Some AR2ax measurements

[speaker dave]

Hi Guys,

From some other posts you may know that I have been restoring a pair of AR2ax's that I inherited from my dad. Last year I redid the cabinets and grilles. Recently I finished the project after new woofer surrounds and some heavy cleaning of the mid pots. I bypassed the tweeter pots at full level.

Once they were up and running I gave them a good listen. First impressions are of a good smooth wide band design. Bass is nice and extended with an overall balance leaning to warm rather than "forward". After some extended listening I did notice a slight mid range prominence and a dullness for the range just above that, the low treble.

So, out came the laptop RTA and omni mic for some pink noise curves which I've attached. PLEASE DON'T TAKE THE FOLLOWING AS CRITICISM OF OUR BELOVED VINTAGE ARs but more as an exploration of what truly good performance was with a speaker of that vintage.

Speakers were in the typical orrientation with long axis vertical and woofer below. Pink noise at a medium level was measured at about 0.5 meters out from the grille (on) at a variety of positions. Most curves have the mids set to my prefered setting of about 1/3 from center dot to max. "Horizontal curves" shows a number of positions swinging from left to right in the plane of the mid and tweet. The green curve is on the best axis roughly centered between the mid and the tweeter or slightly toward the mid. The blue curve shows significant interference from a position just left of the tweeter. The mid and the tweeter overlap considerably so there is a lot of variation from 4kHz on up. The variation around 2k is the midrange and reflections from the box edges. (Neither the woofer or tweeter are contributing at those frequencies.)

"Vertical curve" shows a range of angles up and down (about +-30 degrees) The green curve is on-axis roughly on the center of the box. It looks pretty good with just a slight prominence around 400 Hz. The gold curve is in the upward direction and shows a crossover hole at 1kHz. If you can see the top of the cabinet, this is the response you would hear! So these speakers should be on a tall stand, a stand with considerable uptilt, or placed horizontally. (Hey you might even put them on a bookshelf.) The down curve (blue) starts to show some of the same depression at 1500 Hz.

"Midrange control" shows the total range of the mid pot. Full up is 5 to 6dB hotter than the setting I prefered. Even though it seems flatter it is way too aggressive up at max. Full down shows the woofer and tweeter curves, as the mid is essentially off. Note that the tweeter is quite a bit lower in level (even with its pot at max) and that the mid is more of a wide range tweeter (never rolls off) while the 3/4" dome is a "super tweeter".

"Tweeter only" shows curves with the mid down and the woofer strap removed. I took a few curves at slightly different positions which show cabinet effect. Mounting the tweeter in a circular 45 degree recess give a pretty strong reflection that leads to the comb filtering. A later AR feature the "acoustic blanket" would help here, but the difference might not be audible since the tweeter's level is still under the mid's level.

Again, although more modern speakers, from AR and others, would tend to have less response variation, you can see that the overall balance and the bass extension is pretty good for a speaker with heritage back to the late 50s!

Regards,

David

[genek]

So these speakers should be on a tall stand, a stand with considerable uptilt, or placed horizontally. (Hey you might even put them on a bookshelf.) The down curve (blue) starts to show some of the same depression at 1500 Hz.

There's a reason why they were called "bookshelf speakers." An earlier version of the 2 series actually came with one long face unfinished because it was supposed to be the bottom of the speaker.

[Steve F]

Dave,

Great info. The general shape and slope of your curves is not dissimilar to the "front hemispherical" curve that CBS labs measured in their test of the 2ax for High Fidelity magazine in (I think) Nov 1971.

Your characterization of how it sounds is accurate ("warm...not forward....extended bass") and I think it sums up what it was about the 2ax that made it such a satisfying speaker to live with, day-in and day-out. It was my first 'good' speaker (scrimped and saved and bought them when I was a senior in HS), and I remember them fondly.

Interestingly enough, I set my controls almost exactly how you did yours--tweet full up, mid partway between 'Norm' and 'Max.'

I never tired of them, and I only replaced them 8 years later when I found a mint set of used LST-2's.

Steve F.

[Zilch]

Hello, David.

Pull the mic back to ~1 M, and you may find the drivers integrate better.

Also, try nearfield on the woofer, ~1/4" out from the center of the dome, with the SPL attenuated, as appropriate for being that close.

Since that takes the influence of the room out of the measurement, it'll provide an accurate indication of the bass response, and how the speaker will behave in 2-Pi, alignment, i.e., one boundary.

