Jump to content
The Classic Speaker Pages Discussion Forums
Sign in to follow this  
Howard Ferstler

Review of AR-303 by Julian Hirsch in June of 1995, He A/Bd the speaker against the AR-3a

Recommended Posts

I was pretty happy with the Stereophile review of the 303, (despite wishing for a bit less ambivalence, either way). I certainly have no issues whatsoever with their measurement data.

The 303 was a carefully designed system, using full-custom drivers. During its development, hundreds of measurement curves were taken, optimized and traded-off. If the speaker game was as easy as making any one or two curves, (single or averaged), look better, this would be extremely easy to do with modern electronics.

But this is not the game. In fact, the reality is that speaker designers must consider and balance a huge amount of technical data in an attempt to optimize a product. Similarly, reviewers who attempt to introduce objective criteria into their loudspeaker reviews are faced with the problem of selecting presentations that are in some way useful to the reader and, over the long haul, compatible with their own listening evaluations.

-k

Attached is a copy of Julian Hirsch's review of the AR-303 in the June, 1995 issue of Stereo Review. I have to assume you were happy with his assessment, too.

In it, he does a direct measuring and listening-session comparison between that newer speaker and the AR-3a. Note that he got the crossover specs wrong with the comments on the AR-3a, saying they were 550 and 6.5 kHz, when they were actually 575 and 5 kHz. Well, nobody's perfect. Interesting that the 303 woofer is run up to a somewhat higher (and more directional) frequency than that of the 3a.

The important thing will be his listening/comparing work, and everybody here who owns either speaker will certainly be interested.

Howard Ferstler

Share this post


Link to post
Share on other sites
The important thing will be his listening/comparing work, and everybody here who owns either speaker will certainly be interested.

I tried a pair of the 303s when they were new, side-by-side with my 2ax's. As you would expect, the 303 was as much of an improvement over the 2ax wrt low bass as the 3a. The rest of its sound had a definite "family resemblance" to the older speaker with its level controls turned up a bit. The major factors that caused me not to take them were that they came up short compared to the 2ax when it came to producing a sound that remained stable from anywhere in the room (I experienced the same from a pair of AR-218v tweeters acquired from Carl in my 3a's, BTW) and my distaste for the appearance of the all-black speakers in my living room. If they had been a pair of the rosewood versions with less "beamy" tweeters they might be in my living room today instead of the 3a's.

Share this post


Link to post
Share on other sites
Attached is a copy of Julian Hirsch's review of the AR-303 in the June, 1995 issue of Stereo Review. I have to assume you were happy with his assessment, too.

In it, he does a direct measuring and listening-session comparison between that newer speaker and the AR-3a. Note that he got the crossover specs wrong with the comments on the AR-3a, saying they were 550 and 6.5 kHz, when they were actually 575 and 5 kHz. Well, nobody's perfect. Interesting that the 303 woofer is run up to a somewhat higher (and more directional) frequency than that of the 3a.

The important thing will be his listening/comparing work, and everybody here who owns either speaker will certainly be interested.

Howard Ferstler

I was didappointed by the review. For one thing JH didn't have any FR curves. I was hoping for something more definitive meaning curves for both of them. Obviously his discussion of how bright or muted the two speakers sounded by comparison to each other was inconsistent. Julian must have been nearing the end of his days.

The curve speaker Dave showed for the AR3a was IMO rather grim looking. It was dominated by the midrange. This adds to the disappointment of the muted highs. The curves AR published were for the individual drivers. You'd think that between the skill AR displayed, their attention to detail, their test facilities they'd have been more succesful splicing them to create a smooth overall curve.

Your curves for AR3a and AR303 look nothing like either Speaker Daves for AR3a nor Stereophile's for AR303. The AR3a looks much flatter and the AR303 doesn't have the 60 hz resonant peak and is not underdamped. It also suggests the AR303 should sound brighter. Much has been made of the potentiometer level settings. I'm coming to the opinion that it was right to get rid of them in newer speaker designs and rely more heavily on preamp level equalization which most audiophiles shudder just thinking about. Nevertheless it makes the speakers sound more uniform, more predictable, doesn't rely on misadjustments. The switches on AR9 work very well as a substitute and I think are the best choice. KLH did the same.

Taking all of this into cosideration and with my recent experience with AR2ax, I don't think I ever heard AR3 or AR3a to their best advantage except in the LvR demos with AR3. In all other situations where the level controls and preamp tone controls were set at their indicated flat settings and there were no equalizers, these speakers came off as less than impressive. I'm now certain they had much more to give than I heard from them. Even with a friend's pair we never fully explored what AR3 could do.

Looking at Dave's graphs showing the effects of the moldings around the front of the baffle, they make no sense to me. The greatest effect should be at the lower frequencies especially in the tweeter's range. This is because even for an AR3a tweeter, lateral radiation, almost 90 degrees off axis would decrease as frequency increases. So does material reflectivity. Therefore at the lower end of the tweeter's range the energy reaching the wood would be greater and more of it reflected than at higher frequencies. With the treble rolled off, the tweeters needed all the help they could get anyway even if it were true that the wood reflects that much. The idea that reflections equal distortion is just one more audiophile absurdity. The only place where there aren't reflections is in an anechoic chamber and while that may be useful for making certain measurements, I can't imagine a worse place to listen to music.

