Screen Array

[Zilch]

Vertically asymmetric (20° + 30° x 90°) "Second-generation Optimized Aperture™" mega-PT ("Progressive Transition™") waveguide, 30" W x 15" H, available in both 1" 2374 (126-00177-00) thread-on and 1.5" 2384 (126-00177-01) bolt-on throat variants:

[mikebake]

Not good...................

[Zilch]

Dunno yet.

How low do it go?

What's the SSC™ "Screen Spreading Compensation" about? :dont-know

[Robh3606]

So 50 with a downward tilt?

Rob:)

[Zilch]

Or upward, Rob, as I believe I have it sitting there, the mouth wider at the top.

Think Altec Model 14's asymmetric MantaRay MR931, which was a bit more of an extreme split:

http://www.lansingheritage.org/html/altec/specs/components/mr931.htm

[Minimize floor bounce.... :yes: ]

[Robh3606]

Or upward, Rob, as I believe I have it sitting there.


Right either way depending on the speaker height and the type seating in the venue they are used in. I am thinking flat vs stadium as an example.

Rob:)

[grumpy]

SSC

http://www.jblpro.com/catalog/support/getfile.aspx?docid=864&doctype=3

(p19)

e.g., hotter on-axis vs. off compared to non-SSC design (simplification).

Have fun w/ dat. :)

[Zilch]

e.g., hotter on-axis vs. off compared to non-SSC design (simplification).I'll see what the measurements say, but I read that as saying the HF comp is being done by the waveguide, and I do see it (the means) in the rising DI and narrowing horizontal beamwidth, but it appears to be well and precisely controlled.

Normally, it says, you'd have to add HF comp, but with SSC™, you don't have to? Normally, you'd have to do that anyway for flat on-axis response with a constant-directivity horn or waveguide. Here, it looks like they're progressively beaming the VHF, from 90° to 65° or 70° horizontal beamwidth, instead, to accomplish it, and correspondingly less HF comp should be required.

The Klipsch Tractrix KPT-402 horizontal beams at twice that rate, and rather more messily all the way:

http://audioheritage.org/vbulletin/showthread.php?p=235211#post235211

What am I missing? We're giving up a controlled amount of power response uniformity at higher frequencies so as not to have to apply the requisite additional boost? I'm certainly not afraid of either.

Thanks for digging out the curves, which do not appear in some of the product data sheets, Grumpy.... :thmbsup:

[JBL 4645]

I’d say a safe bet is 500Hz or 800Hz upwards.

[kegger]

Following with interest here Zilch,

Thanks for the headsup, I certainly am looking into these, :dancin:

[Zilch]

What Grumpy said:

[kegger]

Am I correct in reading that CD comp is still needed here?

What drivers have you tried now and what xovers as well?

-------------------

And it looks like the horn has a "boost" at around 20K?

[Skywave-Rider]

Zilch,
Very interested to see what happens w/ BMS 4555 19kHz spike on these, as that would be my option.

Thanks for doing all this.:applaud:

I anticipate a group buy of perf screen for acoustic filtering.

[Zilch]

Am I correct in reading that CD comp is still needed here?Yes. I've shown it with the stock Model 19 passive compenstation set at max, which provides ~9 - 10 dB of "boost" via midband attenuation.

That's not enough with this driver, but it gets this combination into the realm of "listenable." :yes:

[Zilch]

I anticipate a group buy of perf screen for acoustic filtering.Stock JBL Blue grille cloth alone certainly doesn't get it, Gray, Black:

[boputnam]

So 50° (vertical dispersion) with a downward tilt? As Zilch sez, it depends upon physical positioning.


Vertically asymmetric (20° + 30° x 90°)Typically, that would be quoted as -20°/+30°x 90°H. But, you could flip it over so to be -30°/+20°x 90°H, depending upon use, or, if you want something more easily arrayable, rotate it 90° for a -45°/+45°x 20°/30°H...:p

[Zilch]

... if you want something more easily arrayable, rotate it 90° for a -45°/+45°x 20°/30°H....2360 dispersion in a "diminutive" package.... :D



Tracked down by member Publius on AudioKarma:

[grumpy]

...planning on putting a perforated screen in front of the horn?

