone controls,

and once set the speakers remain undisturbed until something changes

again. I've run everything from the 1812 Overture to In-a-Gadda-da-

Vida through them, and have never felt that they were "better suited"

to any particular kind of music. If the sound coming out of the speaker

seems "laid back" or "forward," it's not because the speaker is that

way but because that's the way the record was made.

[soundminded]

In a recent thread, “Ar-3a Tweeter – Carl’s and Roy’s Observations – Potential solution??” Jerry “onplane” comments on the AR-3a.

“the other two drivers suffered from the same syndrome, this might not be a terrible problem. The mid and woofer, however, don't seem to have suffered this same fate. This leaves the high frequencies coming from the Ar-3a somewhat "muted".

In that same thread genek mentioned: So AR speakers have no settings that "boost" mid or treble; "full" is what you use with an amp and recordings that don't have "excessive" highs …

Genek is correct! There is no “boost” setting and for a very good reason. Back when the AR-3a’s were designed the primary source of music was vinyl records and the vast majority of playback devices had ceramic phono cartridges, which were woefully inadequate in accurately reproducing high frequencies. So in order to sell their recordings, the record companies artificially “enhanced” high frequencies to compensate for this deficiency.”

No, this is an incorrect assumption, and it has nothing to do with the balance of high frequencies in AR speakers. When the AR-3a was introduced in 1967, ceramic-phono cartridges were practically non-existent and were only used in the cheapest one-piece music-box sets, and rarely in high-fidelity sound systems. Nearly every quality system had either a Shure (or Stanton, etc.) magnetic cartridge, and some even used moving-coil-type cartridges. Ceramic cartridges became nearly obsolete nearly ten years earlier by the time quality stereo became a reality. The real change to improved sound from the old ceramic versions was the introduction of GE’s variable reluctance cartridge; then along came Shure Brothers and others with their superior moving-magnet cartridges.

The engineering philosophy of AR engineers was without exception to design speakers with the highest-possible sonic accuracy. Remember AR’s famous words in one of the brochures: “The accurate reproduction of sound is one of man’s most benevolent gifts to himself.” AR felt that accuracy required flat acoustic-power response in the listening environment that replicated the sound field one would experience in a concert hall or auditorium, etc., and to get this degree of flat acoustic power, the speaker would have to exhibit extremely wide dispersion at the highest frequencies. Unfortunately, there was a bit of a trade-off between efficiency and wide dispersion in the design of tweeters capable of providing the extended power response required. Therefore, at the time, the ¾-inch AR-3a hard-dome tweeter had somewhat lower sensitivity than either the midrange or woofer, and this resulted in what many people felt was a sense of “reticence” or a “laid-back character.” AR considered padding down the woofer and midrange to match the tweeter’s sensitivity, but this was an unattractive proposition considering the already low efficiency of the AR-3a.

What AR found was really happening, however, was that to obtain sound in the living room that approaches the original spectral balance of live sound in a concert-hall seat, it is actually necessary to turn *down* the level controls of both the mid-range and tweeter units in an AR-3a! The reason is simply this: microphones for both live broadcasts and recording sessions are invariably set up in the near field of the sound source, while concert-goers are in the reverberant field. In concert halls the reverberant field has a far more drastic high-frequency roll off (relative to the bass) than is true of living rooms. This was clearly shown in measurements made by Bolt, Beranek & Newman in their (unpublished but reported) tests of nine concert halls, and these halls were measured when empty! Thus what is put into the speaker systems is a near-field spectrum, usually made even “hotter” because the instruments are aimed at the microphones during the recording process. To produce the *same* spectral balance at the ears of listeners in both concert halls and living rooms, the reproducing system must compensate for the difference between high-frequency roll offs in concert hall and living rooms. For those circumstances requiring truly flat acoustic energy input to a room, such as the playback of live recordings made at home, etc., then the treble tone control might have to be turned to maximum or the use of additional tone control input.