Good up to ~200 Hz, very likely, it'll reveal where the peak is, and how "boomy," as well as the cutoff at the low end, -6 and -10 dB being definitive, though it's often difficult to decide where to call 0 dB as actually occurring as reference....

[onplane]

"Midrange control" shows the total range of the mid pot. Full up is 5 to 6dB hotter than the setting I prefered. Even though it seems flatter it is way too aggressive up at max. Full down shows the woofer and tweeter curves, as the mid is essentially off. Note that the tweeter is quite a bit lower in level (even with its pot at max) and that the mid is more of a wide range tweeter (never rolls off) while the 3/4" dome is a "super tweeter".

Regards,

David

Dave, thanks ever so much for sharing your work.

Your observations on the tweeter output match what I hear. If you happen to have an old low or mid power amp/receiver (I have several in my basement), you could try bi-amping.

The idea here is to cut the mid back to the 50% position OR less. Now, this puts the mid way behind the woofer and about even with the tweeter. Just doing this would create one heck of a bass heavy speaker (a kind of super New England sound).

With bi-amping, however, you simply bring up the mid (via the volume control on the mid/tweeter amp) to bring it back in balance with the woofer. The tweeter, however, now sees more voltage and it will produce more sound. So instead of being "quite a bit lower", it will be more in balance with the mid and should take some of the edge off the pronounced mid frequencies.

This is what I do with my 3a's, see:

Coaxing more SPL out of AR tweeters

Again, to bi-amp all you need is two solid state, common ground amps (99% of solid state is common ground). You use the three terminals provided by AR.

Just let me know, Dave, and I'll send you directions on how to test and wire the bi-amp.

Regards,

Jerry

[Zilch]

"Vertical curve" shows a range of angles up and down (about +-30 degrees) The green curve is on-axis roughly on the center of the box. It looks pretty good with just a slight prominence around 400 Hz. The gold curve is in the upward direction and shows a crossover hole at 1kHz. If you can see the top of the cabinet, this is the response you would hear! So these speakers should be on a tall stand, a stand with considerable uptilt, or placed horizontally. (Hey you might even put them on a bookshelf.) The down curve (blue) starts to show some of the same depression at 1500 Hz.

+/- 30° would define an extreme window in the vertical dispersion. Compute the angle from the listening position, and I believe you'll find the difference between sitting down and standing up is more like +25° from 6 feet back, and less from further away. You're showing "worst case."

RTA measures in-room response with reflections summed, and thus, generally, a downward-sloping response curve which would be flatter if measured anechoically, or quasi-anechoically with the reflections removed (see attachment below).

I'm saying these measure well, consistent with accepted standards for "ideal" in-room frequency response, though as I mentioned above, 0.5M is too close. See Toole, Fig. 18.1. 1M is itself a compromise, and 2M is better, but the greater the distance, the more room reflections interfere with the measurement, and the objective, presumably, is to characterize the speaker, not the room....

[speaker dave]

I'm glad this has gotten a lively discussion going. Let me respond to some of your comments.

Genek: Yes we forget that speakers used to be placed on bookshelves (or on the floor), before audiophiles had us putting them on stands well away from the walls. When Allison did his tests of 3as in the Boston area, every speaker tested was touching at least one boundary.

Steve: And do you remember your first girlfriend that well or that fondly?

Zilch: I took a dustcap measurement of the woofer per your suggestion. (Attached) Nice flat response down to 50Hz and -6dB around 40Hz. Thats pretty good for the younger brother of the AR3. One caveat about nearfield curves is that they represent true 2 pi mounting (buried flush in the ground or a wall) and as such can't show box diffraction effects. When I was at JBL years ago we got in trouble with the L212 that was designed to be near perfect buried flush in the ground. when measured by High Fidelity in an anechoic chamber, the bass drooped and a midrange mound developed, just as the 4 pi to 2 pi difference would have predicted. For a speaker near a wall, below 200 as you suggest, I think the curve is representative.

Jerry: Thats an interesting idea about biamping and getting a better level match mid to tweeter. I am sure that would work, although if I went that far I would want to get a more classic crossover between mid and tweeter. With the tweeter at the current level (well under the mids level) it tends to reduce interference effects. A little inductance on the mid wouldn't hurt, but where do you stop in redesigning?

