Stereophile Review of AR3a...

[genek]

On one hand, I agree that a high preference rank is a very important indicator of accuracy. On the other hand, it is already established that some number of alternative products with readily distinguishable sound can all fall within the most preferred bin, and so we haven't fully determined a scale of absolute accuracy, or clear design goals.

Unless the listening panel is consistently mistaking the reproduced sound for the real thing, how is it any indication of "accuracy?"

In the absence of a "scale of absolute accuracy or clear design goals," it seems to me that a high preference rank is only an indicator of popularity. Again, a perfectly valid criteria for manufacturing and marketing decisions but not much else.

[Zilch]

Let's not confuse accuracy with perceived realism or preference. Accuracy is objectively quantifiable. Toole & Olive have found that listener preferences are highly correlated with accuracy, but it could easily be otherwise.

Realism is separate, in my view. AR3a may produce a realistic rendering of a concert hall presentation of orchestral music in small listening spaces, or a small ensemble in a highly reverberant one, independent of accuracy. By objective standards, it is not an accurate loudspeaker, rather, has a characteristic coloration specific to its intended use. Similarly, JBL 100 is voiced specifically to enhance its presentation of rock music, and is not accurate....

[genek]

Let's not confuse accuracy with perceived realism or preference. Accuracy is objectively quantifiable.

Only if there is consensus on what it is, i.e., what Ken describes as "a scale of absolute accuracy." I'll leave it to the pros to debate what that is or should be. For me, it's enough to know that manufacturing and marketing decisions are currently being made to some criteria that ends up producing new products I have no interest in.

[Zilch]

The problem with the whole direct field/reverberent field debate is that none of our observations truly prove our respective points. Most of the speakers we would point to have a fairly flat anechoic response and a gradually falling power response. The room adds increasing absorption (towards the high end) and so the end result without EQ is a room curve that falls off from LF to HF. Howard would say "gently falling in-room response. That's why it sounds good." Zilch and I would say "the underlying direct field is flat. That's why it sounds good." As long as the power response falls predictably, i.e. the directivity index rises gently, the direct field can be flat while the room curve follows popular notions of what in-room curves should be.

Independent of the issue of anechoic vs. in-room response, AR design philosophy asserted that power response (their thankfully unfilled objective being that it should be flat) was more definitive of what the listener would hear than on-axis frequency response. Allison and Berkovitz presented both, the only known publication of actual AR system frequency response from the vintage era, along with a spatially averaged (heroically, over multiple locations in multiple rooms) and smoothed in-room power response.

Which does it better match?

[speaker dave]

Independent of the issue of anechoic vs. in-room response, AR design philosophy asserted that power response (their thankfully unfilled objective being that it should be flat) was more definitive of what the listener would hear than on-axis frequency response. Allison and Berkovitz presented both, the only known publication of actual AR system frequency response from the vintage era, along with a spatially averaged (heroically, over multiple locations in multiple rooms) and smoothed in-room power response.

Which does it better match?

Now the AR3a is an interesting case.

It really does have wide dispersion, generally only constrained by the baffle to 180 degrees. Its axial response (aside from interference and edge effects) can closely mimic its power response. Rather than making it flat on axis and flat in power AR gave it the same power response as most other speakers and a falling on axis response.

If only power response mattered, the AR3a, which matches modern speakers in power response, would sound just the same. Truth is that most people that know it acknowledge that it is considerably rolled off relative to "modern speakers" (such as the KLH 6!). That they like the roll off and feel it suits orchestral music isn't the issue. It is that a conventional power response doesn't prevent the rolled off direct response from giving the system a duller sound.

David

[speaker dave]

I just D/L'd Olive's paper on room correction. I'm not yet convinced it reveals that much in the way of loudspeaker design goals, at least in the areas we are talking about. OK, we know that peaks, dips and resonances are a bad thing, and that 1/3-octave, quasi-CB resolution is not enough. We know that increasing smoothness and bass extension will generally lead to a preferred speaker.

My problem with it is that it is still "steady state" based. As with all such systems you have to have some sort of room curve. JBL seems to have used a 1dB per octave slope. The disliked systems (except #6) are nearly as smooth but have chosen a different room curve, so success or failure is based on choosing the right room curve.

I have experience with the Snell RCS1000 (predecessor to the Tact and the Lyngdorf) and have also played and measured the Audesey (was not impressed). With the Snell you could get a good result if you did a lot of listening and then massaged the target curve until it sounded good. I sat through some Audesey sessions and it was the same thing: they would tweak a dB here and a dB there until the golden ears thought it was right.