Just curious what the goal is other than verifying JBL's curves...
and playing with toys :D. A narrowing beam with increasing frequency
above 4KHz doesn't sound like an obvious home-use desire-ment, but
does sound like radial horn behavior (albeit less extreme in the narrowing).

[Zilch]

... planning on putting a perforated screen in front of the horn?It's implicit in the patent and intended use that full constant directivity is resored behind a perfed screen. I'm thinking that other potential grill material candidates such as transparent sunscreen, which is available in different open area specs, might mimick that behavior. As Skywave suggests, perf screen itself might not be a bad option, or typical JBL Pro perf metal, perhaps. Do we know how that behaves in this respect?


Just curious what the goal is other than verifying JBL's curves ... and playing with toys :D.They're PT waveguides that will play low, apparently. JBL crosses them @ 1.3 kHz with 2418H-1 in 3622, but 750 Hz (18 dB/oct) using 2430H in 4622. I can only speculate from the above curves what is possible using true large-format drivers, but the dimensions suggest what might be practical. OA design criteria say 800 Hz, at least, but somewhat, if not considerably lower may be possible.

Dance5 uses 2451H:

http://www.jblpro.com/catalog/support/getfile.aspx?docid=810&doctype=3

The 1.5" mounting flange is factory drilled for both mounting bolt circle diameters.

2426H is used in another three-way Screen Array, 4632. There are options here, once the SSC issue is resolved, assuming there is one. :hmm:


A narrowing beam with increasing frequency above 4KHz doesn't sound like an obvious home-use desire-ment, but does sound like radial horn behavior (albeit less extreme in the narrowing).A Klipshorn upgrade, perhaps? ;)


Very interested to see what happens w/ BMS 4555 19kHz spike on these, as that would be my option.Measurements are up on AK now, Skywave. The spike is certainly there on-axis, but substantially supressed off-axis.... :yes:

[kegger]

Also I'm curious if some of the drivers that have a droop in the extreme top end
even after compensation might benefit from that built in "boost" there? Hummmm..

I believe some of the pro drivers with beefier diaphragms fall off up there pretty fast?

[grumpy]

oop. sorry.. didn't go back and re-read entire thread when new page
got started. Carry on :).

[Zilch]

O.K., who understands precedence compensation as implemented using 2405H in Everest I? (I'm not certain that I do.)

Here's the proposal:

Set up the L/R 2384s per the "Geddes alignment," i.e., toed-in 45° and crossing in front of the listening position. A listener on center is at 30° off-axis to both of them, with the frequency response EQ'd or comp'd flat at that angle, Green, below.

As the listener moves to one side or the other (or, as other listeners might be aligned,) they move more off-axis but closer to the proximal speaker, and the SPL remains constant, while concurrently, more on-axis to the distal one, and it remans constant also. This mimicks the defined directivity Everest (also S2600 and S3100) stationary image effect, broadening the "sweet spot" substantially. It works with any constant directivity horns/waveguides deployed using extreme toe-in, in fact.

As I understand it, to overcome the concomitant precedence effect, Everest I employed the supplemental 2405H UHF driver to disproportionately increase the VHF from the distal speaker at the same time. The result was a veritable festival of phase interference which the Screen Array waveguide with its integral SSC™, behaving similarly with similar disproportionate VHF boost, would not suffer, Red, below.

Do I have this right, or somehow backwards or otherwise screwed up? :dont-know

[Robh3606]

The way the Everest Horns work is defined directivity. They are designed to cover a rectangular area and have the same relative SPL in the front and back of the room. They shift where you would expect the highest SPL to be evened out at the back of the room or on their sides to the middle of the room. As you move left or right they work in reverse to the precedence effect because of their unique directivity.


Everest I employed the supplemental 2405H UHF driver to disproportionately increase the VHF from the distal speaker at the same time.

When the 2405 is aimed like it is in the Everest it does what Geedes does with his waveguides. They are aimed to cross in front of you. Seems to me that it would work even without the 2405 like it does in the newer smaller systems.