Howard: I like that approach to getting an area integrated curve. I've done a similar technique to map out the listening window average. Cudos for using stereo noise also. A lot of people overlook that fine point and don't realize they are measuring interference effects between 2 speakers.

Regarding the measuring distance, I agree that 0.5 meter is way too close to see a fully integrated response. You can even see in the 0.5m vertical curves how one measurement is nearer the tweeter end of the cabinet so bass response drops. At the other end the bass is back up and the HF level drops. I just wanted better resolution of the direct sound, with less interference from the room, because I was exploring the individual contributions of the drivers and the effectiveness of the crossover network.

Attached are the vertical curves re-done at 2 meter distance. As expected, the curve is rougher, expecially for lower frequencies. The HF end droops more both due to falling power response and greater room absorption as you go up the frequency range. But the crossover effect is still apparent at 2m (see curve). I checked the angles of max crossover dip a little more carefully this time and they were very close to + and - 35 degrees. This is a very wide angle between nulls and surprisingly symmetrical in the time before Linkwitz/Riley crossover thinking, so no room for criticism on this regard.

Finally, to toss out a little heresy in the AR forum: The notion of measuring the drivers individually, ignoring cabinet diffraction edge/grille reflections or crossover interference effects in the direct sound, is very old school. Even AR evolved well away from this philosophy after the LST. Maybe it was the influence of the UK design group but the AR3a improved was clearly an effort to design a crosover with better free field performance. Ditto the acoustic blanket, which would clean up the direct response without changing the power response. The dome mid and tweeter on a shared faceplate (AR 90?) was to improve the vertical polars, again with no effect on power response. That is one of the major trends of modern speaker design, to find ways to better integrate driver response and reduce the effects of the cabinet on the direct response of the drivers.

(Let the flames begin!)

David

[genek]

Finally, to toss out a little heresy in the AR forum: The notion of measuring the drivers individually, ignoring cabinet diffraction edge/grille reflections or crossover interference effects in the direct sound, is very old school.

Well, most of the measurement curves of individual drivers are old documents. It's not as if Edgar Villchur has been spending any time since he left AR in 1967 remeasuring his old speakers...

[Steve F]

No flames needed or even contemplated. This is a very interesting discussion.

First, I do remember my first girlfriend, but not as fondly as my 2ax's. I mean here it is, 37 years later, and I'm still talking about the 2's, not her.....

The 'Dual-Dome' arrangement was intro'd on the AR-9LS. That was a major advancement over the original 9. They also went to a 10" internal/12" direct woofer arrangement, instead of the 9's two side-firing 12's.

There was a technical paper on the 3a that showed a full system on-axis FR curve, not merely the individual drivers, but that doc was not as widely distributed as the individual drivers doc.

My personal opinion is that since AR really believed (at that time) that far-field reverberant power response dominated the listener's perception of the speaker and that they knew a closer-distance full system FR measurement would not be "pretty" or explainable, they made the conscious decision not to publish conventional full-system FR curves.

But neither Bose nor Advent did either (although probably for different reasons).

Steve F.

[Zilch]

(Let the flames begin!)

The flames have died down somewhat. Many believe it was me who was snuffed:

http://www.classicspeakerpages.net/IP.Boar...?showtopic=5243

Howard wants to rekindle them and perfect the immolation:

http://www.classicspeakerpages.net/IP.Boar...?showtopic=5267

But neither Bose nor Advent did either (although probably for different reasons).

The same reasons, actually, based upon an erroneous belief that the reverberant field is dominant in typical home listening spaces.

I have posted the AR3a system response curves as presented by Allison (1970) here:

http://www.classicspeakerpages.net/IP.Boar...ost&p=78007

[Zilch]

It is with good speakers. Your problem is that you have discovered that speakers with narrow dispersion manage to have the direct field dominate, provided the listening distances are not too great and the room is padded down enough for most of the remaining off-axis energy to be absorbed.

I have demonstrated that my "narrow dispersion" speakers have wider dispersion than AR3as, and do not corrupt the off-axis response in achieving it.

Unfortunately, the result is stellar clarity and a non-live-music kind of sound. If that works for you, fine. But do not say that your approach is the only one that matters, particularly when, by definition, the goal of high-fidelity sound reproduction is to simulate live music in the best way possible.