Share this post


Link to post
Share on other sites
I agree with you about the close-reflections issue, but obviously some people take them very seriously. They may matter in some rooms, with some recordings, and with the listener sitting close.

The entire industry has taken them seriously for 25 or more years; NOBODY builds speakers like that anymore.

I have amply demonstrated the adverse effects of the Smaller Advent's edge overhang upon the direct response out to +/- 45°, and that's with a moderate-dispersion constant-directivity waveguide. Using a cone or wide-dispersion dome, the impacts would be even more significant.... ;)

Share this post


Link to post
Share on other sites
The entire industry has taken them seriously for 25 or more years; NOBODY builds speakers like that anymore.

I have amply demonstrated the adverse effects of the Smaller Advent's edge overhang upon the direct response out to +/- 45°, and that's with a moderate-dispersion constant-directivity waveguide. Using a cone or wide-dispersion dome, the impacts would be even more significant.... ;)

Not trying to start (or re-open) any kind of food fight, but----

All the discussion we've had in these pages over the last year or two has amply demonstrated that despite AR's attempts with the 3a/5 at 'max dispersion' or cd or whatever you'd like to call it, their non-flush driver mounting and wildly intrusive cabinet molding--combined with only a mid-'60's understanding of everything-- conspired to sabotage their efforts.

Really, the only relevant, realistic, true assessment of AR's original "wide dispersion as benefit to the home listener" philosophy can be made with the AR-11/10 Pi. Period.

These speakers corrected all the peripheral shortcomings of the 3a:

- truly flush-mounted drivers

- a truly non-intrusive, acoustically transparent grille, and

- no intrusive cabinet molding whatsoever

The 11/10pi's design gets all the "sub-space interference and distractions" of the 3a out of the picture, and lets AR's original design premise come through, unfettered by secondary interference.

Couple the 11's 'corrections' of the 3a's mistakes with its ff-cooled tweeter (that allowed the HF response to be brought up in level correctly with the rest of the spectrum--finally), and you have a speaker with which a true evaluation of the original AR premise can be made.

The only fair, scientifically-valid way to evaluate whether or not we think AR's 'wide dispersion/cd approach' is a valid approach for home loudspeakers is to listen to and assess the performance and sound of a speaker that correctly embodies that design.

Such a loudspeaker is the 11, not the 3a.

I'm tired of reading arguments as to whether the 3a was 'good' or 'outmoded' because of it wide dispersion 'intentions.' The 3a had so many well-documented problems that prevented its wd/cd attempts from fully getting through that to try to eval it on that basis is a waste of time.

Get some 11's in good condition, listen to those, compare them to other designs embodying other radiation philosophies and make a determination from there.

Using the 3a as a representative of 'good' max dispersion design is scientifically flawed and therefore of no value.

I further submit that we limit this discussion/eval to forward radiating/single baffle loudspeakers only (no triangular cross-section Allisons of multi-panel LSTs). This will limit the discussion/eval to the most commonly-used home speaker type, and eliminate unnecessary variables from clouding our discussion and conclusions.

In short: Use the 11, not the 3a. Then decide.

Steve F.

Share this post


Link to post
Share on other sites
Not trying to start (or re-open) any kind of food fight, but----

All the discussion we've had in these pages over the last year or two has amply demonstrated that despite AR's attempts with the 3a/5 at 'max dispersion' or cd or whatever you'd like to call it, their non-flush driver mounting and wildly intrusive cabinet molding--combined with only a mid-'60's understanding of everything-- conspired to sabotage their efforts.

Really, the only relevant, realistic, true assessment of AR's original "wide dispersion as benefit to the home listener" philosophy can be made with the AR-11/10 Pi. Period.

These speakers corrected all the peripheral shortcomings of the 3a:

- truly flush-mounted drivers

- a truly non-intrusive, acoustically transparent grille, and

- no intrusive cabinet molding whatsoever

The 11/10pi's design gets all the "sub-space interference and distractions" of the 3a out of the picture, and lets AR's original design premise come through, unfettered by secondary interference.

Couple the 11's 'corrections' of the 3a's mistakes with its ff-cooled tweeter (that allowed the HF response to be brought up in level correctly with the rest of the spectrum--finally), and you have a speaker with which a true evaluation of the original AR premise can be made.

The only fair, scientifically-valid way to evaluate whether or not we think AR's 'wide dispersion/cd approach' is a valid approach for home loudspeakers is to listen to and assess the performance and sound of a speaker that correctly embodies that design.

Such a loudspeaker is the 11, not the 3a.

I'm tired of reading arguments as to whether the 3a was 'good' or 'outmoded' because of it wide dispersion 'intentions.' The 3a had so many well-documented problems that prevented its wd/cd attempts from fully getting through that to try to eval it on that basis is a waste of time.

Get some 11's in good condition, listen to those, compare them to other designs embodying other radiation philosophies and make a determination from there.

Using the 3a as a representative of 'good' max dispersion design is scientifically flawed and therefore of no value.

