Line Array Speakers : Modular, Expandable, Out-of-the-Box Design for in-home use

[benanders]

Double no.

Lol okay.

Incidentally, I was playing with XDir for the two woofers (per speaker monitor), and realized size does not seem to be a factor. Either I’m missing something, or that could be an issue for accuracy of the software’s predictions.

If graph lines are the most important factor in one’s choice of effective speakers, seems this calculator could be pretty misleading in some cases?

 

[benanders]

Ghosted ( of Christmas Past? ) thread, so posting for my own edification after my usual holiday absence from ASR (off work = no work breaks, = busy, or busy listening to music; no threads

I sent out a holidays card to some local audio friends. It read: “My tree is 8 feet tall, but it sounds way taller!”

 

[Divasson]

Hi, just saw your magnificent line array and read the conversation.

Regarding lack of crossover: the tweeter is feed by a transformer to raise the voltage, in series with 10 uF capacitor network before the transformer. However: the tweeter is a pure capacitor itself, measuring 0,10 uF, so the capacitor network acts here as a voltage divider, not a "proper" high pass filter. I'd still call it a crossover.

The woofers are direct to the input, so no crossover here at all. ("woofer" for a 3" driver may sound optimistic)

I am using them (not the whole line array, but the Stradas 2 as a stereo pair) on Mapleshade stands, crossing with the subwoofer at potentiometer midpoint with very good results. The spatial prowess of the ensemble (3D images and immersive soundstage) are better than any other brand I have tested. However, these would always be weak on the mid bass area, regardless of subwoofer setup (IMHO, of course)

I see in the background some Gallo Reference 2: the best overall speaker I had. If you are in France, are you willing to sell them? (Je suis à Saillagouse, au 66)

 

[benanders]

Hi, just saw your magnificent line array and read the conversation.

@Divasson thank you. And welcome to ASR.

Regarding lack of crossover: the tweeter is feed by a transformer to raise the voltage, in series with 10 uF capacitor network before the transformer. However: the tweeter is a pure capacitor itself, measuring 0,10 uF, so the capacitor network acts here as a voltage divider, not a "proper" high pass filter. I'd still call it a crossover.

Seems to be the popular consensus. No argument from me.

The woofers are direct to the input, so no crossover here at all. ("woofer" for a 3" driver may sound optimistic)

Ha! Fair enough. We can increase to 4” if you count the surrounds…

I am using them (not the whole line array, but the Stradas 2 as a stereo pair) on Mapleshade stands, crossing with the subwoofer at potentiometer midpoint with very good results.

Yes - the 12:00 position of the subwoofer crossover works well for me, too. For stereo pair and also 6x per channel Strada.

The spatial prowess of the ensemble (3D images and immersive soundstage) are better than any other brand I have tested. However, these would always be weak on the mid bass area, regardless of subwoofer setup (IMHO, of course)

Yes, integration between stereo Strada 2 and TR3D subwoofers can be an issue. Have you tried elevating the subwoofer(s) above the floor by 30-45 cm?

I see in the background some Gallo Reference 2: the best overall speaker I had. If you are in France, are you willing to sell them? (Je suis à Saillagouse, au 66)

The Reference 2’s are great speakers. The line arrays are quite different in character from the three main Gallo designs (stereo Strada 2, Reference 2, Reference 3). I do not wish to sell the Reference 2’s, but if it’s consolation to you, the speakers are located on the other side of the planet from Europe, anyway.

 

[songOVERsound]

Don’t beat me up too hard over this..

 

[thewas]

View attachment 279122Don’t beat me up too hard over this..

That's not a line array.

 

[songOVERsound]

That's not a line array.

Can you explain why they aren’t and what they would need to be line array? Thanks

 

[thewas]

Can you explain why they aren’t and what they would need to be line array? Thanks

In few words it needs several mid/high drivers spaced closely vertically over each other to create a narrow vertical radiation while that PMC has only one mid driver and one tweeter.
In detail https://en.wikipedia.org/wiki/Line_array

 

[songOVERsound]

In few words it needs several mid/high drivers spaced closely vertically over each other to create a narrow vertical radiation while that PMC has only one mid driver and one tweeter.
In detail https://en.wikipedia.org/wiki/Line_array

Ahhh so the PMC’s are, if anything, closer to a line source rather than a line array?

 

[thewas]

Ahhh so the PMC’s are, if anything, closer to a line source rather than a line array?