The goal of high-fidelity sound reproduction is to faithfully reproduce the program content, in which context generating an artificial soundstage, while pleasing to many listeners, clearly represents an adulteration of the prime objective. See Bose.

I continue to wonder why you are writing on this site at all, since the design rationale for the vast bulk of the speakers that are featured, not to mention the sound-reproduction tastes of most of the people who participate, are at odds with your entire approach to audio.

It'd be good if you got past your biases, Howard; the purposes of this site, and the objectives of its membership, are best served by the sharing of information and knowledge. You've as much as told Speaker Dave that his measurements, and by implication, this thread, are all but worthless. I'm sorry, but, contrary to your opinion and assertions, he IS able to measure the vertical off-axis response in these speakers, and locate the nulls therein. It is also regrettable that none of us have ears at both ends of the couch, as you further suppose.

Are you selling something, per chance?

The truth, primarily, and it's clear you'd prefer that those of us in pursuit of it here went away....

[Steve F]

"QUOTE (Steve F @ Apr 3 2009, 09:42 AM)

But neither Bose nor Advent did either (although probably for different reasons).

The same reasons, actually, based upon an erroneous belief that the reverberant field is dominant in typical home listening spaces."

I was trying to be low-key in my answer and sort of soft-pedal the truth about Advent and Bose. I know the head designers at both companies quite well, both professionally and personally. In Advent's case, they took a completely opposite view to AR. They believed that first-arrival, on-axis FR was the prime determinant of speaker quality and listener perception. The original Large Advent was voiced according to Kloss' view of the proper on-axis 'octave-to-octave' balance. This may have led the Advent's on-axis response to be somewhat less than great-looking, so they didn't publish the curves--as you can see, for a completely different reason than why AR didn't publish theirs. AR was because they favored far-field power and placed less importance on near-field on-axis curves; Advent was because although they DID favor on-axis, near-field curves, their curves were not necessarily terrific-looking, because of their O-T-O philosophy. Same result--not published--but for completely different reasons. I know this first-hand--it's not open to question.

Bose just didn't trust that the magazines and reviewers could properly interpret their D/R work at Bose and make sense out of it with conventional FR curves. So rather than set themselves up for criticism by publishing curves which then were not confirmed by the mags, the just didn't publish anything. Additionally, Bose endeavored to stay 'above the fray' of audiophile mania and market their products not to the 5% audiophile slice, but to the 95% casual user segment who just wanted good sound and a design philosophy they (the customer) could get their heads around and buy into--both literally and figuratively. Again, it's completely different reasoning than AR (who courted the engineer/audiophile very hard), but with the same result: system FR curves not published. I know this first-hand--it's not open to question.

You can agree or disagree with whether or not these three companies SHOULD have published complete system FR curves or not, but these--factually--are the reasons why they DIDN'T. And they're all different reasons.

Steve F.

[Zilch]

You can agree or disagree with whether or not these three companies SHOULD have published complete system FR curves or not, but these--factually--are the reasons why they DIDN'T. And they're all different reasons.

Is the Bose "White Paper" available? Clearly, Holt's synthesis of it reads as if Allison wrote the original:

http://www.stereophile.com/historical/425/index1.html

Holt provides a more modern analysis of how the 901 does what it does which is further refined in Toole.

I never analyzed Kloss's "Octave to octave" design approach, nor have I seen it definitively documented.

Bottom line, pragmatically speaking, while they may each have in detail had a different rationale, they did not publish system frequency response data for the same more fundamentally good reason -- it was "problematic," and would not stand up under objective scrutiny....

[soundminded]

It looks like someone couldn't make up his mind;

http://www.audioheritage.org/html/profiles/jbl/l100.htm

Kinda like a juggling act. Or three card monty

According to Sam's Audio Engineering Reference, there really was a West Coast Sound, it wasn't a myth. It seems to have descended from the motion picture industry. For some people, an electronically amplified rock or pop concert is their idea of live music. Not my idea but....for many this was also a landmark speaker. One of them was anyway.

[Zilch]

For some people, an electronically amplified rock or pop concert is their idea of live music. Not my idea but....for many this was also a landmark speaker. One of them was anyway.

Yeah, "Fake but FUN."

It was a long, hard road moving L100/4311B enthusiasts past getting all bent thinking I was bashing them and into understanding just what it is they like.

Same with Altec 604 fans, when informed that L100s were designed to mimic their inaccuracies.