I further submit that we limit this discussion/eval to forward radiating/single baffle loudspeakers only (no triangular cross-section Allisons of multi-panel LSTs). This will limit the discussion/eval to the most commonly-used home speaker type, and eliminate unnecessary variables from clouding our discussion and conclusions.

In short: Use the 11, not the 3a. Then decide.

Steve F.

IMO AR's chief shortcoming was its spectral balance. Playing comercially produced recordings with tone controls set in their sacrosanct flat position and midrange and tweeter controls set at their indicated dotted position they simply did not produce enough high frequencies relative to the bass and midrange. This made them sound muffled. Musical instruments did not sound the way they were supposed to. The speakers lacked clarity and naturalness. The argument Villchur et al made about concert halls whether as a design goal or to justify a performance shortcoming was wrong. Furthermore studies show people like hearing high frequencies. It stimulates their brains. The precise reproduction of high frequencies is among the most critical and difficult of all aspects of sound reproduction to get right. Perhaps that explains why there are probably more kinds of high frequency reproducers than all other loudspeaker driver types combined. AR11 like AR9 does not correct the problem, it merely trades one set of problems for another. The problem still hasn't been understood or solved. But this is the area where KLH beat AR and why so many like me bought them in preference to AR even though AR measured better and got better reviews. AR should have spent more time listening to real recordings and real musical instruments. It should have devoted more time trying to make recordings sound more accurate than trying to make speakers measure closer in anechoic chambers to the theoretical paradigm of the day.

Share this post


Link to post
Share on other sites
The 11/10pi's design gets all the "sub-space interference and distractions" of the 3a out of the picture, and lets AR's original design premise come through, unfettered by secondary interference.

Couple the 11's 'corrections' of the 3a's mistakes with its ff-cooled tweeter (that allowed the HF response to be brought up in level correctly with the rest of the spectrum--finally), and you have a speaker with which a true evaluation of the original AR premise can be made.

I never heard the AR-11, but did hear a pair of 10pi's side-by-side with 3a's at my dealer when the ADDs first came out. The 10pi definitely had the ability to be set "brighter" than the 3a, but it was not hard at all to set the HF and MR switches to match sound with the 3a, and once that was done there was no discernible difference, either sitting in one spot or moving around the listening space. I expect there probably were differences you could have detected with the right measuring equipment, but I've never been in the habit of carrying instrumentation with me when going to audio dealers.

Share this post


Link to post
Share on other sites
I never heard the AR-11, but did hear a pair of 10pi's side-by-side with 3a's at my dealer when the ADDs first came out. The 10pi definitely had the ability to be set "brighter" than the 3a, but it was not hard at all to set the HF and MR switches to match sound with the 3a, and once that was done there was no discernible difference, either sitting in one spot or moving around the listening space. I expect there probably were differences you could have detected with the right measuring equipment, but I've never been in the habit of carrying instrumentation with me when going to audio dealers.

While both Soundminded and Gene make interesting points/observations about the 10 Pi/11's spectral balance, that is not the exact point I was raising.

I'm actually targeting my post to Z, HF, k, Speaker Dave to try to bring about a discussion/evaluation of the 'rightness' or 'wrongness' of the original AR design premise: the premise that said max dispersion was the correct approach in home loudspeakers. As Z has quite accurately pointed out, home loudspeakers haven't been designed that was for quite some time now. But can we use the 3a as an example of why or why not? I say no--the 3a is not the speaker with which that determination can be made.

My contention is that the 3a/5 have so many peripheral problems with non-flush drivers, intrusive cab molding, not truly acoustically-transparent grilles--and then on top of it all, non-ff-cooled tweeters that were far down in level--that the original AR premise of max dispersion cannot be given either a 'thumbs up' or a 'thumbs down' by eval'ing the 3a. Too many other of the 3a's problems get in the way.

Use the 11. (Of course, you're going to have to find some good 11's first. I suppose we could use 91's or 78LS's or 303's, but I'd rather stick with the 11, since that was the 3a's immediate successor.) Don't focus on its 'spectral balance' or your personal feelings about the 11's sound relative to the 3a. Focus on its dispersion, the way it engages the room, all the factors that we've discussed here before. The 11 is 'pure' enough in the secondary areas ( diffraction, etc.) to let its design premise come through. When you hear an 11, you'll hear AR's philosophy clearly for the first time--free of the cab molding issue, free of the non-flush driver issue, and free of the depressed HF energy issue.

Now we can eval AR's max dispersion premise properly, and you can come to a more scientifically-valid conclusion.

Let's stay on topic.

Steve F.

Share this post


Link to post
Share on other sites
While both Soundminded and Gene make interesting points/observations about the 10 Pi/11's spectral balance, that is not the exact point I was raising.

I'm actually targeting my post to Z, HF, k, Speaker Dave to try to bring about a discussion/evaluation of the 'rightness' or 'wrongness' of the original AR design premise: the premise that said max dispersion was the correct approach in home loudspeakers. As Z has quite accurately pointed out, home loudspeakers haven't been designed that was for quite some time now. But can we use the 3a as an example of why or why not? I say no--the 3a is not the speaker with which that determination can be made.