Automatic room correction indeed.

Wasn't the Berkovitz box more time domain based (EQed the direct field)?

David

[genek]

Now the AR3a is an interesting case.

It really does have wide dispersion, generally only constrained by the baffle to 180 degrees. Its axial response (aside from interference and edge effects) can closely mimic its power response. Rather than making it flat on axis and flat in power AR gave it the same power response as most other speakers and a falling on axis response.

If only power response mattered, the AR3a, which matches modern speakers in power response, would sound just the same. Truth is that most people that know it acknowledge that it is considerably rolled off relative to "modern speakers" (such as the KLH 6!). That they like the roll off and feel it suits orchestral music isn't the issue. It is that a conventional power response doesn't prevent the rolled off direct response from giving the system a duller sound.

I have to question whether the rolloff had that much to do with my original selection of ARs over the other speakers I considered way back when. I listen to other kinds of music besides orchestral, and I do occasionally turn up my treble controls. The thing that stands foremost in my memory about the speaker listening exercise was that the ARs (I listened to 3a. 5 and 2ax, not any of the smaller models) seemed to fix the apparent sources of sounds better as I moved around the listening rooms...no wandering musicians...and that there was nothing that the speakers seemed to deliver too much or too little of that couldn't easily be adjusted to my satisfaction with amplifier tone controls. This was long before the internet and CSP came along to provide me with access to all the old AR white papers and multiple analyses of dispersion, "flat power" and room effects, btw, and the only reading materials I had available to me at the time were AR's single-sheet flyers (the ones with musicians on them), some reviews in Audio magazine and a book of collected past reviews from High Fidelity.

[Zilch]

If only power response mattered, the AR3a, which matches modern speakers in power response, would sound just the same. Truth is that most people that know it acknowledge that it is considerably rolled off relative to "modern speakers" (such as the KLH 6!). That they like the roll off and feel it suits orchestral music isn't the issue. It is that a conventional power response doesn't prevent the rolled off direct response from giving the system a duller sound.

My point exactly. Room absorption mooted the wide dispersion, the in-room curve tracks the direct field response, NOT the total power as measured in the reverberant chamber, and the core thesis of listeners in a dominant reverberant field per Beranek, the linchpin of AR design, fails.

Run the +/- 15° power response, and I suspect it will match the measured in-room curve quite nicely.

The thing that stands foremost in my memory about the speaker listening exercise was that the ARs (I listened to 3a. 5 and 2ax, not any of the smaller models) seemed to fix the apparent sources of sounds better as I moved around the listening rooms....

Constant directivity. Read the EconoWave listener evaluations for similar.

[Zilch]

The solution to this problem (and perhaps the only solution, since simple equalization cannot alter the contrast between the direct-field output and the reverberant-field output) is to use drivers that have wide and uniform radiation over their entire crossover-controlled operating ranges. Done right, this keeps the on-axis, off-axis, and wide off-axis outputs all smooth.

That puts the room in control, which is entirely unnecessary.

[Though it does provide amply expanded fodder for subjectivist audiophool blather, the stock and trade of reviewers.... ]

[genek]

Constant directivity. Read the EconoWave listener evaluations for similar.

So are these waveguides strictly DIY parts, or is anyone using them in a speaker I might have a chance of finding in a dealer demo room somewhere?

[Zilch]

So are these waveguides strictly DIY parts, or is anyone using them in a speaker I might have a chance of finding in a dealer demo room somewhere?

Haven't we been here already? They are in JBL Pro products. You can hear some of them at Guitar Center, or equal, and the better ones at a JBL Pro dealer. See also Best Buy Pro, which may have the QSCs, and a policy allowing you to try a pair of those for 30 days.

OR, you can spend $100 and get a waveguide, inexpensive compression driver like Selenium D220Ti, and the eWave crossover to play with....

[genek]

Haven't we been here already? They are in JBL Pro products. You can hear some of them at Guitar Center, or equal, and the better ones at a JBL Pro dealer. See also Best Buy Pro, which may have the QSCs, and a policy allowing you to try a pair of those for 30 days.

OR, you can spend $100 and get a waveguide, inexpensive compression driver like Selenium D220Ti, and the eWave crossover to play with....

I remember that the JBLs use waveguides, but off hand I don't recall you mentioning that they are the exact same parts as eWave.