And VOTT and its derivatives doing the X-curve not translating well for playing music at home.

And....

[genek]

I don't know what those other folks thought about their speakers or where their ideas came from, but there's more than enough AR reference materials on CSP to relieve anyone who cares to read it of the misconception that AR speakers were "flat," that they were intended to be that way or that their designers even thought they should be. The references, especially those that Edgar Villchur authored and put his byline on, make it fairly evident that the speakers were intended to deliver what Villchur believed to be realistic, uncolored sound reproduction at his ears, and that whatever measurements he made in the process were intended by him to be predictors of whether the speakers he was producing would achieve that goal, not to demonstrate to anyone that they met anyone's "standard" for "good speakers" but his own. If you look at the pre-Teledyne materials, driver curves are used almost exclusively in discussions of quality control or measurement methodologies (of course, all that changed after Villchur left and Teledyne took over; they plastered those driver curves on just about everything they printed up.)

Debates about whether the speakers measure "flat" at 0.5M, 1.0M or any other distance are pretty much a waste of energy, because we alredy know they never did and weren't intended to. If you're listening to classic-era AR speakers and your listening preferences are not the same as Villchur's were at the time they were made (and from his most recent interview, it doesn't appear that his have changed since the old days), you're just listening to the wrong speakers.

[soundminded]

I don't know what those other folks thought about their speakers or where their ideas came from, but there's more than enough AR reference materials on CSP to relieve anyone who cares to read it of the misconception that AR speakers were "flat," that they were intended to be that way or that their designers even thought they should be. The references, especially those that Edgar Villchur authored and put his byline on, make it fairly evident that the speakers were intended to deliver what Villchur believed to be realistic, uncolored sound reproduction at his ears, and that whatever measurements he made in the process were intended by him to be predictors of whether the speakers he was producing would achieve that goal, not to demonstrate to anyone that they met anyone's "standard" for "good speakers" but his own. If you look at the pre-Teledyne materials, driver curves are used almost exclusively in discussions of quality control or measurement methodologies (of course, all that changed after Villchur left and Teledyne took over; they plastered those driver curves on just about everything they printed up.)

Debates about whether the speakers measure "flat" at 0.5M, 1.0M or any other distance are pretty much a waste of energy, because we alredy know they never did and weren't intended to. If you're listening to classic-era AR speakers and your listening preferences are not the same as Villchur's were at the time they were made (and from his most recent interview, it doesn't appear that his have changed since the old days), you're just listening to the wrong speakers.

For a long time I thought that "East Coast sound" and "West Coast sound" was a myth until I read it in Sam's Audio Engineering Reference Handbook. The West Coast sound seems to have evolved from the motion picture industry. It started when wealthy people acquired theater speakers for large homes. Manufacturers like JBL and Altec saw a niche market they could exploit first dressing up their theater models and then scaling them down to more modest sized homes and modest sized budgets. But theater speakers were never intended as high accuracy high fidelity sound reproducers for use in homes. And they never struck me as sounding that way.

OTOH, that was the claim made by the manfacturers of East Coast sound loudspeakers and they didn't strike me as particularly accurate either. For the longest time I just accepted it, there was real music and there were recordings. That is why I was so surprised at the live versus recorded demos I heard with AR speakers at trade shows. It could be done...under some circumstances in very carefully controlled environments, but wider experience drew me to the conclusion that this was the rare exception, not the rule. How interesting to find out decades later that they had to advance the treble control on pre-amplifiers to make it work. That is just the way AR speakers sounded to my ears, the sacrosanct "flat" positions of tone controls resulting in what was often a rolled off indistinct high end. (Tests I'd seen later also showed many receivers and amplifiers had their flattest FR with their tone controls set away from their indicated flat positions. Perhaps that's why tone control defeat switches were introduced.) Anyway, I've spent considerable time and effort trying to understand why this is so and I think I've gotten some pretty good insight into the problem.

I am not belittling the efforts of either group, both made very valuable contributions to the state of the art of sound reproduction. Had it not been for the pioneering work done on the west coast to make practical loudspeaker drivers reliable and affordable, Villchur never could have gotten off the ground. On the other hand, I will not give undeserved praise either since IMO none of them have come close to solving the overall problem they claim they have. I've always thought it one of life's great ironies that after Villchur left the audio equipment businesss, he went into the hearing aid business. I've wondered how many of his customers for his later products acquired the need for them by abusing their hearing using his earlier products. Reminds me of a dentist I had as a kid who gave each of his patients a piece of candy when they left his office (true story.)