My contention is that the 3a/5 have so many peripheral problems with non-flush drivers, intrusive cab molding, not truly acoustically-transparent grilles--and then on top of it all, non-ff-cooled tweeters that were far down in level--that the original AR premise of max dispersion cannot be given either a 'thumbs up' or a 'thumbs down' by eval'ing the 3a. Too many other of the 3a's problems get in the way.

I got your point, and I think you missed mine, which was that once the 10pi's crossover was set to compensate for spectral differences, there was no discernible difference in any other aspects of the two speakers' sound including those attributable to dispersion, such as "imaging," "soundstage," "airiness," etc. IOW, the 3a's supposed peripheral "problems" that the 10pi supposedly "solved" either weren't solved by the 10pi or were not really problems in the first place under reasonably normal listening conditions.

I don't know if anyone else here ever had the opportunity to hear classic and ADD versions of the same models side-by-side. My guess is that Ken and Tom probably did at one time or another. Maybe they can share their recollections as well for historical purposes.

Share this post


Link to post
Share on other sites
I got your point, and I think you missed mine, which was that once the 10pi's crossover was set to compensate for spectral differences, there was no discernible difference in any other aspects of the two speakers' sound including those attributable to dispersion, such as "imaging," "soundstage," "airiness," etc. IOW, the 3a's supposed peripheral "problems" that the 10pi supposedly "solved" either weren't solved by the 10pi or were not really problems in the first place under reasonably normal listening conditions.

I don't know if anyone else here ever had the opportunity to hear classic and ADD versions of the same models side-by-side. My guess is that Ken and Tom probably did at one time or another. Maybe they can share their recollections as well for historical purposes.

I was always convinced the AR3/AR3a approach was correct. My restored AR2axs with the same tweeter convince me more than ever now that the system balance is improved. The other problems Steve cited are IMO of minor importance and were either not objectionable or could have been overcome. A different grill cloth, increased tweeter output to overcome the grill, all were unimportant. So were the reflections off the molding which are nearly 90 degrees off the tweeter axis. IMO the three most important factors in speaker performace are frequency response, frequency response, and frequency response. That being said, frequency response is a far more complex issue than is commonly understood. An electrical signal is a scalar. A sound field is a three dimensional vector. Trying to make some simple comparison between the two, the vague generalizations about dispersion and room reflections is a gross oversimplification that obscures the facts about auditory perception and the engineering challenges. If there was one serious problem that plagued that design, it was trying to match a 12 inch AS woofer that reaches down below 40 hz with a 1 1/2" dome midrange driver. In AR3 the problem was the upper range of the woofer, in the AR3a it was the lower range of the midrange driver. AR9 which added an 8" LMR was the ultimate answer to that problem. Unfortunately AR9s tweeter does not have as much dispersion as AR3a and not nearly as much as AR LST. It's output still has an audible rolloff and the balance is still wrong with all controls set flat. The lack of high output capability also left AR3a's tweeter vulnerable to frequent burnout by people who overdrove it. Using 4 of them in LST was an improvement but not enough for some. They should have been fused. A higher crossover frequency and steeper slope would have helped too. Say a third order filter at 7 khz. Even with these problems, AR3 and AR10 pi were the only systems that ever successfully took on the challange of the LvR demo before an audience of more than a few even if those demos were highly contrived. Four decades later with products costing many times their price and with much improvement in available materials, they have yet to be successfully duplicated. Until I hear it for myself, my hunch is that with current design concepts they can't be.

Share this post


Link to post
Share on other sites
So were the reflections off the molding which are nearly 90 degrees off the tweeter axis. IMO the three most important factors in speaker performace are frequency response, frequency response, and frequency response.

I don't know how many times it has to be posted here, but the design objective of the AR3a tweeter was max dispersion, and the measurements on a flat baffle indicate the extent to which that was achieved. There is plenty of off-axis energy for reflections to play havoc with the direct field frequency response.

I'm actually targeting my post to Z, HF, k, Speaker Dave to try to bring about a discussion/evaluation of the 'rightness' or 'wrongness' of the original AR design premise: the premise that said max dispersion was the correct approach in home loudspeakers.

Toole SPECIFICALLY targeted that question fully 25 years ago; it was the very foundation of his work. Part 1 of that study explores the core theses of the design philosophies, and Part 2 tests them out via his statistical listener preference "end run."

Answer: The Villchur, Bose, Allison, et al. acoustic power approach is not supported by the data, and no researcher has published to the contrary since.... ;)

Share this post


Link to post
Share on other sites
There is plenty of off-axis energy for reflections to play havoc with the direct field frequency response....

Hi Z,

Is it, therefore, a fortunate degree of early reflection "havoc" that is appealing to 3a fans, or do they simply enjoy them in spite of the havoc ? ;)

Roy

Share this post


Link to post
Share on other sites
I don't know how many times it has to be posted here, but the design objective of the AR3a tweeter was max dispersion, and the measurements on a flat baffle indicate the extent to which that was achieved. There is plenty of off-axis energy for reflections to play havoc with the direct field frequency response.