If I see or hear something I like I may decide to buy it, or if something I already own breaks down and there's nobody but me around to fix it I might do that. But I don't have the time or ambition to build things anymore.

[Carlspeak]

You guys may be neghbors (of sorts). Both live in CA. Maybe Zilch could invite you over to his lair for an e-wave listening session perhaps?

[genek]

You guys may be neghbors (of sorts).

I wish. Zilch, Ken and I used to live within commuting distance of each other, but now I'm down in San Diego.

[Shacky]

I heard a pair of Indignias at a local AK meet. Not the econowaves but I know Zilch has put a lot of work into both designs. They were very nice sounding. Very nice bass for their size. They would not be my cup of tea. Too bright for my ears. And we had a local speaker maker on site who noted that most "flat" speakers sound bright to people.

We played Stevie Ray Vaughn's Little Wing/Third Rock track on them in the small listening room. It was way too loud for me to listen directly for more than a minute or two. But it was very nice sounding.

Now in the large room we had my AR 3a's set up driven by a Sherwood SS Reciever. Same SRV track blew the crowd away. Was one "better" than the other? The Indignias were probably more revealing. But no where near the awe inspiring bass as when Bobby Shannon kicks in on the beginning of that track. You knew everyone liked the AR's as people had been mulling around till they came on then everyone sat down to do some serious listening.

This was in a large square room that was not the most acoustically correct setup. But the AR's were shining.

So while I know Zilch has contributed to very nice low cost deisgns like the econowave and indignia projects, I'll take my AR 3a's any day. As I would bet about 80% of the AK'ers we had in attendence. The AR's need work to be at their best, lots of power and probably high damping factor amp. But when they are on they are really on.

So theory aside, I'll put my AR 3a's up soundwise to just about anything out there.

[speaker dave]

I ran some of our material (no names were mentioned) past Roy Allison, and got the following comment. (Roy is a man of few words.)

"Interesting how people just don't seem to get that first arrival primarily determines localization, but in typical conditions cumulative energy for 30 msec or so determines perceived spectral balance."

Howard Ferstler

So in a room with 0.5 to 1.0 second RT, how much of the reverberent energy do we get to discard?

(Answer: Quite a bit.)

[Zilch]

"Interesting how people just don't seem to get that first arrival primarily determines localization, but in typical conditions cumulative energy for 30 msec or so determines perceived spectral balance."

Geddes says 10ms, Kantor says 20 ms, and Allison says 30ms; whatever way, we don't want high-IACC early reflections adulterating the spectral balance. There are means to enhance spatiality WITHOUT allowing the room to degrade the imaging and corrupt the program content. Wide dispersion is NOT the ticket to excellence in sound quality....

[Zilch]

As I noted, the angle from 45 degrees off axis to 90 degrees off axis covers an AREA almost two and a half times as large as as what we have covered from zero to 90 [45°?] degrees off.

Exactly why we don't want it.

You talk as if there is this wall of infinite reflections impinging upon our ears. It's not that at all; the number of reflective paths at any given location in space is finite, and one need only attenuate the major ones to alter the presentation.

So, Zilch, while you may think I am obsessed with 1970s technology and you and Geddes are the wave of the future, I think it is you guys who are operating in a dated-performance audio world. Allison and Villchur rendered your approach obsolete decades ago.

BAH!

Footnotes, easily replicated.

[Alas, poor Howard, nobody wants it today, the very definition of obsolescence..... ]

[genek]

Interesting chicken-vs-egg argument.

Are wide-dispersion stereo speakers no longer offered by audio manufacturers because consumers prefer multi-speaker surround sound and don't want them anymore, or is the decline of stereo and the growth of multi-speaker surround sound the result of consumers trying to find some other way to fill their rooms with sound the way the wide dispersion speakers they can no longer get used to do for them? Which came first, the disappearance of wide dispersion stereo speakers or the growth of surround sound?

Wide dispersion stereo speakers are certainly gone from the consumer market, but consumers aren't exactly mobbing audio dealers for narrow focus stereo speakers either. It's pretty clear that one way or the other, consumers want sound coming at them from more directions than just straight ahead.

[Zilch]

Which came first, the disappearance of wide dispersion stereo speakers or the growth of surround sound?

Easy one. Allison knows the answer.

There are enough reflections to make speakers listened to indoors (even the most directional models) sound radically different from the same speakers listened to outdoors. If what you say is true, putting some absorbing materials on the room walls to eliminate those "reflective paths" would have the place sounding just like an anechoic chamber. Your reflection-attenuation technique basically tries to unsuccessfully turn speakers into oversized headphones. You cannot get away from the room, and it is better to work with it than against it. Of couse, for this to work best one at least a good room to begin with.

