Revisiting "Imaging"

[Robh3606]

[QUOTE=Ian Mackenzie;446895]Hi Rob,

You posted the Array was not time aligned when in fact JBL literature says it was.

“ The Arrays are not time aligned and yet they manage to image well as an example.


It was misinformation while casting your own interpretations. That just confused everyone reading your post. That is why l posted the attachments.

The only facts relevant are about original design as outlined in the attachments.
I unfortunately don’t have time to go through the rest of your post right now.

The same thing applies to anyone attempting to opinion their bias on a particular design as to whether they believe it or not? What that does it pollutes the whole thread with half arsed polarising opinions which then makes others very reluctant to post about their experiences with that system or something else. Why would you bother?

This isn’t your bottom of river US political campaign.

By definition they are not time aligned.

It is not misinformation.

The step response cannot be denied.

As far as half arsed polarizing opinions

Just read one.

Rob

[Ian Mackenzie]

[QUOTE=eso;446891]

Hi Rob,

Here attached is far more insightful and relevant information from the Project Array Designer himself.

See the attached documentation

https://support.jbl.com/on/demandwar...20Brochure.pdf




Being fully adjustable was a primary requirement for the Cogent main arrays.

Early on Rich and I worked on locating the balance points of each horn/driver assembly and mounting those assemblies around their balance points while having the drivers physically aligned became the basic layout for the racks.

Zero position was suspending the mid horn and driver at it's balance point with a pivot and struts that allowed a good +/- range on the vertical axis. The midbass assembly is able to move forward or backward to maintain physical alignment with the mid while moving up or down as needed to focus with the mid and the desired distance.

Finally the tweeter needed to be able to move in yet another axis to align with both of the other drivers. Sourcing all of the various hardware took a while, and once we had it Rich's machining know how was needed to adapt everything.

And then other features just help to make a cohesive system. One side of the racks distributed the signal wiring and the other side fed the field supplies. Crossover network mounted in the base of the rack.

I made a variation on these racks for my system here. I needed something shorter

eso

Hi eso,

Beautiful work.

I wish l could organise this.. really clever. I found the Tad 4003 really too heavy to balance on the Bill Woods A700. I’ve since moved the Tad’s on and plan to use a lighter Neo 1.5 inch driver. They are amazing horns. It’s all about the way the wave front travels through them. They are the last ones Bill made apparently. Steve Schell put me onto Bill. I have negativity engineered the OMA Monarch as a diy project in Hornresp. It has a throat with an acoustic high pass filter on the bottom woofer to avoid comb filtering. It’s about 105 db from 35-700 hertz with dual AE 15M woofers. I must revisit this project.

[Ian Mackenzie]

Hi Rob,

I have attempted to interpret what your technical explanation about.

I honestly don’t understand that your trying to get across?

Your coming at this from a very specific point of view. But where’s the relevance and all the supporting data. There’s no documentation so l don’t understand. There’s no labelling so l don’t see how you can expect me to comprehend this?

The best thing to do is post this another thread and provide a comprehensive coverage of the whole thing.

Btw the commentary from Greg only refers to the horn and the woofer.

This thread is not deep dive into a particular loudspeaker. But you seem to think you have a point to make. Do it in another thread.

[BMWCCA]

[Robh3606]

Hi Rob,

I have attempted to interpret what your technical explanation about.

I honestly don’t understand that your trying to get across?

Your coming at this from a very specific point of view. But where’s the relevance and all the supporting data. There’s no documentation so l don’t understand. There’s no labelling so l don’t see how you can expect me to comprehend this?

The best thing to do is post this another thread and provide a comprehensive coverage of the whole thing.

Btw the commentary from Greg only refers to the horn and the woofer.

This thread is not deep dive into a particular loudspeaker. But you seem to think you have a point to make. Do it in another thread.

Well lets keep this simple.

The 2 systems you referenced are time aligned the Dunlavy SC-IV and the Wilson Alexx V

Since this is a thread about imaging and most speakers are not time aligned it seemed prudent to point this out. Time alignment means all drivers arriving at the same time to the listening position.

I am not making this about a specific speaker. It's really about two different design philosophies.

You don't need to have time alignment to image well. It may help but it's not an absolute.

The Array is not and you can see this in the step response. All 3 drivers have very clearly separate arrival times. Tweeter, Mid and Woofer in that order.

The Wilson is better but has a driver out of phase so the negative spike

Finally the Dunlavy which gives you a perfect step response where all drivers are in phase, shown as a single triangular spike, where you cannot see the drivers individually.