--Tom Tyson

"In concert halls the reverberant field has a far more drastic high-frequency roll off (relative to the bass) than is true of living rooms. This was clearly shown in measurements made by Bolt, Beranek & Newman in their (unpublished but reported) tests of nine concert halls, and these halls were measured when empty!"

Tom, do you have a link to a document stating this or know when Beranek first made it known? I came to this same conclusion analyzing the problem in 1974, I'd be interested to see when Beranek first saw it. I've got data for about 200 concert halls around the world and they show a typical pattern of 1khz RT of around 2 seconds (or more) and an 8khz RT of around 1 second. This is due to the more rapid decay of high frequencies than mid and low frequencies. This is the result of 1) the frequency selective characteristic reflection/absorption of the materials which create the acoustic effect and 2) the frequency selective transmission of sound in air, humidity playing an important role. I'm currently studying two papers from Beranek's web site, one comparing 59 concert halls and the other comparing 20 opera houses both trying to correlate the preferences of conductors and other "golden ears" with measured parameters. Relative rate of HF decay is not one of them.

The fly in AR's argument of course (if it really was their argument) is that in a live musical performance, the initial arrival of sound via the shortest path between musician and listener carries the full high frequency compliment of overtones in a proportion similar to what close miking picks up while the change in spectral balance occurs over the time of decay of each note. This is critical in understanding and reproducing the musical timbre of instruments as they would be heard at live concerts, their characteristic sound is one of both clarity due to sharp intitial transient attacks resulting from the presence of high frequencies and mellowness due to their increasingly relative absence over time during decay which modifies their timbre. What this means is that the same musical instrument will have a different subjective timbre depending on what specific acoustical environment it is heard in and that it is impossible to reproduce the timbre of the instrument as it would be heard in any concert hall without also recreating the reverberant sound field as a dynamic event at the same time, they are inseparably linked being different aspects of the same phenomena. Therefore no single FR of a sound reproducing system in a small room no matter what it is will work because the timbre is controlled by spectral balance of each note over time which is not captured on the recording. The other problem then is that when used to reproduce music which is normally heard in a small venue often at a closer distance where this effect is much less a part of the total sound because mid frequency RT is much shorter, say under 1 second, the speaker designed for "average" spectral balance of a concert hall is much too mellow, too little high frequency energy is present unless it is re-equalized.

The problem of reproducing sound of musical instruments as they would be heard performing in your own listening room and as they would be heard at a concert venue are actually two different problems the second being much harder to solve than the first. (Neither are close to being solved so far. There are obviously critical elements even in the first, the easier of the problems which have been overlooked.) Then there are the variables of what the acoustics of the listening room does to the sound generated by the loudspeakers before it reaches the listener which is not really addressed at all in any designs and the variables of the spectral balance of recordings (my own experience being that even with cds, they are all over the map even from the same recording company) and it is not hard to understand why at the current state of the art, it is impossible to buy or build an electronic recording reproducing sound system which sounds to critical ears like actual music.

In fact, of the billions of sound systems in existence, it is probable that no two sound exactly alike playing the same recording. From this perspective the ceaseless pursuit of audio perfection by some audiophiles with no limits on cost or practicality is laughable and pathetic if for no other reason than it is hopeless. What they lack in knowledge and understanding of what is possible, they more than make up for in enthusiasm, energy, determination, and money.

What is also remarkable is that over 40 years ago, AR was able to create demonstrations which produced sound from its loudspeakers very similar to live music played in the same room even though it must be admitted that the demonstrations had to be very carefully constructed and the entire process very highly contrived. It's a feat few if any have duplicated successfully in the intervening years which speaks volumes as to how little real progress in this science has actually been made. In fact, in some ways we may actually have gone backwards, the engineers having given up and all but admitted that the problem has beaten them, they simply cannot solve it.

[genek]

The fly in AR's argument of course (if it really was their argument) is that in a live musical performance, the initial arrival of sound via the shortest path between musician and listener carries the full high frequency compliment of overtones in a proportion similar to what close miking picks up while the change in spectral balance occurs over the time of decay of each note.