Rather a point source approximation with spacely distributed woofers for better room excitation in the bass region.

 

[Vladimir Filevski]

That title…
Sounds impossible, right?
Possible, and sounds right.

The project has spanned some years of (1) excessive backlit screen reading, (2) many in-person listening sessions to diverse high end systems, and (3) defining my preferences/expectations, then pursuing them with this particular speaker model (Gallo Acoustics Strada 2) through haphazard trial and error. The system may look finished, but it’s not. So let the judgement begin on where else to begin

Nice looking speakers, but that symmetrical (and large) distance between woofers is not optimal acoustically. If you bring Strada 2 modules vertically closer together as physically as possible (woofer spheres almost touching), then center-to-center spacing between woofers and tweeters will be smaller and likely to produce better results. The symmetrical visual beauty of the speakers will be lost, but it is a small price to pay for better sound quality.
You have to invest in measuring microphone, and time to learn free ARTA (or REW) mesuring software and VituixCad simulation software - measurements and simulation are infinitely faster than trial-and-error method.

 

[benanders]

Nice looking speakers, but that symmetrical (and large) distance between woofers is not optimal acoustically. If you bring Strada 2 modules vertically closer together as physically as possible (woofer spheres almost touching), then center-to-center spacing between woofers and tweeters will be smaller and likely to produce better results. The symmetrical visual beauty of the speakers will be lost, but it is a small price to pay for better sound quality.
You have to invest in measuring microphone, and time to learn free ARTA (or REW) mesuring software and VituixCad simulation software - measurements and simulation are infinitely faster than trial-and-error method.

Thanks for your comments @Vladimir Filevski . My thinking was same as yours when I assembled my second version, with four monitors per channel. I spaced neighboring spheres precisely 1.5mm apart, thinking it best for theoretical coupling of as much of each 4” driver’s frequency band as possible.

The sound was dull, flat, whatever word best conveys lack of perceived dynamics (relative to the first “line array” of two monitors per channel).

Despite seeming unlikely, the paired 4” drivers on each monitor interact according to the length of the tweeter between them.
That’s why I put the gap between each monitor in this version - tweeter present or not, the space between the 4” drivers must exist.

Main reason I haven’t bothered to mic and DSP things yet is because I haven’t felt it necessary. A friend did this with his equipment early on for me, and while it made things sound a bit more detailed in a small area, the overall effect of the speakers and their ability to address the whole room decreased.

I get how line arrays generally demand EQ even more (obviously) than traditional point sources to perform correctly and thus sound “correct,” but in this case all first-hand perceptions as-is, have suggested very little need. “Lucky strike,” maybe.

 

[Vladimir Filevski]

Comparison must be taken with only one parameter changing - the same number of Strada 2 modules/monitors, while changing the distance between them!
Six monitors with 4" distance compared to four monitors with 1.5 mm distance have two different parameters changed. Which parameter is more relevant to the sound quality - the sheer number of Strada 2 monitors (6 compared to 4, which is 50% increase!), or the different distance (4" compared to 1.5 mm)?
Actually, there is a third parameter - the height of the line array! Line array comprised of 6 monitors with 4" distance is 47" taller than the shorter version of 4 monitors with 1.5 mm distance. Taller line array is always better!
I would speculate that 6 (or 7) monitors with 1.5 mm distance may sound better than your current version of 6 monitors with 4" distance. Of course, measurement and simulation are needed for optimal design.

 

[JRS]

As suggested by the software analysis (thank you @voodooless), the Strada 2 monitor should sound problematic. NRD test results also support that prediction.

But, Strada 2 (stereo pair) sounds excellent at high frequencies, despite any contradictory performance test predictions / results. Seems an inexplicable (truly not explicable, grin) discrepancy between what’s on paper and what’s in reality. I don’t profess to understand why, but the listening experiences among reviewers/owners in the past decade remain consistently positive. That’s partly why I chose to try a line array with the Strada 2 (wasn’t exactly my idea first, if you search Gallo 5LS…).

WHY a Line Array INside??

I wanted two main things from what could be an ever-evolving project.

(1) Very large / tall sonic image and believable dynamics achieved without blasting SPL. IOW, a convincing wall of sound.
(2) Service a long room at both sitting and standing heights, again without blasting SPL.

Even the crudest big line array should permit those two goals, but life is not just about wants; I had several logistical restrictions, too (none of which must apply in a dedicated-forever listening room).