Toole SPECIFICALLY targeted that question fully 25 years ago; it was the very foundation of his work. Part 1 of that study explores the core theses of the design philosophies, and Part 2 tests them out via his statistical listener preference "end run."

Answer: The Villchur, Bose, Allison, et al. acoustic power approach is not supported by the data, and no researcher has published to the contrary since.... ;)

"I don't know how many times it has to be posted here, but the design objective of the AR3a tweeter was max dispersion, and the measurements on a flat baffle indicate the extent to which that was achieved. There is plenty of off-axis energy for reflections to play havoc with the direct field frequency response."

With AR3a's 15khz signal down 5 db (70%) only 60 degrees off axis how much havoc could a reflector 90 degrees off axis play? How much havoc do the reflective surfaces near my Steinway play? How about the reflections off a steel music stand a foot away from a violin? Anyone who has listened to real music in an acoustically dead room compared to a live room knows why rooms with few reflections are called dead. If you spent more time studying how musical instruments make sound you might get a lot more insight into the problems of reproducing that sound. On the other hand if you are looking to reproduce the sound of a Peavy guitar speaker-amp, perhaps that's the ideal setup for you.

You wouldn't believe how much indirect high frequency energy is deliberately directed away from the listener and at reflective surfaces of my room in my re-engineered AR9s. It's over 95%. For the Steinway when the top is down...it's nearly 100%. Except for diffraction (the real thing, not just reflections someone mistakenly calls diffraction) virtually all of its sound is indirect, that is reflected at least once between the point where it is generated and a listener's ears if he is seated. That is one of the elements that makes it sound so magnificent and why the AR9s can't duplicate it. Everything my AR9s produce from above 200 hz to 6 khz is directed right at the listener. That's the source of where its worst distortion lies. Still it's far better now than the way it sounded when it was used as it came from the factory.

Share this post


Link to post
Share on other sites
Is it, therefore, a fortunate degree of early reflection "havoc" that is appealing to 3a fans, or do they simply enjoy them in spite of the havoc ? ;)

Do we suspect that is also the reason the LvR demos were successful? :lol:

Personally, I feel it's fortunate that AR3a did not achieve its max-dispersion, flat power response design objective. To what extent these "compromises" play a role in that is as yet indeterminate.

SpitWad is done. Here is the directivity data out to 90° (hemispherical beamwidth); I'll get the same on AR3a before moving on. Don't know if anybody's getting it, but at this resolution, 5° increments, it's 37 measurements compiled into each of these plots:

post-102716-1276455165.jpg

post-102716-1276455211.jpg

post-102716-1276456167.jpg

post-102716-1276457492.jpg

Share this post


Link to post
Share on other sites
You wouldn't believe how much indirect high frequency energy is deliberately directed away from the listener and at reflective surfaces of my room in my re-engineered AR9s. It's over 95%. For the Steinway when the top is down...it's nearly 100%. Except for diffraction (the real thing, not just reflections someone mistakenly calls diffraction) virtually all of its sound is indirect, that is reflected at least once between the point where it is generated and a listener's ears if he is seated. That is one of the elements that makes it sound so magnificent and why the AR9s can't duplicate it.

COOL!

Bose got it right, then.

[Who'd have guessed...? ;) ]

Share this post


Link to post
Share on other sites
"Amply demonstrated" how? Measurements aside, did you set the things up in a good listening room and do level-matched (with good-recordings) A/B comparisons between them and some good mainstream wider-dispersion models?

There ARE no mainstream wider-dispersion models.

That be your gig.

[Alas, you have retired from it.... ;) ]

Share this post


Link to post
Share on other sites
COOL!

Bose got it right, then.

[Who'd have guessed...? ;) ]

Bose got a lot of things right...often for reasons different than he said. He also got some things very wrong. One thing he got wrong was that a 4 inch midwoofer makes an awful tweeter. IMO 901 does not produce anything close to acceptable high frequencies either in level or quality. What little it produces beams directly forward, an inevitable result of a 4" cone. Bose himself said he had a strong aversion to the shrillness of other speakers, especially in the way they distorted the sound of violins. Like poorly reproduced bass, take your pick, bad distortion at some frequencies or no output at all.

His original speaker design and Series II can produce the deepest bass with very low distortion but it requires enormous power and additional equalization to do it. It also requires being used in multiples. Within its power handling capabilities it noses out JBL Paragon.

Bose 901 cannot come even remotely close to reproducing the reverberant sound field of a concert hall. In that regard it is no better than any other speaker. But it can reproduce sound fields a lot closer to the way instruments would be heard in the listener's room.

Who would have guessed? His accountants. While there are considerable differences between the original execution of the design and the current version, due to their commonality and uniqueness if taken together they are likely the longest selling (42 years) single audio product, probably the most units sold and the most profits on any one model in the history of the industry. This fact demonstrates that whatever it is doing right is very important to many people who will overlook its shortcomings. And worst of all, it flies in the face of every idea its competitors and audiophiles hold to be absolutely true. Perhaps that is the reason they hate him so much. That alone is enough for me to admire him. A one billion dollar a year privately owned company that started with nothing and built originally around a single product. Yes he did at least something very right.