I think it is worth keeping in mind that AR's "arguement" was merely an explanation of why they chose to place the dots for "norm" where they are, and has not all that much to do with their speakers' on-axis spectral balance. The design target for classic AR speakers was not the sound of any particular concert hall, but flat on-axis frequency response in an anechoic chamber with all the pots or switches at what AR deemed "flat" (full increase rather than at the dot), and most of the response curves I saw in various audio publications of their period appeared to indicate that that's pretty much what they achieved.

For most owners who had amps with tone controls, where to set the pots was a relatively trivial issue that they spent maybe half an hour fiddling over the first time they took their speakers out of the box.

[soundminded]

I think it is worth keeping in mind that AR's "arguement" was merely an explanation of why they chose to place the dots for "norm" where they are, and has not all that much to do with their speakers' on-axis spectral balance. The design target for classic AR speakers was not the sound of any particular concert hall, but flat on-axis frequency response in an anechoic chamber with all the pots or switches at what AR deemed "flat" (full increase rather than at the dot), and most of the response curves I saw in various audio publications of their period appeared to indicate that that's pretty much what they achieved.

For most owners who had amps with tone controls, where to set the pots was a relatively trivial issue that they spent maybe half an hour fiddling over the first time they took their speakers out of the box.

As I recall it (going back 40 years or more) the main thrust of AR's sales pitch was that they sold the most accurate speakers available at any particular price point and their most expensive speaker was the most accurate speaker they knew how to build. This was at a time when people still listened to live unamplified music, went to real concerts where there was no lip sync, when there was still music education in public schools, and when the general level of musical knowledge was probably higher than it is today. AR never targeted the heavy metal, acid rock, punk rock, deafen yourself while you're stoned on drugs segment of the market which was left to Altec, JBL, and Cerwin Vega. It targeted more "thoughtful" people.

In those days, there actually was a somewhat agreed to meaning to the term "accuracy" even if there wasn't agreement on which inaccuracies were least objectionable. Of course, due to the variables I cited in my previous posting, any speaker sounding accurate playing one recording might sound inaccurate playing another but that's the nature of the products of the recording technology that had to and still has to be contended with. There were legitimate technical and marketing reasons why inaccuracies were deliberately built into many recordings. That said, generally the market agreed with AR which explains why AR was so successful at least for a time during this golden era. Not only did it own a large chunk of the market as a percentage of sales, many professional musicians and music ciritcs chose AR speakers for their own private use and they were often chosen for professional use. So did institutions. The speakers in the main listening room in the music department at my college were AR3s. One of my professors of electrical engineering had just bought AR3as when I took his circuit design course. To many people, these were the state of the art. Were they perfect? Obviously no but even by today's standards, when new they were likely more accurate than most of what is out there at any price. Accuracy is not a matter of preference, it's a matter of being able to tell the difference between the reproduction and the original. This was the marketing value of the live versus recorded demos. Accuracy is not a moving target, just one that's been impossible so far to hit. This is what I meant when I said we are going backwards, it strikes me that today's products miss that mark farther than AR did which is sad because we still know how to do everything AR did and much more. The advantage of having nothing resembling an objective goal to meet is that you can create an endless stream of products each advertised as superior to others including your own previous models and there is no way to dispute it. And as long as the price keeps getting higher, a lot of people will believe it. BTW, this is why I still contend as I said in a thread on tweaks that a determined hobbyist can build better (more accurate) sounding speakers than professionals if he tries long enough, hard enough, and gains enough insight but he will have to do an awful lot of rethinking and be very determined to get there.

[Steve F]

A few more random thought on this very interesting topic—

1. The AR-3 L v R demonstrations featured a string quartet recorded outdoors (in “God’s anechoic chamber,” as I think Villchur put it), then replayed in a concert hall, so both the live quartet and the anechoic recording were influenced for the first and only time by the hall’s acoustics. This was a perfectly legitimate way to go about things.