The figures are Array Wilson and Dunlavy in that order.

Rob

[Ian Mackenzie]

I don’t see any labels or the source of your data?? No proof or evidence. Fail.

Were your school assignments this vague? I hope this is a joke.

I have attached the essentials of the Array 1000 so anyone looking can make sense your quandary. In my earlier posts l referred to the designer’s comments. They stand.

The matter of the tweeter isn’t relevant because it wasn’t mentioned by the designer or by me. I will go over that latter ….yawn. It’s 1.36 am.

The primary crossover point is the woofer and the horn. Technically and in practice to enable a smooth crossover transition in the horizontal axis and through the vertical axis (over a defined angle) this is best done by alignment of the woofer and the horn acoustic centres. That way the Linkwitz 24 db acoustic crossover transformation (per JBL document) can be achieved with both drivers in acoustic and electrical phase (per the designer).

The main mission is to keep any nasty nasty dips in the vertical polar pattern outside the defined vertical listening window at the crossover point.

So the designer tweaks this with the horn by testing various positions until a satisfactory compromise is met. Now the location of the horn acoustic centre may NOT be exactly in alignment of the woofer. The (business ) impact on the smoothness of the crossover region takes precedence over minor ( and l mean minor) time alignment discrepancies.

To explain the 750 hertz crossover point as a wavelength of 458 mm. A half wavelength is 229 mm which is the point of a 180 degree phase shift. Now if the horn acoustic centre was 15 mm behind the woofer to obtain the best vertical polar response this would equate to an 11.7 degree phase shift. Relating this 15 mm the actual time delay looks like this. A 1 ms delay is 0.344 of a metre. Therefore a 15 mm is 0.04 ms. This is an example of what this is all about. If it was 150 mm that is a 111 degree phase shift. Referring to the previous example that is about 0.33 of a ms. Well below the BL group delay threshold at 750 hertz of 2.5 ms.

According to your ancient BL Threshold for group delay it’s 2.5 ms at 750 hertz. Therefore l suggest your technical measurement is irrelevant.

Now the tweeter horn. I don’t have the drawings of the tweeter placement but let’s say it 150mm forward of the compression driver. Referring back to your BL group delay threshold curve at the 8000 hertz tweeter crossover point the threshold is about 1.8 ms. That expressed as a distance is 0.6192 of metre. Is the tweeter time alignment therefore important? No it isn’t.

A pertinent point is that the tweeter is there to improve the subjective performance of the stand alone horn above 8000 hertz. In JBL language everything is done for a good reason. It has a relatively narrow vertical polar pattern. This means it unlikely to be heard when the listener is standing up. To an extent the horn vertical polar pattern will fall off when the listener stands up. This is of course dependent on the listener distance.

For all intents and purposes this is a time aligned system in so far as the BL group delay threshold is concerned. If this system was very precisely time aligned then it would seriously impact on other audibly performance criteria.

If the horn crossover point was much higher like 2 or 3 khertz then everything would need to be re evaluated.

It’s important to point at that a high frequency resonance or diaphragm break up an occur in a 0.1 ms to 6 ms range. Are these things audible ? Yes they can be and that’s why little horn tweeters are sometimes used.

[Robh3606]

I don’t see any labels or the source of your data?? No proof or evidence. Fail.

Were your school assignments this vague? I hope this is a joke.

I have attached the essentials of the Array 1000 so anyone looking can make sense your quandary. In my earlier posts l referred to the designer’s comments. They stand.

The matter of the tweeter isn’t relevant because it wasn’t mentioned by the designer or by me. I will go over that latter ….yawn. It’s 1.36 am.

The primary crossover point is the woofer and the horn. Technically and in practice to enable a smooth crossover transition in the horizontal axis and through the vertical axis (over a defined angle) this is best done by alignment of the woofer and the horn acoustic centres. That way the Linkwitz 24 db acoustic crossover transformation (per JBL document) can be achieved with both drivers in acoustic and electrical phase (per the designer).

The main mission is to keep any nasty nasty dips in the vertical polar pattern outside the defined vertical listening window at the crossover point.

So the designer tweaks this with the horn by testing various positions until a satisfactory compromise is met. Now the location of the horn acoustic centre may NOT be exactly in alignment of the woofer. The (business ) impact on the smoothness of the crossover region takes precedence over minor ( and l mean minor) time alignment discrepancies.