Share this post


Link to post
Share on other sites
If edge-diffraction effects were important the Villchur LvR demos would have been flops. They were not, as numerous audio- and music-magazine reviews clearly pointed out.

Howard Ferstler

You love to fall back on this as if satisfying an audience in a well conducted LVR demo is proof that speaker perfection had been achieved and further technical improvements are empty advances. Of course we keep reminding you of the many successful LVR demonstrations over the decades from 1916 on yet apparently none of those count, only the single one you want to give credence to. Not exactly an unbiased view.

I think Steve's question was about AR design philosophy and speaker dispersion. First note that as an industry insider I'll tell you that engineering and marketing are always entertwined. What the engineers can achieve the marketers will want to tout as monumental advances. After acoustic suspension EV developed the dome mid and tweeter and since they measured quite well this became part of AR marketing and mantra. They were able to show measurements of drivers that had ideal performance under a particular condition: when measured individually on large smooth baffles they looked great. In fact they might have measured better that way than any other manufacturers drivers at the time, its hard to say since others didn't rise to the challenge (well there were some ADC models that copied the ad approach).

To AR's credit, when they did the AES papers, they showed the whole system and we can see that there were technical problems with driver blending, edge reflections (both internal reflections and defraction related), tweeter sensitivity and woofer response in 4pi.

My question is can we assume that since AR achieved wide dispersion, and advertised wide dispersion, do we really know that they believed it was an essential parameter?

Regarding power response, my feeling is that they found it secondary to individual driver frequency response. Clearly the drivers and networks were optimized to be totally direct-field-flat. That these individual curves were so clearly optimized implies it was their primary objective. The power response would follow on based on driver directivity and individual driver sensitivities. The first dome tweeter was inherently heavy and would never match the sensitivity of the woofer or mid. As such power response would have to roll off. You can show a power curve and compare it to a concert hall RT curve and say "Its okay, Beranek approves", but comparing radiated power response and concert hall curves is tenuous science. Wouldn't the recording have been made in the concert hall with reduced treble reverberation? Why would we want to roll it off again when playing through the speaker? Also, no one ever points out that the power curve they compare to a concert hall is measured with mid and tweeter levels at max, a position that AR advises against.

I tried to calculate the directivity index curve of the AR3a a few months ago. You should be able to subtract the frequency response curve from the power response curve, as published in the various papers, and get the measure of directivity. I scanned and digitized curves and did the subtraction but the results made no sense. I would guess that the curves weren't accurate enough or came from somewhat different samples. It was apparent, though, that directivity was fairly constant while frequency response and power response both rolled off at HF. In the end the power response of an AR3a is similar to any other speaker, but it is achieved by falling frequency response rather than by rising directivity. This leads to its known, relatively dull balance, but is probably better than achieving flat power response and sounding way too bright.

David

Share this post


Link to post
Share on other sites

These various responses are mostly concerned with either the spectral balance preferences of the 3/3a/11/LST or a re-hash of why the 3a's edge molding meant that it wasn't a true 'max dispersion' speaker or why some don't think it actually mattered.

I'm suggesting that the 11 be eval'd as a relatively 'pure' representative of the Classic AR design approach. It corrects all the faults that Z has pointed out about the 3a on many occasions and I don't need to cite the 11's attributes vs. the 3a for a 3rd time here.

From a purely scientific standpoint, a valid experiment is possible only when the extraneous variables are removed, so that one can be sure the experiment is going to reveal the data about the subject you're really interested in.

The experiment we want to conduct is:

Market verdict aside (because there are many factors that determine market success, some having to do with listener preference/spkr performance, some having nothing whatsoever to do with those factors), the premise we want to examine is:

Was AR's max dispersion approach valid or not?

Forget what Toole and the "market" have said. We want to examine it. The Forum. Let's decide for ourselves, from a clean sheet of paper, so to speak.

Z has illuminated the extraneous variables of the 3/3a which make those speakers unsuitable test subjects.

The 11 is a 'clean' test subject.

Get a pair of good 11's, compare them in every way imaginable to your standard-bearers of today, and report back to the Forum as a whole.

Enough time-wasting pontification in the abstract.

Conduct a scientifically-valid experiment with a worthy test subject, and report the findings.

Steve F.

Share this post


Link to post
Share on other sites
You love to fall back on this as if satisfying an audience in a well conducted LVR demo is proof that speaker perfection had been achieved and further technical improvements are empty advances. Of course we keep reminding you of the many successful LVR demonstrations over the decades from 1916 on yet apparently none of those count, only the single one you want to give credence to. Not exactly an unbiased view.

As we discussed earlier, we can reasonably conclude from the contemporary accounts that each of the historical LvR demos "proved" that the technology being presented was able to deceive an audience with the discriminating abilities of its time, and that each succeeding LvR demo "proved" that audience discriminating abilities had improved since the previous demo to the point where the technology that achieved the same level of deception represented a "technical improvement."

What is missing is some test or demo "proving" that a contemporary audience now requires further technical improvements to achieve the same result, and in the absence of such proof there is no reason to assume that any improvements made since the most recent LvR have done anything to advance the goal of achieving that same result. The fact that nobody in today's audio industry is pursuing that goal or even seems to give a damn whether it can be achieved doesn't help much, either.