The fact that a string quartet had limited HF energy and a concert hall tended to naturally attenuate HF energy at the listening position both contributed to the AR-3’s ability to pull off such a demo convincingly—which it did. The AR-3 would NOT have been able to pull off a convincing L v R demo of, say, the 1964 John Coltrane Quartet in a smaller venue. Elvin Jones’ drums/cymbals would not have been reproduced by the AR-3 with the power and intensity required foe a lifelike presentation in the near field.

2. By 1976, AR had the technology that they didn’t have in the early 1960’s. The AR-10Pi/Neil Grover L v R demos of Grover playing his jazz drumkit in the nearfield to a critical BAS audience was totally convincing. I am a jazz drummer myself and I worked for Zildjian Cymbals for a time, and met Neil Grover at a professional percussion tradeshow. We spoke at some length about the AR demo, and he recalled fondly how difficult it was to get everything just right. The speakers, it turns out, were the least of their worries. Amps kept blowing, tape recorders kept saturating (even at 15 ips), mic positions were critical, etc. The 10Pi’s never skipped a beat (no pun intended) and were actually the easiest part of the reproduction chain. At that BAS demo in 1976, I remember asking how they (AR and Grover) dealt with the fact that the varnish wears off the drumstick tip after a while, drastically changing the sound. It was a difficult factor to deal with, as they had to have several pairs of sticks on hand, matched for sound characteristics, and be careful to only use any one pair for just so long.

When I spoke to Neil about this almost 30 years after the original event, he grew wide-eyed and said to me, “So YOU were the one! What a difficult thing the sticks turned out to be. We were amazed that someone in the audience would pick up on that! It’s good to meet you, after all this time.”

I would hazard a guess and say that the 10Pi (and 11) could do a convincing job on the string quartet, but not the other way around.

In any event, from 1954 to 1978 or 80-ish, AR was trying their best to make accurate speakers, even though the notion of what, exactly, constituted “accuracy” was a somewhat amorphous concept.

Steve F.

[genek]

As I recall it (going back 40 years or more) the main thrust of AR's sales pitch was that they sold the most accurate speakers available at any particular price point and their most expensive speaker was the most accurate speaker they knew how to build. This was at a time when people still listened to live unamplified music, went to real concerts where there was no lip sync, when there was still music education in public schools, and when the general level of musical knowledge was probably higher than it is today. AR never targeted the heavy metal, acid rock, punk rock, deafen yourself while you're stoned on drugs segment of the market which was left to Altec, JBL, and Cerwin Vega. It targeted more "thoughtful" people.

I don't see why that should make a difference. Sure, feeding the output of a rock group's mixing boards and amps through AR speakers will produce a different sound compared to feeding it into the other brands you mention, but setting the recording mikes up in front of the speakers at one of these groups' concerts shouldn't be terribly different from putting them in front of a symphony orchestra. If the sound in the hall is raucous and shrieking, the recording should play back the same way. I can make my AR's shriek just as much as any other brand's, all I have to do is feed them a signal that sounds that way (the speakers can handle it much better than my ears will). The fact that this can't be done in reverse, i.e., make a modern speaker most commonly used for modern amplified music reproduce the sound of nonamplified music as a classic speaker can, is an indication that things have, as you say, gone backwards.

As for whether an amateur could conceivably equal the results of the classic speakers' professional designers, the original speakers were designed without the benefit of modern technology, and the major tools involved were listening and the willingness to make a great many false steps. So I would agree that there's no reason a non-pro with a good ear, lots of time and patience to experiment and really deep pockets to fund a lot of failed trials and errors could not do the same eventually.

[genek]

The fact that a string quartet had limited HF energy and a concert hall tended to naturally attenuate HF energy at the listening position both contributed to the AR-3’s ability to pull off such a demo convincingly—which it did. The AR-3 would NOT have been able to pull off a convincing L v R demo of, say, the 1964 John Coltrane Quartet in a smaller venue. Elvin Jones’ drums/cymbals would not have been reproduced by the AR-3 with the power and intensity required foe a lifelike presentation in the near field.

Could any other speaker available at the time have done it, or was it just beyond the state of the art in 1964?