I think there is a design brief on the site.

http://www.audioheritage.org/vbulletin/showthread.php?t=5671

Rob:)

[subwoof]

That horn looks intriguing as a "modern" replacement for the 2360. I just assembled a party system at the college rental house I own and while the 2360 parked on top of the 31" wide 813 cabinets certainly looks cool, the depth and rigging required makes for a compromised living space.

That 2384 1.5" horn, with a 2450SL ( 2452SL but older ) on it, mounted inside a box that had a screened front would still give that "wow" look and really cut down on the delay required ( and give better HF extension than the mile-long 2" throat ).

Oh well, more money more money. I'll check with my parts guy for the cost today.

sub

[grumpy]

Hmm... 1) Anyone have the rest of the figures for that paper?

2) seems like the frequency dependent directionality for the new PT's might sort of do
what is being suggested, but with an additional 'lobe' (not asymmetric in the horizontal
plane, as the 2346 horn as used in the DD55000) that would need to be squelched.

[Zilch]

The way the Everest Horns work is defined directivity.2384 is similarly defined directivity, but only in the VHF, which is why you must deploy them in the Geddes alignment.


As you move left or right they work in reverse to the precedence effect because of their unique directivity.They normalize the SPL only, much like adjusting the balance control for an off-center listening location, which, by itself, does not defeat the precedence effect, the time-of-arrivial cue. The direct sound from the distal speaker remains delayed, hence the supplemental 2405H providing disproportionately boosted VHF to further combat the effect of lateral displacement.


That 2384 1.5" horn, with a 2450SL ( 2452SL but older ) on it, mounted inside a box that had a screened front would still give that "wow" look and really cut down on the delay required ( and give better HF extension than the mile-long 2" throat ). Good point. The depth is significantly reduced due to the rapid flare rate; it's more like 2380A. "Throatless" 2450 should bolt right up to the 1.5" throat opening.


Oh well, more money more money. I'll check with my parts guy for the cost today.I believe you will find them to be surprisingly modest in cost, less than half that of 2352, for example.


Seems like the frequency dependent directionality for the new PT's might sort of do what is being suggested, but with an additional 'lobe' (not asymmetric in the horizontal plane, as the 2346 horn as used in the DD55000) that would need to be squelched.2384 does NOT produce that lobe, correct?

I believe that's what I'm suggesting here, anyway..... :yes:

[grumpy]

in the UHF (using the DD55000 paper terminology), you'd have the one narrower
on-axis lobe from the PT horn, and the HF (below 4-5KHz) might behave more like
a conical horn, pattern-wise. (skip the 'lobes' reference, my mind had wandered...)

I guess I don't (yet? :)) see the -midrange- image restoration function of the 2346
horn in just toeing in the speakers. I didn't get from that paper that -all- of the image
restoration (vs lateral listening position) came from the 2405's.

In that an asymmetrical amplitude response was used to offset lateral position
(inverse pan-pot effect), and was designed to be large enough to mask positional
and delay effects, I'd speculate that this is why -some- people find this effect to
sound artificial or unnatural... I wouldn't go that far, but it is different and noticeable.

That off-axis listening to the PT SSC horn would help restore a flatter response curve, I do see.
Usefulness of a shorter, smaller-mouthed horn than a 2360, I do see.
Seems like a nice behind-screen solution worth ringing out
if you have the time and inclination.

[Zilch]

All of this leads me to conclude that accurate imaging results when the audio signal from left and right channels arrives at the listening position according to (at least) three theretical aspects:

1. frequency response must be the same (including any reflections)
2. amplitude must be the same
3. arrival time must be the same

It has been observed that signals that are the same in frequency response and that arrive within a relatively small time window, ten(?) milliseconds or so, are combined by our brains into a single sound. So, for practical purposes we might be able to ignore distance differences to each of the two speakers for listeners who are not equidistant from the speakers in a home listening environment.