[Steve F]

Zilch 4/3:

"The same reasons, actually, based upon an erroneous belief that the reverberant field is dominant in typical home listening spaces."

Zilch 4/4:

"....they did not publish system frequency response data for the same more fundamentally good reason -- it was "problematic," and would not stand up under objective scrutiny...."

Two completely different reasons by Zilch. Fist he says that Advent and Bose didn't puublish FR curves because they believed that far-field reverberant was correct.

I proved that that wasn't the reason in either company's case at all.

Now Zilch says they didn't publish because they were afraid their curves would not stand up under scrutiny. If by this, Zilch means that both Advent and Bose didn't want to set themselves up for criticism/misinterpretation over their FR curves, this may be somewhat closer to the truth.

But it's a COMPLETELY different reason than Zilch gave on 4/3. You can't have it both ways. Advent CERTAINLY didn't believe that the reverberant field was dominent in the home listening space and Zilch needs to come clean and admit his error if his credibility is to be maintained.

Steve F.

[speaker dave]

Steve:

I think the main reason why manufacturers typically don't show high resolution response curves is that they are rarely flattering. For decades many of the marketing departments would hand draw smooth (just enough wiggle to look plausible), flat, purely ficticious curves. Even legitimate companies would do this, such as the Altec A7 curve that Davis and others had published. These became fixed in consumer's minds as what typical speaker performance was. By comparison even an excellent speaker, shown warts and all, would look inferior.

At my first industry job I spent my spare hours measuring any competitive product hanging around the lab. I remember measuring a first Large Advent and, assuming it was defective, grabbing the second only to see that the curve was identical. It wasn't bad, really, but (my recollection) was the the tweeter response was very uneven above 10kHz. First it bumped up 4-5dB then dropped into a large hole around 15kHz. Otherwise the response was fairly flat (I don't remember any particular octave to octave balance, but its been a couple of years!) What shocked me was that I had a conditioned expectation of a level of performance and every speaker I measured was well below that level.

In this context, it is understandable for AR to publish individual driver curves taken under ideal mountings that are (yes) a little misleading of what the total system performance would be. The curves also well illustrate AR's strength at that time, wide dispersion from dome drivers in an era where larger cones and horns were the norm. The complete AR3a curve in Allison's paper are probably better that 90% (if not 100%) of the rest of the market at that time, but would confuse customers. To much honesty.

For Howard: How best to measure a loudspeaker (direct free field response, power response, in-room curve, etc.) remains the most interesting and most important question remaining in our field. I do believe that a concensus has been gelling over the last 20 years that the direct field is the stronger indicator, or better still that the ear uses the direct response at high frequencies with a time window that opens up at mid frequencies to include floor and ceiling bounces and finally at low frequencies is wide enough to encompass the room and steady state curves. I have certainly taken a lot of in-room curves in my time but I think you need to be very careful with them because it is impossible to define a universal far field room target curve. Most of the researchers that dig into the audibility of a speaker's power response (integral to a live room measurement) conclude that a flat sound power curve is very undesirable and that large holes in system power response are fairly benign. They do stop short of saying that variations in power response are totally inaudible.

Just my two cent worth.

"Everyone is entitled to my opinion."

David

[soundminded]

Zilch 4/3:

"The same reasons, actually, based upon an erroneous belief that the reverberant field is dominant in typical home listening spaces."

Zilch 4/4:

"....they did not publish system frequency response data for the same more fundamentally good reason -- it was "problematic," and would not stand up under objective scrutiny...."

Two completely different reasons by Zilch. Fist he says that Advent and Bose didn't puublish FR curves because they believed that far-field reverberant was correct.

I proved that that wasn't the reason in either company's case at all.

Now Zilch says they didn't publish because they were afraid their curves would not stand up under scrutiny. If by this, Zilch means that both Advent and Bose didn't want to set themselves up for criticism/misinterpretation over their FR curves, this may be somewhat closer to the truth.

But it's a COMPLETELY different reason than Zilch gave on 4/3. You can't have it both ways. Advent CERTAINLY didn't believe that the reverberant field was dominent in the home listening space and Zilch needs to come clean and admit his error if his credibility is to be maintained.

Steve F.