You should sign on with one of the upscale purveyors of room treatments, Howard; your perspective is patently elitist.

Toole tells how to do it, if required, but through intelligent selection and deployment of loudspeakers, mitigation of early lateral reflections is easily accomplished with minimal intervention, while preserving effective spatial cues AND precise imaging. This can be accomplished in virtually any room, including the padded cell you presume to comprise my primary listening space....

[speaker dave]

I ran your comment above past Roy and he said:

"This fellow must own a concert hall to get an Rt approaching 1sec. Small domestic listening rooms typically have 0.25sec at very low freq's, 0.15 at HF. Medium to large domestic living rooms 0.35 at LF, 0.25 at HF. Nothing is thrown away is such rooms for listening spectral balance detection."

End of his typically brief comment.

Howard Ferstler

Mine is more that .5 sec midband and not overly large or reverberent. I'll pull out the curves. I've measured it twice.

It seems to have escaped everyones notice that Roy has apparently had a change of mind on the very issue that we have long been debating.

He is now accepting that perception of frequency balance is based on an early-sound time window (that he states as) 30ms long. This is distinctly different than the previous notion that balance is not a time windowed phenomonon and is purely based on the steady state response, the total cumulation of direct and reverberent energy.

From the 1970 AES paper: " These results can be summarized as follows. 1)That we are convinced that home listeners perceive the spectral balance of the sum of the direct and reverberent fields, and that the very small time differences between them have no effect on this perception of balance."

Now he states: "In typical conditions cumulative energy for 30 msec or so determines perceived spectral balance."

Its perfectly fine for him to modify his views on the subject. It has been 40 years and a large number of papers have been published that talk about our ability to focus in on direct and early sound while ignoring the reverberent tail. I have repeatedly cited a few by Kates, Bech and Lipshitz and Vanderkooy.

This doesn't diminish the worth of the "Sound Field" paper in documenting the combination of a quality loudspeaker and the typical domestic listening room, although conclusions about the perceived effect of high frequency rolloffs should be reconsidered.

[Zilch]

It seems to have escaped everyones notice that Roy has apparently had a change of mind on the very issue that we have long been debating.

Slam-dunked right here in The Kitchen just short of a year ago, actually:

http://www.classicspeakerpages.net/IP.Boar...ost&p=80407

[kkantor]

Geddes says 10ms, Kantor says 20 ms, and Allison says 30ms; whatever way, we don't want high-IACC early reflections adulterating the spectral balance. There are means to enhance spatiality WITHOUT allowing the room to degrade the imaging and corrupt the program content. Wide dispersion is NOT the ticket to excellence in sound quality....

Well, see, the Law of Averages proves me right!

What I really think is that one can establish guidelines, but that it is not possible to define a hard and fast number. Certain characteristics of the source material effect the results a great deal, such as how centrally pitched the signal is, and what its amplitude envelope looks like.

I think I said in my older pubs that using music with typical power envelope modulation rates, 0 to 600us is transient localization territory, 600us to 2ms is the general range defining transient timbre, 600us to 10ms is source localization and 2ms to 30ms is tonality, but don't quote me.

-k

I am Seattle for an evening. The hotel I am staying at is hosting meetings of DARPA and The America Tattoo Society. Hanging out in the lobby is interesting....

[speaker dave]

Well, see, the Law of Averages proves me right!

I think I said in my older pubs that using music with typical power envelope modulation rates, 0 to 600us is transient localization territory, 600us to 2ms is the general range defining transient timbre, 600us to 10ms is source localization and 2ms to 30ms is tonality, but don't quote me.

-k

I am Seattle for an evening. The hotel I am staying at is hosting meetings of DARPA and The America Tattoo Society. Hanging out in the lobby is interesting....

Most of the research I've seen favors a variable time window. Long enough to be essentially the steady state response at LF. Short enough to be the direct response at HF, and (per Soren Bech) just enough to catch the floor bounce at mid frequencies.

Tattoos eh? Sounds like staying at the CES hotel that hosts the AVN awards in Vegas. Always interesting!

David

[speaker dave]

Slam-dunked right here in The Kitchen just short of a year ago, actually:

http://www.classicspeakerpages.net/IP.Boar...ost&p=80407

Sorry, but I missed that the first time around.

So if it has all been settled, why are we still debating it?