These asymmetric horns, project Everest I, and the other comments from folks on various fora suggest that when sound arrives in equal amplitude (#2) as well as equal frequency response (#1), good imaging is achieved. Project Everest and these asymmetric horns apparently were an attempt to have output vary with distance off-axis, so that for a listener positioned more directly in front of one speaker would have that speaker's sound attenuated to a level that would match, in level, the sound coming from the speaker that is further away.

Any constant directivity horns/waveguides will produce this effect when deployed using the Geddes extreme toe-in listening alignment I described above, HOWEVER,


Simply correcting the distance offset in dB between the loudspeakers and the listener by altering the polar patterns is not sufficient. Another factor, the precedence effect, comes into play. Transient (high frequency) information is localized at the earlier loudspeaker, even when the two loudspeakers are reproducing the sound at the same level at the listener. Thus, we designed into the system a degree of "over-correction" to counteract this.

Figure 7 shows the general characteristics of the precedence effect in terms of the level imbalance required to restore phantom center localization for a delay between loudspeakers. This data is derived from the studies of Haas, and there is a good bit of variability to it, depending upon the details of the test setup and the precise nature of the program. In our studies, we found that approaching the Haas values of level imbalance, as a function of delay difference, was most useful above 4 kHz. This ultimately resulted in our angling the UHF transducer sixty degrees from the normal straight-ahead position. Extensive listening tests in a variety of environments confirmed the validity of this.

The overcorrection at high frequencies is shown in Figure 10. Note that the overcorrection is apparent only above 4 kHz. At 16 kHz, it only approximates the Haas data shown in Figure 7. The need for this correction was estimated at the outset of the design cycle, but the precise amount was determined empirically, as we stated earlier.

Figures 7 - 10 appear at the bottom of the last page of the tech note, but are apparently not the ones cited by Eargle, above. Still, if I am reading those correctly, the HF gets "hotter" with increasing lateral distance (and consequently, angle) from either side, and with frequency, much as is "built into" 2384 on-axis response, and since the differential occurs smoothly and gradually over angles between 30° and 45° off axis with 2384, the precendence correction should be very similar to that of original Everest I specification.

I'm suggesting that 2384, suitably deployed, may not be the "problem" waveguide heretofore supposed, but instead and in fact, may produce an improved Everest I effect.... :yes:

[grumpy]

well, we all know you're going to try it... so, 'on with the show' :)

(don't think I ever implied they were problem horns... just a solution tuned to a problem)

[Robh3606]

Hello Zilch


They normalize the SPL only, much like adjusting the balance control for an off-center listening location, which, by itself, does not defeat the precedence effect, the time-of-arrivial cue. The direct sound from the distal speaker remains delayed, hence the supplemental 2405H providing disproportionately boosted VHF to further combat the effect of lateral displacement.

Ok the two work in conjunction and that is unique to Everest as they didn't use a UHF driver the the newer 2600/3100 systems.


I'm suggesting that 2384, suitably deployed, may not be the "problem" waveguide heretofore supposed, but instead and in fact, may produce an improved Everest I effect.... :yes:

Well you have me curious, how?? I am having trouble visualizing how you could aim the horn for it to work. You are talking 30-45 degree's off it's 90 degree axis?? or its 50 degree axis??

Rob:)

[subwoof]

Yeah - I was following this thread but I must of taken the wrong exit or a post was removed / edited.

Are you saying mount these vertically to get the desired effect? that would mean an inclusive angle of 50 degrees and that would remove them from normal DIY status due to the 30" height!

Any side view of a typical theater would show that the horn needs to have a greater *upward* pattern of whatever parameter is in play which is just the opposite of the 4346 ( everest 1 ) horns application in the 4660 church system.

So in your living room simply install the 2384 upside down so the "sweet" angle is firing *down* into the lowboy couch.

As a side note, the horn's vertical is specified 20/30 degrees - what is the pattern in the 2346's "unequal" axis when used in the everest 1 application? Maybe I missed it but this computer has none of my handy bookmarks..

sub

BTW I ordered a pair of these to "play" with. The dance5 system uses the perf metal grills with foam and I can make grills just like them to get the CD effect(?).