[Steve F]

No, there probably was not another speaker of that time period that could reproduce a jazz quartet with spectral accuracy at lifelike SPLs, even if the requisite recording and amplification equipment existed. (Remember, the recording and amplification equipment was barely—just barely—up to the task in 1976.) Other speakers of the early ‘60’s time period were too uneven in their response and did not have the deep bass extension with the same lack of distortion as the AR-3 to have performed as well in an L v R test.

The AR-3 had exceptional bass-midrange response in terms of accuracy, smooth, even response, lack of distortion, transient behavior, etc. The high frequencies were certainly as good as the bass and MR in terms of the quality and accuracy of reproduction. In fact, AR even said that within its frequency range, the new HF dome radiator achieved the highest degree of accuracy of any of the three radiators.

However, the 1 3/8” tweeter simply did not have the output necessary to achieve lifelike SPLs in small (essentially nearfield) acoustic spaces. So what worked for a string quartet in a concert hall would not have worked for a jazz group in a small club setting. The 3a would have been better, but still would not have done as well as the 10Pi/11.

It would have been interesting to have heard an AR-9 under those same L v R conditions. I remember Victor Campos (the Head of AR’s Engineering, who managed the 10Pi/Grover demo exercise) telling me that they tried the LST, and it didn’t do as well as the 10Pi—too much overlap and interference from the multiple mid-high drivers. The single MR and HF units of the 10Pi—coupled with their exceptionally wide dispersion and amazing power-handling—proved to be the key to replicating “reality.”

Steve F.

[dynaco_dan]

No, there probably was not another speaker of that time period that could reproduce a jazz quartet with spectral accuracy at lifelike SPLs, even if the requisite recording and amplification equipment existed. (Remember, the recording and amplification equipment was barely—just barely—up to the task in 1976.) Other speakers of the early ‘60’s time period were too uneven in their response and did not have the deep bass extension with the same lack of distortion as the AR-3 to have performed as well in an L v R test.

The AR-3 had exceptional bass-midrange response in terms of accuracy, smooth, even response, lack of distortion, transient behavior, etc. The high frequencies were certainly as good as the bass and MR in terms of the quality and accuracy of reproduction. In fact, AR even said that within its frequency range, the new HF dome radiator achieved the highest degree of accuracy of any of the three radiators.

However, the 1 3/8” tweeter simply did not have the output necessary to achieve lifelike SPLs in small (essentially nearfield) acoustic spaces. So what worked for a string quartet in a concert hall would not have worked for a jazz group in a small club setting. The 3a would have been better, but still would not have done as well as the 10Pi/11.

It would have been interesting to have heard an AR-9 under those same L v R conditions. I remember Victor Campos (the Head of AR’s Engineering, who managed the 10Pi/Grover demo exercise) telling me that they tried the LST, and it didn’t do as well as the 10Pi—too much overlap and interference from the multiple mid-high drivers. The single MR and HF units of the 10Pi—coupled with their exceptionally wide dispersion and amazing power-handling—proved to be the key to replicating “reality.”

Steve F.

Hi Steve;

Thank you for your very interesting write-up.

[Diamonds&Rust]

The single MR and HF units of the 10Pi—coupled with their exceptionally wide dispersion and amazing power-handling—proved to be the key to replicating “reality.”

Since you were there, do you recall if the tweeters were the later ferrofluid black cloth, or the earlier lighter colored ones?

I have a suspicion, but I'd like you to confirm it for me.

Bret

[Steve F]

> Since you were there, do you recall if the tweeters were the later ferrofluid black cloth, or the earlier lighter colored ones?

> I have a suspicion, but I'd like you to confirm it for me.

If I said I was absolutely 100% positive about the color of the tweeters, I'd be lying. The first 10Pi/Neil Grover L v R was at the Summer CES in Chicago in June 1976, I believe. The BAS demo followed that one by several months, spurred on by great local Boston demand for it after the publicity of that first demo. So the 10Pi was probably 18 months or so into its life, if that helps you nail down the tweeter variant.

I still find the most fascinating aspect of this subject to be that the LST didn't do as well.

Steve F.