Have you seen any indication that the management at Simex-Iwerks has any interest in producing a convincing "you are there" simulation experience, or are they satisfied with "wow, wasn't that a great ride?"

Share this post


Link to post
Share on other sites
What is missing is some test or demo "proving" that a contemporary audience now requires further technical improvements to achieve the same result, and in the absence of such proof there is no reason to assume that any improvements made since the most recent LvR have done anything to advance the goal of achieving that same result.

Nor reason to assume otherwise, either, other than the untenable supposition that the relentless forward march of improvements in loudspeaker quality and listening discernment capabilities of the public stopped dead in its tracks in the 70's. ;)

The midrange behavior of your vintage model might still be nearly as good as new (not too far off, at least) but your tweeter measurements indicate that the tweeter in your unit might be in trouble. Remember, those here who have fooled a lot with that tweeter have indicated that virtually all AR-3a tweeters will have deteriorated.

So, now, being familiar with how acoustic instruments sound in concert hall live performance isn't good enough, I must also experience this from the best seats in the house, the best house, preferably, which only Allisons can deliver, presumably.

You make a good case for AR3a's generally sucking, but let's wait for the actual data, at which point you can make up some rationale for the normalized directivity having deteriorated over the intervening years.

[Remember also that I have some Allisons here, too.... :lol: ]

Share this post


Link to post
Share on other sites
You love to fall back on this as if satisfying an audience in a well conducted LVR demo is proof that speaker perfection had been achieved and further technical improvements are empty advances. Of course we keep reminding you of the many successful LVR demonstrations over the decades from 1916 on yet apparently none of those count, only the single one you want to give credence to. Not exactly an unbiased view.

I think Steve's question was about AR design philosophy and speaker dispersion. First note that as an industry insider I'll tell you that engineering and marketing are always entertwined. What the engineers can achieve the marketers will want to tout as monumental advances. After acoustic suspension EV developed the dome mid and tweeter and since they measured quite well this became part of AR marketing and mantra. They were able to show measurements of drivers that had ideal performance under a particular condition: when measured individually on large smooth baffles they looked great. In fact they might have measured better that way than any other manufacturers drivers at the time, its hard to say since others didn't rise to the challenge (well there were some ADC models that copied the ad approach).

To AR's credit, when they did the AES papers, they showed the whole system and we can see that there were technical problems with driver blending, edge reflections (both internal reflections and defraction related), tweeter sensitivity and woofer response in 4pi.

My question is can we assume that since AR achieved wide dispersion, and advertised wide dispersion, do we really know that they believed it was an essential parameter?

Regarding power response, my feeling is that they found it secondary to individual driver frequency response. Clearly the drivers and networks were optimized to be totally direct-field-flat. That these individual curves were so clearly optimized implies it was their primary objective. The power response would follow on based on driver directivity and individual driver sensitivities. The first dome tweeter was inherently heavy and would never match the sensitivity of the woofer or mid. As such power response would have to roll off. You can show a power curve and compare it to a concert hall RT curve and say "Its okay, Beranek approves", but comparing radiated power response and concert hall curves is tenuous science. Wouldn't the recording have been made in the concert hall with reduced treble reverberation? Why would we want to roll it off again when playing through the speaker? Also, no one ever points out that the power curve they compare to a concert hall is measured with mid and tweeter levels at max, a position that AR advises against.

I tried to calculate the directivity index curve of the AR3a a few months ago. You should be able to subtract the frequency response curve from the power response curve, as published in the various papers, and get the measure of directivity. I scanned and digitized curves and did the subtraction but the results made no sense. I would guess that the curves weren't accurate enough or came from somewhat different samples. It was apparent, though, that directivity was fairly constant while frequency response and power response both rolled off at HF. In the end the power response of an AR3a is similar to any other speaker, but it is achieved by falling frequency response rather than by rising directivity. This leads to its known, relatively dull balance, but is probably better than achieving flat power response and sounding way too bright.

David

"I think Steve's question was about AR design philosophy and speaker dispersion. First note that as an industry insider I'll tell you that engineering and marketing are always entertwined."

Yes but it depends who the market is. The market for a Corvair was very different from the market for a Corvette. Chevy can try to impart a psychological connection between the two by inventing a similar name but that's where the similarity stops.

"Of course we keep reminding you of the many successful LVR demonstrations over the decades from 1916 on yet apparently none of those count, only the single one you want to give credence to. Not exactly an unbiased view."

I attended two of those AR demos and before hearing the first one I was biased...against AR. I had no explanation as to why the LvR tests were so different from what I heard in dealer showrooms and elsewhere with AR3 and later 3a until I learned here a few years ago that the preamp treble control had been boosted. And now that the microphone placements had to be experimented with to get acceptable results. I've also said that my impression at the time was that the sound from the speakers was not a dead ringer for the real thing, merely surprisingly close.

"My question is can we assume that since AR achieved wide dispersion, and advertised wide dispersion, do we really know that they believed it was an essential parameter?"

It appears that under the direction of Villchur and Allison they at least did. Later on there may have been a different assessment due either to technical or marketing reasons.