[grumpy]

what is the pattern in the 2346's "unequal" axis...

http://www.jblpro.com/pub/obsolete/4660.pdf

[Zilch]

Well you have me curious, how?? I am having trouble visualizing how you could aim the horn for it to work. You are talking 30-45 degree's off it's 90 degree axis?? or its 50 degree axis??You toe-in both speakers at 45°. With a 90° dispersion pattern, there are no sidewall reflections from the wall nearest each speaker, because the pattern boundaries are parallel to it, and also to the front wall. Yes, the boundaries are not absolute, but those are the -6 dB points, and any reflections that do occur from outside the controlled directivity window are substantially attenuated. At -10 dB, they are effectively not even present.

On-axis, then, the two speakers are crossfired with the included angle at the apex of the triangle being 90°. That's not where you listen, rather, you listen somewhat further out from where they cross, i.e., off-axis, outwardly, to each speaker. It may be seen, then, that as you move laterally, you are moving more off-axis to the speaker in the direction you move, and are moving closer to, so the SPL from that speaker remains constant. At the same time, you are moving more ON-axis of the speaker you are moving away from, and despite the now greater distance, the SPL remains the same from that speaker, as well.

Thus, you are, in effect, normalizing the SPL from both speakers to lateral movement in either direction, and the phantom image stays in the same location, midway between the speakers, replicating the effect Everest I, S2600, and S3100 achieve with their defined directivity horns and minimal, if any, toe-in.

A concomitant benefit is that the earliest reflection comes from the OPPOSITE speaker reflecting off either wall, which reflection is attenuated by the distance it must travel to get there in the first place. Additionally, it's only inwardly off-axis reflections from the opposite wall that ever get to the listener. Those from on-axis have an angle of incidence to the opposite wall of 45°, are reflected at 45° additional, for an included angle of 90°, and thus bypass any listener(s) within any distance closer than as much as 3 x the "height" of the crossfire triangle.

Draw it on paper and you'll see. Try it with CD speakers, and you'll hear. As described by Eargle, precedence compensation further enhances this "trompe l'oreille." :yes:


Are you saying mount these vertically to get the desired effect?No, normal orientation, but toe-in 45° each side, and you listen from outside the triangle; they crossfire in FRONT of the listener. The vertical axis, you choose according to the relative heights of waveguide and listener, just as you suggest.


BTW I ordered a pair of these to "play" with. The dance5 system uses the perf metal grills with foam and I can make grills just like them to get the CD effect(?).I'll be very pleased to learn that a foam-backed perforated metal grille is all that is required to accomplish this.... :thmbsup:

[Robh3606]

Thus, you are, in effect, normalizing the SPL from both speakers to lateral movement in either direction, and the phantom image stays in the same location, midway between the speakers, replicating the effect Everest I, S2600, and S3100 achieve with their defined directivity horns and minimal, if any, toe-in.

That's understood and the easy part, you can do that with any CD horn like you say.


may produce an improved Everest I effect.... :yes:


I don't get why?? The 2405's were aimed and crossed well before the Everest's larger horn. Looking at the geometry I can understand how they were able to use them to help compensate for the Hass effect like you noted. This makes the Everest system unique.

I don't understand why these horns would be better than any other CD horn such as the ones used in the 2600/3100 set-up like you described.

What makes these horns unique, that's what I am not getting???

Rob:)

[Zilch]

I don't understand why these horns would be better than any other CD horn such as the ones used in the 2600/3100 set-up like you described.

What makes these horns unique, that's what I am not getting???S2600/3100 did not have precedence compensation*, and thus, if Eargle was correct regarding its importance, were compromised from the outset.

In Everest I there were phase and time-alignment issues associated with the separate UHF driver and the deeply offset horn driver. I did not make the "festival" thing up of whole cloth.

In 2374/84, the precedence comp, if it's correct or even merely workable, is integral to a significantly shallower waveguide, both advantageous.

I don't consider the toe-in a detriment, either. All of those earlier products used it, too, just not done to this extreme.