To explain the 750 hertz crossover point as a wavelength of 458 mm. A half wavelength is 229 mm which is the point of a 180 degree phase shift. Now if the horn acoustic centre was 15 mm behind the woofer to obtain the best vertical polar response this would equate to an 11.7 degree phase shift. Relating this 15 mm the actual time delay looks like this. A 1 ms delay is 0.344 of a metre. Therefore a 15 mm is 0.04 ms. This is an example of what this is all about. If it was 150 mm that is a 111 degree phase shift. Referring to the previous example that is about 0.33 of a ms. Well below the BL group delay threshold at 750 hertz of 2.5 ms.

According to your ancient BL Threshold for group delay it’s 2.5 ms at 750 hertz. Therefore l suggest your technical measurement is irrelevant.

Now the tweeter horn. I don’t have the drawings of the tweeter placement but let’s say it 150mm forward of the compression driver. Referring back to your BL group delay threshold curve at the 8000 hertz tweeter crossover point the threshold is about 1.8 ms. That expressed as a distance is 0.6192 of metre. Is the tweeter time alignment therefore important? No it isn’t.

A pertinent point is that the tweeter is there to improve the subjective performance of the stand alone horn above 8000 hertz. In JBL language everything is done for a good reason. It has a relatively narrow vertical polar pattern. This means it unlikely to be heard when the listener is standing up. To an extent the horn vertical polar pattern will fall off when the listener stands up. This is of course dependent on the listener distance.

For all intents and purposes this is a time aligned system in so far as the BL group delay threshold is concerned. If this system was very precisely time aligned then it would seriously impact on other audibly performance criteria.

If the horn crossover point was much higher like 2 or 3 khertz then everything would need to be re evaluated.

It’s important to point at that a high frequency resonance or diaphragm break up an occur in a 0.1 ms to 6 ms range. Are these things audible ? Yes they can be and that’s why little horn tweeters are sometimes used.

OK so basically JBL is not in the time alignment camp like most other manufacturers. You just admitted it in your explanation.

Like I said in my original post. As long as the design meets the B+L graph I originally posted. That means below the curve not at 2.5 Msec at 750Hz where it becomes audible.

Like the 3 systems graphed that are below the threshold.

Those are Stereophile step response graphs buy the way.

Why don't you just admit the 1400 is not time aligned/coincident as in the review? What's the big deal?

Text from the Array Review

"In the time domain, the 1400 Array's step response on the tweeter axis (fig.8) indicates that all three drive-units are connected with positive acoustic polarity, and that the tweeter output arrives at the microphone half a millisecond before that of the midrange, which in turn arrives half a millisecond before that of the woofer.

This is definitely not a time-coincident design, though the fact that the ear/brain does integrate arrivals over a longer period than 1ms should mean that this won't matter much.

LG was impressed by the stability and accuracy of the JBLs' imaging, which you'd think might be adversely affected by the lack of time coincidence.

But as far as the lower-frequency units are concerned, the 1400 Array's step response is at least time-coherent, in that the overshoot of the midrange unit's step smoothly leads into the woofer's step.

This suggests an optimal crossover implementation."


The Key is optimal crossover implementation which again I said in my post. Smooth driver integration and good polar response being the main focus.

Essentially all well designed systems are time-coherent. There is a difference they are not the same.

Rob

[Mr. Widget]

Well lets keep this simple.

The 2 systems you referenced are time aligned the Dunlavy SC-IV and the Wilson Alexx V

Since this is a thread about imaging and most speakers are not time aligned it seemed prudent to point this out. Time alignment means all drivers arriving at the same time to the listening position.

I am not making this about a specific speaker. It's really about two different design philosophies.

You don't need to have time alignment to image well. It may help but it's not an absolute.

The Array is not and you can see this in the step response. All 3 drivers have very clearly separate arrival times. Tweeter, Mid and Woofer in that order.

The Wilson is better but has a driver out of phase so the negative spike

Finally the Dunlavy which gives you a perfect step response where all drivers are in phase, shown as a single triangular spike, where you cannot see the drivers individually.

The figures are Array Wilson and Dunlavy in that order.

Rob

Thanks Rob!

While it is true JBL did some time alignment work on the 1400 Array... physically pushing the mid/tweeter array forward, it is obvious from the impulse response that they needed to move it further. That said, for practical reasons it is also obvious why they wouldn't want to do that.

That Dunlavy impulse response is VERY impressive... but I bet it is only achieved when measuring on tweeter axis vertically and horizontally.

I don’t see any labels or the source of your data?? No proof or evidence. Fail.

Were your school assignments this vague? I hope this is a joke.

Dude... what is your problem? Geez!