Which is important, the direct field from a loudspeaker or the reverberent field created by the room? IMO both are. In an interesting experiment, I tried modifying only the reverberant field of the top octave of 901. Try it sometime. I got a very interesting and unexpected kind of acoustic distortion heard only on some material. Try it and see what you get and we'll compare notes. In its current version, modified 901 propagates well over 95% of its highest frequencies away from the listener. But without that few percent aimed directly, things go very wrong. I'm trying to design a purely reflective primary speaker but having no luck. Results so far have been disappointing. Another surprise was how important reflections off the ceiling are. I know this flies in the face of the literature but that's what I've found.

In examining horn type instruments (I played the clarinet many years ago) these are also usually aimed away from the listner. A clarinet aims all of its sound straight down at the floor. In an anechoic chamber played this way a listener would probably hear little or nothing coming out of it. The first sound you hear from it in a real room is a reflection off the floor. I therefore don't find the terms direct and reverberant quite accurate anymore, instead I think of them as anisotropic and apparently isotropic. The reverberent field isn't really directionless, in fact it is critical how the echoes arrive, that is from which direction and when in relation to the corresponding field coming from the direction of the listener. There are no speakers I've ever seen or head of that are engineered to adjust these two fields independently of each other. Tweeter and midrange controls adjust both simultaneously, as does the Bose 901's or any other equalizer. Practically no engineering effort has been made to integrate the performance of a loudspeaker into the acoustics of the different rooms it might be used in. Small wonder it not only don't they perform their intended function very well but never sounds the same in any two different places.

[Steve F]

Speaker Dave:

" For decades many of the marketing departments would hand draw smooth (just enough wiggle to look plausible), flat, purely ficticious curves...."

I'm shocked--SHOCKED!!--that a Marketing person might do this.

(Although 1/3-octave-smoothed curves are already pretty nice looking. Especially if you compress the scale.....)

In the very latest AR Classic lit, they showed both the individual driver curves and the power response curves for each model. Those power response curves always looked a little too good to me, very convieniently degrading just the right amount as you went down from one model to the next.

Actually, you really had to hunt around in AR's lit for the 'truth,' but it was there if you looked hard enough. There was a graph of the 2ax's drivers' response where all three drivers were overlaid on the same graph. Sure enough, the tweeter is down in level relative to the W-M and never reaches their level, EXACTLY as Speaker Dave found in his measurements. I'll try to find it and post it.

I think what we sometimes forget, at least to a certain degree, was that these companies were for-profit enterprises, not non-profit entities serving only the public (audiophile) good. They were entitled to 'stretch' things a little and present them in the most flattering light (like any company in any industry) so that people would BUY them. That was the point. Even for Classic AR.

Steve F.

[Zilch]

Now Zilch says they didn't publish because they were afraid their curves would not stand up under scrutiny. If by this, Zilch means that both Advent and Bose didn't want to set themselves up for criticism/misinterpretation over their FR curves, this may be somewhat closer to the truth.

But it's a COMPLETELY different reason than Zilch gave on 4/3. You can't have it both ways. Advent CERTAINLY didn't believe that the reverberant field was dominent in the home listening space and Zilch needs to come clean and admit his error if his credibility is to be maintained.

My original statement was no less "broad brush" than yours, and served to elicit more information. Let's find definitive documentation of Kloss's design approach, and discuss that. With respect to Bose and AR, I have presented evidence in support of my statement.

I do have experience with large Advents, and they don't image worth a whit, though neither my measurements nor any others I've seen have thus far indicated why. I'd sure like to know what's up with them, though that's another thread.

I care far less about my credibility than getting at the facts. If you want an admission, fine: I'm ignorant with respect to the fundamentals of Advent design. Show me (and everybody else) the way....

Edit: The curves you reference are here, mistitled:

http://www.classicspeakerpages.net/library...frequency_resp/

[Zilch]

For Howard: How best to measure a loudspeaker (direct free field response, power response, in-room curve, etc.) remains the most interesting and most important question remaining in our field. I do believe that a concensus has been gelling over the last 20 years that the direct field is the stronger indicator, or better still that the ear uses the direct response at high frequencies with a time window that opens up at mid frequencies to include floor and ceiling bounces and finally at low frequencies is wide enough to encompass the room and steady state curves. I have certainly taken a lot of in-room curves in my time but I think you need to be very careful with them because it is impossible to define a universal far field room target curve. Most of the researchers that dig into the audibility of a speaker's power response (integral to a live room measurement) conclude that a flat sound power curve is very undesirable and that large holes in system power response are fairly benign. They do stop short of saying that variations in power response are totally inaudible.