(1) Expandable - add / remove units as needed based on ceiling heights.
(2) Modular - disassemble / pack away in little space and, preferably, ship easily, too.
(3) Non-obstructive to front wall view - no sense in blocking a big window (see attached pic).

Again - project not finished, may never be and doesn’t need to be. But baby-stepping in the right direction for my goals.
Software testing? I’ll get there yet!
Experimentally-sound ( = statistically robust) qualitative sampling? WIP and can elaborate, if of any interest.

Regarding integrative evidence (both empirical and qualitative) supporting indoor LA’s, has anyone else looking here actually read every page and associated links of @Wesayso diyAudio thread? Don’t attempt it in an afternoon - it works better for those of us who’ve tagged along for years.

I look at that and wonder why not orient horizontally and dtavk. Cost a fortune but with level taper and good EQ could be a killer. I'd add some (1 to 3 ) true tweeters in the center to fill end the highs.

 

[benanders]

From what I typed before, I see how you assume those issues @Vladimir Filevski . Thank you for detailing them. Here’s some clarifying:

Comparison must be taken with only one parameter changing - the same number of Strada 2 modules/monitors, while changing the distance between them!

Correct, and I did it this way with 2x Strada per channel and 4x Strada per channel. I did not try it with 6x Strada per channel after consistent results in 2x and 4x setups.

Six monitors with 4" distance compared to four monitors with 1.5 mm distance have two different parameters changed. Which parameter is more relevant to the sound quality - the sheer number of Strada 2 monitors (6 compared to 4, which is 50% increase!), or the different distance (4" compared to 1.5 mm)?

Good question.

2x Strada per channel with gap sounded better than 4x Strada per channel without gaps. Half as much driver surface area in the 2x Strada (with gap) shouldn’t be better unless a chunk of frequencies had destructive interference in the 4x Strada (without gaps).
Of course, there are still insurmountable confounding factors (e.g. impedance, outermost drivers’ spacing from vertical boundaries, etc.) in that type of comparison.

Actually, there is a third parameter - the height of the line array! Line array comprised of 6 monitors with 4" distance is 47" taller than the shorter version of 4 monitors with 1.5 mm distance. Taller line array is always better!

Despite the need to alter one variable at a time, it’s not possible with indoor LA driver spacing. You basically say as much. Changing one variable can automatically alter another, too. Different spacing between drivers alters LA length (height), which in turn shifts how close the outermost drivers are from vertical boundaries.

I would speculate that 6 (or 7) monitors with 1.5 mm distance may sound better than your current version of 6 monitors with 4" distance. Of course, measurement and simulation are needed for optimal design.

Regarding the dogma of nearest possible driver spacing in LA’s, I get it, heck I shared your expectations, but look at it this way: if spacing Strada’s 4” drivers as closely as possible gave better sound, then the design of the Strada should have the two spheres close together and the tweeter mounted above or below the spheres, instead of in a gap between the spheres.

Using 3, 5 or 7x Strada units per channel is not an option for a passive speakers setup: wiring / impedance constraints.
I tried a semi-powered speaker setup, driving the outermost monitors (top/bottom unit of 6x per channel) with a second amplifier. The electrical clutter-to-sound quality ratio didn’t offer noticeable advantage. However, if I find a more suitable compact power supply to replicate x6 or x12, I’ll be trying again. That would offer better flexibility on # Strada per channel and reduce room height-related boundary/driver spacing issues, among other potential benefits (e.g. power tapering without DSP).

 

[benanders]

I look at that and wonder why not orient horizontally and dtavk. Cost a fortune but with level taper and good EQ could be a killer. I'd add some (1 to 3 ) true tweeters in the center to fill end the highs.

If you mean rotate each monitor 90°, can’t do it. Would restrict the tweeters’ dispersion too much ( ~30° vertical by design), and cause serious lobing in their now-vertical plane ( ~180° horizontal by design). I suspect having the horizontally-paired mid drivers would also lend too much directivity / beaming, but they might tame sidewall issues, too.

The old manufactured version of my modular copycat (the 5LS) did have a tweeter in every gap between mid driver pair of spheres. I was worried fewer tweeters in mine would be a problem, but surprisingly, no one is noticing so far.

When I was toying with early mock-ups, I did muse what the company’s horizontally-configured “center-channel” monitor would look like in a LA configuration. The image shows the spheres much too tightly spaced; it looks like a monster and indeed would cost a lot!