Widget

[eso]

[QUOTE=Ian Mackenzie;446898]

I found the Tad 4003 really too heavy to balance on the Bill Woods A700. I’ve since moved the Tad’s on and plan to use a lighter Neo 1.5 inch driver. They are amazing horns. It’s all about the way the wave front travels through them. They are the last ones Bill made apparently. Steve Schell put me onto Bill. I have negativity engineered the OMA Monarch as a diy project in Hornresp. It has a throat with an acoustic high pass filter on the bottom woofer to avoid comb filtering. It’s about 105 db from 35-700 hertz with dual AE 15M woofers. I must revisit this project.

Bill is a brilliant man.

When we first started experimenting with the conical horns Bill had cast some of his throats for the Cogent drivers (their throat is a unique size, ~2.7"). But using Bill's throat design limits the petal thickness to 3/4". For purely aesthetic reasons I thought the petals should be thicker to visually balance with the mass of the drivers. So I asked rich to machine an adapter that bolts to the drivers with a ring with set screws mounted in the horn. I think it's a much cleaner attachment and the horn mass is a nicer balance.

And also this allowed the previously mentioned tweak Rich and I implemented on the mid horns. I know there's a thread here about horn/driver interactions at the throats and volumes elsewhere as well. We were seeking to make a perfect match for our efforts.

ΩMA still uses the 3/4" petals with Bill Woods' throats now badged ΩMA.

eso

[rusty jefferson]

Thanks Rob!

While it is true JBL did some time alignment work on the 1400 Array... physically pushing the mid/tweeter array forward, it is obvious from the impulse response that they needed to move it further. That said, for practical reasons it is also obvious why they wouldn't want to do that.

That Dunlavy impulse response is VERY impressive... but I bet it is only achieved when measuring on tweeter axis vertically and horizontally.....


Widget

Excellent example of everything in speaker design being a compromise. 1400 are really good sounding speakers. Appreciate the knowledge Rob.

I've never seen a speaker with passive networks have an step response like that, only active. I'll have to read that article.

[toddalin]

I've never seen a speaker with passive networks have an step response like that, only active. I'll have to read that article.

https://i.ebayimg.com/images/g/-SMAA...a/s-l1600.webp

[jmpsmash]

Very interesting discussion!

I am on the time alignment camp. Hard to achieve given all drivers are of different shape and depth. One would have to do what Dunlavy did, or have to resort to waveguide, or DSP.

When I was playing around with the 2397/2441/2216ND1 (still remember that thread?), I was just a beginner and couldn't figure out why I was never able to get clean FR like I see on many speakers. Then over the years I spent a lot time learning about crossover design and then I realized the depth of the horn caused the acoustic center to be way far behind the woofer. And with computer tools, this can be easily simulated. The following are 2 theoretically ideal driver, originally perfectly aligned but was pushed so that they are one crossover wavelength (2000Hz/17.1cm) apart. Still phase aligned but not time aligned. The excess phase causes big ripples in the FR.

For anyone who has DSP active filters, it is very quick to do a similar experiment. change the delay to keep the drivers phase aligned but not time align, listen to it, and then change it again to be time aligned, the image will suddenly become much more defined.

[Robh3606]

https://i.ebayimg.com/images/g/-SMAA...a/s-l1600.webp

Have you seen any step response measurements on them? I had an 811C as a center for years and it was a great vocal speaker.

Rob

[toddalin]

Have you seen any step response measurements on them? I had an 811C as a center for years and it was a great vocal speaker.

Rob

No, but I remember seeing one opened up on C/L and it was packed with caps and chokes.

Also, there are these for sale in Long Beach.

https://images.craigslist.org/00l0l_...2_1200x900.jpg

https://images.craigslist.org/00P0P_...M_1200x900.jpg

There is also a very good chance that my Super Big Reds are time aligned through the dedicated electronic crossover. But they still don't image like the Mermans. Maybe they would if located in the same places.

[1audiohack]

....The Array is not and you can see this in the step response. All 3 drivers have very clearly separate arrival times. Tweeter, Mid and Woofer in that order.

The Wilson is better but has a driver out of phase so the negative spike

Finally the Dunlavy which gives you a perfect step response where all drivers are in phase, shown as a single triangular spike, where you cannot see the drivers individually.

The figures are Array Wilson and Dunlavy in that order.

Rob

Hello Rob;

The Dunlav response is an ETC measurement, unlike the two above that are impulse responses. There is no return through zero in an Energy Time Curve display. Dominated by the HF energy it looks great but it is not at all a fair visual comparison.

Barry.