*Well, I don't recall them having any response "anomaly" that would provide it, though I admit I was never looking for such. If anyone could describe what that might be, I could re-measure to verify, I suppose, since I'm sitting here looking at all of these horns, presently.... :yes:

[kegger]

So Zilch, now that you've got "some" of the technical worked out how do they sound?

Do they stack up to some of the better horns you've heard?

Are there any subjective qualities that come out?

[Zilch]

So Zilch, now that you've got "some" of the technical worked out how do they sound?I'm going to have to clear sufficient space to set up the pair properly, Kegger. I will probably mate them with B380 bass bins which are what I originally used to evaluate the Everest I and S2600/3100 horns.

What I am hearing presently from a single on the test stand is clean and clear with "big horn" transparency, even though I am only crossing at 1.2 kHz. Listening for it, I do hear the SSC™ compensation as a slight change in timbre moving on-axis, but it may be more prominent to others who hear VHF better; I know it's there.

Needless to say, the imaginary gonadotrophic effect is huge -- these babies BEG to be cranked.... :rockon1:

[Russellc]

Following with interest here Zilch,

Thanks for the headsup, I certainly am looking into these, :dancin:
Quite interested here as well....:blah:

Russellc

[kegger]

Hey Russ, how yu doin,,....:)

--------------------------------

Well I gots a few bills and what not to take care of then I am planning on
getting a pair of these big ole horny guy's. Thanks for the thoughts Zilch.

Not helping that someone put a pair of S99's on the bay neer by here, with
local pickup only, I won um for $100, so I got some LE14's comin, whoo hoo!

(Thinking can I fit a small wavegiude "like the 18sound" in the S99 cab). . humm
Yah then cut out the bottom and "extend" the cab, Zilch your creating monsters!

[kegger]

Got through to JBL Pro Parts,

The 1" thread on version is $131.00 each, and I was told there were plenty in stock. http://www.audiokarma.org/forums/images/smilies/thumbsup.gif

Unfortunately I did not ask about the 1.5" bolt on version, http://www.audiokarma.org/forums/images/smilies/no.gif

[Zilch]

1.5" Bolt-on, $113 MSRP.... :thmbsup:

[Zilch]

Transparent shade selectively attenuates 2384 VHF on-axis:

[kegger]

Any idea how low they might be able to be crossed Zilch?

Just wondering about there possable suitabilty in a 3-way using them for mids.

[Zilch]

Any idea how low they might be able to be crossed Zilch?

Just wondering about there possable suitabilty in a 3-way using them for mids.Being "Optimized Aperture," 800 Hz nominal for the 1.5" version.

My initial measurements, above, show somewhat lower may be possible.

The 1" throat variant may be a different story, as it has an actual throat.... :dont-know

[kegger]

I'm not familiar with how the throat changes things Zilch, are you suggesting
the 1" variant with it's smaller throat "may" go lower by virtue of the throats
size just being smaller? Care to elaborate?

Yes still a lot to learn about horns here! :)



Being "Optimized Aperture," 800 Hz nominal for the 1.5" version.

My initial measurements, above, show somewhat lower may be possible.

The 1" throat variant may be a different story, as it has an actual throat.... :dont-know

[Zilch]

http://www.jblpro.com/catalog/support/getfile.aspx?docid=298&doctype=3

[kegger]

Thanks, I'm reading, :applaud:

[Zilch]

We have many horn theorists here who can fill in the details, but, generally, how low a horn/waveguide will play well varies with the mouth size. 2374/84 certainly has that part down.

With respect to the rest of it, I defer to those others who actually know something.... ;)

[Zilch]

BTW I ordered a pair of these to "play" with. The dance5 system uses the perf metal grills with foam and I can make grills just like them to get the CD effect(?).YO, Subwoof!

Did you get this to happen?

Second generation, now, with 2432H drivers.

[Renewed interest in these on another forum.... :yes: ]

[Mostlydiy]

Any news on how low 2384 will go? Would it be possible to cut them off to meet my midbass horns at 500? If I go with the 2384 they will be mounted in a wall so the big baffle might help it on its way down to 500? Any thoughts on the difference betweem 2352 and 2384 soundwise?

/Mostly