I find Toole Fig. 18.5 (p. 375) and the accompanying text astounding, though I'm not sure I've completely got it yet, the implication being that the room is all but irrelevant. That's somewhat consistent with my own experience, but it's certainly inconsistent with prevailing wisdom as to how speakers and rooms interact....

At my first industry job I spent my spare hours measuring any competitive product hanging around the lab. I remember measuring a first Large Advent and, assuming it was defective, grabbing the second only to see that the curve was identical. It wasn't bad, really, but (my recollection) was the the tweeter response was very uneven above 10kHz. First it bumped up 4-5dB then dropped into a large hole around 15kHz. Otherwise the response was fairly flat (I don't remember any particular octave to octave balance, but its been a couple of years!) What shocked me was that I had a conditioned expectation of a level of performance and every speaker I measured was well below that level.

Like this?

http://www.audiokarma.org/forums/showthrea...073#post1913073

Scroll down to #21; you've gotta put the grille on....

[soundminded]

Speaker Dave:

" For decades many of the marketing departments would hand draw smooth (just enough wiggle to look plausible), flat, purely ficticious curves...."

I'm shocked--SHOCKED!!--that a Marketing person might do this.

(Although 1/3-octave-smoothed curves are already pretty nice looking. Especially if you compress the scale.....)

In the very latest AR Classic lit, they showed both the individual driver curves and the power response curves for each model. Those power response curves always looked a little too good to me, very convieniently degrading just the right amount as you went down from one model to the next.

Actually, you really had to hunt around in AR's lit for the 'truth,' but it was there if you looked hard enough. There was a graph of the 2ax's drivers' response where all three drivers were overlaid on the same graph. Sure enough, the tweeter is down in level relative to the W-M and never reaches their level, EXACTLY as Speaker Dave found in his measurements. I'll try to find it and post it.

I think what we sometimes forget, at least to a certain degree, was that these companies were for-profit enterprises, not non-profit entities serving only the public (audiophile) good. They were entitled to 'stretch' things a little and present them in the most flattering light (like any company in any industry) so that people would BUY them. That was the point. Even for Classic AR.

Steve F.

"I think what we sometimes forget, at least to a certain degree, was that these companies were for-profit enterprises, not non-profit entities serving only the public (audiophile) good. They were entitled to 'stretch' things a little and present them in the most flattering light (like any company in any industry) so that people would BUY them"

I'm shocked! SHOCKED! They lied. And deliberately too. They are no better than all the others. If you can't sell the steak, sell the sizzle. Which leads me only to repeat what I've posted before. The live versus recorded tests were remarkable, not because they produced exactly the same sound but because under contrived conditions that they came as close as they did, that it could be made to work as well as it did at all. And their contribution is not to be belittled but not to be taken as the ultimate solution to the larger and far more difficult problem. And we could say exactly the same about what is being offered to the market today. Only at far higher prices and with even more pretentious hype. On a scale of 0 to 100 where 0 represents even a casual listener with normal hearing being able to tell instantly he is hearing a recording and not the real thing and 100 being even experienced concertgoers with excellent hearing and good acoustic memory being unable to tell most or all of the time they were hearing a recording and not the real thing, the state of the art remains sadly at.....0.

[genek]

How best to measure a loudspeaker (direct free field response, power response, in-room curve, etc.) remains the most interesting and most important question remaining in our field.

Maybe to those of you who actually work in the field. As someone whose only role in it is to listen to products and spend money, I know that the fact that you all can't agree on a single, standardized method is what makes most speaker measurements pretty much useless as a predictor of the actual listening experience. If all speakers were measured the same way, then the results might be as useful as the EPA mileage numbers on cars: by themselves they'd tell you little about real-world performance in your living room and nothing at all about whether you will like that performance, but once you had taken the time to listen to a lot of product and decide which ones you like and don't like, they might make it possible to predict whether an unknown product will be closer to the ones you like or the ones you don't.

So ether pick just one method and use it on everything, or use them all on everything. Anything else just presents those of us who are not in your field with the apparent spectacle of a bunch of dueling know-it-alls repeatedly pelting each other with apples and oranges.