 

[Vladimir Filevski]

Correct, and I did it this way with 2x Strada per channel and 4x Strada per channel. I did not try it with 6x Strada per channel after consistent results in 2x and 4x setups.

If I understand this correctly, you tried 4 x Strada in one column, varying the spacing between between drivers only? Then, it is indeed the correct way to determine which variant gives the best sound.
Although, variant which gives the best sound for 4 x Strada (or 2 x Strada) doesn't necessarily make it the best variant for 6 x Strada, because you are moving (slowly) from "point source" territory into the "line array" territory. Of course, there is no fixed point when one becomes the other, because neither 2 x Strada is a "pont source", nor 6 x Strada is a "line array".

Regarding the dogma of nearest possible driver spacing in LA’s, ...

Well, that is not a dogma - but it refers to real line arrays only! Your 6 x Strata is not a real line array, nor a real column speaker. Nevertheless, you may get good sound from your loudspeaker even if it doesn't belong to some defined category.
To become the real line array, one loudspeaker system must fulfill several strictly defined necessities, which are impossible for (multiple) Strada 2 modules, whatever the spacing is. Bearing that in mind, picture above (horizontal Strada Centres, stacked) is not a line array, because of too high crossover frequency. It may become one, only if "shading" of modules is used.
These are theoretical/practical musings only, I wish you many hours enjoying the sound of your loudspeakers.
If you can make frequency measurements in your room, it will be very interesting for the majority of the members here.

 

[benanders]

If I understand this correctly, you tried 4 x Strada in one column, varying the spacing between between drivers only? Then, it is indeed the correct way to determine which variant gives the best sound.
Although, variant which gives the best sound for 4 x Strada (or 2 x Strada) doesn't necessarily make it the best variant for 6 x Strada, because you are moving (slowly) from "point source" territory into the "line array" territory.

Correct. Iterations included 2x Strada (no gap; with gap), 4x Strada (no gaps; with gaps), and 6x Strada (with gaps only - it was a lot of precision drilling by that point, ha! ).
The deciding comparison was how dynamic 2x Strada with gaps seemed, compared to 4x Strada without gaps. In both cases of comparing an equal number of monitors per channel (2, 4), with gap(s) also sounded better than without gaps.
All comparisons were still flawed to their core for my need to re-mount / re-wire / reconnect everything between “samples,” but I’m okay with that.

To become the real line array, one loudspeaker system must fulfill several strictly defined necessities, which are impossible for (multiple) Strada 2 modules, whatever the spacing is.

This brings up an uncertainty - one that confounds my understanding of how the Strada 2 drivers deliver their “goods.”
Conceptually (and where my thinking is now), since the two 4” drivers of a Strada 2 should be coupling (at least partially), those unlikely physics should also extend to additional drivers located above/below the Strada 2. Hence my use of the phrase line array, despite the inconsistencies you point out.

The Gallo M-T-M arrangement, dimensions-wise, dates back nearly 20 years to the first Reference 3 model; seems everything from then until ~2012 had drivers roll off ~2 kHz and tweeters pick up there.
Strada 2 changed the driver and pushed its roll off up to ( !!! ) 6 kHz. I was skeptical of the shift to 6 kHz, and understood why some folks refused to even try Strada 2 (vs. their previous Gallo models).
Two 4” drivers spaced ~4” apart (the gap from the tweeter in between) should suffer serious issues in frequencies above 1.5-2 kHz, theoretically. To me, theoretically, the Strada 2 seems like it should sound a mess even without extra units. What worked well for previous versions shouldn’t work for Strada 2, dimensions-wise. I admittedly do not understand why Strada 2 not only seems to manage, but seems to sound (subjectively) better than the previous Gallo models. Perhaps friends and I are just all broken ears of a feather flocking together, grin.

If you can make frequency measurements in your room, it will be very interesting for the majority of the members here.

All I can plead here is negligence. If I did have the basic equipment, measurements would’ve already been taken for sheer curiosity. My hesitation is from lack of desire for DSP, and also this suspicion: if the available on-paper results of Strada 2 are accurate (they’re not great), it’s unlikely that increasing # units in an untreated room will show on-paper improvements.
But I’ll get to it sooner or later.

 

[dasdoing]

I wonder if you could stack a batch of radioshack realistic minimus 0.6 (horizontaly). should be very cheap to find in the US

 

[benanders]

I wonder if you could stack a batch of radioshack realistic minimus 0.6 (horizontaly). should be very cheap to find in the US

That should take the probable high frequency issues of this project and amplify them considerably.