"As such power response would have to roll off. You can show a power curve and compare it to a concert hall RT curve and say "Its okay, Beranek approves", but comparing radiated power response and concert hall curves is tenuous science."

No it is junk science. It is a comparison between a steady state measurement and a purely transient phenomenon. If Beranek approves...he'd have been wrong too.

"Wouldn't the recording have been made in the concert hall with reduced treble reverberation? Why would we want to roll it off again when playing through the speaker?'

Recordings are made with microphones near the sources of sound. The amount of reverberation on a recording, even one that sounds relatively reverberant coming from a pair of louspeakers is infinitesmal compared to the reverberation heard in the audience at a live performace. The spectral balance on a recording is closer to the source than an average in the hall. But whether too bright or dull, to make the recorded/reproduced instruments sound like the real thing, they need to have a relative spectral content similar to the way they'd sound close up. The current paradigm makes up for the 90% of the reverberant sound that is missing by playing the sound that is there ten times louder. It doesn't work. The current methods cannot duplicate the tone (timbre) of musical instruments as they are heard at live performaces.

"Also, no one ever points out that the power curve they compare to a concert hall is measured with mid and tweeter levels at max, a position that AR advises against."

There was a lot more deception in their advertising than seemed evident at the time. At its flattest with the tweeter control at max and the midrange control below midway without additional equalization AR2ax still sounds muffled on all recorded material. Not until the treble is brought up to compensate for its rolloff does the speaker sound accurate. When properly adjusted it can sound very accurate. Only after restoring a pair last fall did this ever occur to me. I have to wonder what other "adjustments" Allison had to make to make AR3 sound as accurate as it did.

"In the end the power response of an AR3a is similar to any other speaker, but it is achieved by falling frequency response rather than by rising directivity. This leads to its known, relatively dull balance, but is probably better than achieving flat power response and sounding way too bright."

AR3a achieves neither flat frequency response nor flat power output. Even taking into account the tweeter's mismatched sensitivity its total power radiation as a funcion of frequency rolls off as frequency increases, just not to the same degree as more directional tweeters. Whether on axis response falls off too is something I'm not quite sure of yet but I suspect it does. In 1967 the only remedy was a treble control. Today there are other more effective means. The design is salvagable with effort. The more directional single tweeter designs are not. Snell AII created an interesting form of distortion trying as I pointed out in another thread.

Share this post


Link to post
Share on other sites
Nor reason to assume otherwise, either, other than the untenable supposition that the relentless forward march of improvements and listening discernment abilities of the public stopped dead in its tracks in the 70's.... ;)

There is every reason to believe that the discernment abilities of the public have not stopped dead, but that just leads to a discussion of whether they've improved or eroded. In the absence of subsequent attempts to test, the only assumption that is reasonably safe to make is that interest in knowing is very low.

As for how this applies to individual purchasing decisions, today's consumer has lots of objective guidance as to what choices will best produce an "accurate" playback of what's on a source recording, but none as to how to achieve a convincing "you are there" experience, and no shopping options other than trial and error.

Share this post


Link to post
Share on other sites
I attended two of those AR demos and before hearing the first one I was biased...against AR. I had no explanation as to why the LvR tests were so different from what I heard in dealer showrooms and elsewhere with AR3 and later 3a until I learned here a few years ago that the preamp treble control had been boosted. And now that the microphone placements had to be experimented with to get acceptable results. I've also said that my impression at the time was that the sound from the speakers was not a dead ringer for the real thing, merely surprisingly close.

We know better, now, no?

There was a lot more deception in their advertising than seemed evident at the time.

"Disingenuous" is the word I used, and "Promotional stunt." "Trick" and "Fooled" are now in play; it seems only Howard still ascribes probative significance to the LvR "entertainment" events.

[it's all that's left in his arsenal, alas. :lol: ]

Admittedly, some modern horn designs do not have the saw-toothed artifact, and that certainly is a good thing. Good for them.
Certain horn designs may do similar things, and if that is the case, good for them.

Horn/waveguide enthusiasts everywhere certainly appreciate this hard-won concession on the part of Howard Ferstler.... ;)

Share this post


Link to post
Share on other sites
There is every reason to believe that the discernment abilities of the public have not stopped dead, but that just leads to a discussion of whether they've improved or eroded. In the absence of subsequent attempts to test, the only assumption that is reasonably safe to make is that interest in knowing is very low.

As for how this applies to individual purchasing decisions, today's consumer has lots of objective guidance as to what choices will best produce an "accurate" playback of what's on a source recording, but none as to how to achieve a convincing "you are there" experience, and no shopping options other than trial and error.

As I have said many times, Bose sells in a couple of weeks what AR sold in its best years. Microsoft, Apple and Sony spend more on seasonal ad campaigns than any pure speaker company has ever sold in a year. "Hifi," and for that matter, "classical music," has always been a fringe thing. I love obscure artists and visionary scientists, but let's not confuse them with commerce.

Acoustically, the corner may be one's friend. Rhetorically, it should probably be avoided.

-k

http://en.wikipedia.org/wiki/List_of_fallacies

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.
Sign in to follow this  

×
×
  • Create New...