It would be different because everything about it is different. The OP's subs are several times larger than the ones Bob simulated, so Bob's sim has much tighter center to center distances. Also Bob simulated either 12 or 24 subs (hard to tell if they are double stacked or not) and only showed it at 80 hz, above 80 hz it will get decidedly sketchier looking. It's also zoomed out quite far, that line of subs is probably 48 or so feet wide which means:
a) you can't see as much detail as a more zoomed in shot
b) it's a fully functioning array down to a fairly low frequency
And I can't recall if he mentioned the resolution on that sim either - between 1/3 to 2 octave smoothing is available in the software so it could be anywhere from lightly to heavily smoothed.
To be sure, it looks a lot better than dual subs placed at a distance but even in the nice looking horizontal array there's maybe 10 db of variation from the bright red to the dull orange areas, and probably 20 db or more variation from the dark red to the yellow areas.
My single sub placement (vertically stacked) had a lot less variation than this horizontal array. So there is destructive interference happening here.
Run this sim again for OP's much larger subs which can't be as closely coupled and see what you get.
Bob's pics from post 589 -
Last edited:
Hi 'just a guy',
The physics for idealized 2Pi model are explained in the following AES paper
AES Convention 103 paper nr: 4581
Title: Loudspeakers, Mutual Coupling and Phantom Images in Rooms
Authors: Holland, Keith R.; Newell, Philip R.
“As can be seen from equation (3), the frequency up to which the mutual coupling occurs is determined by the distance between the two sources; as the propagation distance approaches half a wavelength the phase of the pressure from the second source is no longer in phase with the velocity.”
Instead of physics for idealized field models, a more practical 'rule' can be found in many sources.
Renkus-Heinz
Line-array white paper
“A second requirement is that the sources on the line have to be less than 1/2 wavelength apart. This is the inverse of the first requirement. Olson calculated in the 1940’s that two adjacent sources radiate a spherical polar pattern (i.e. sum coherently) when they are less than 1/4 wavelength apart. Between 1/4 and 1/2-wavelength spacing the pattern narrows, but side lobes (interference patterns, caused by destructive interference) do not appear until the spacing is greater than 1/2 wavelength."
Electro-Voice
In many data sheets of EV you can find the following phrase:
“Mutual coupling occurs when the frequency is such that the center-to-centre distance is less than about one-half wavelength”
L-Acoustics
ARCS (operation manual)
"Double row configurations are typically used to improve low frequency impact due to the enhanced low frequency coupling obtained since all 15'' loudspeakers satisfy WST Condition #2 (acoustic center separation is less than half wavelength over the entire operating bandwidth)."
Genelec
Answers Library | Genelec.com
“For mutual coupling the subwoofers must be place within ½ a wavelength of one another”
Regards,
Djim
The physics for idealized 2Pi model are explained in the following AES paper
AES Convention 103 paper nr: 4581
Title: Loudspeakers, Mutual Coupling and Phantom Images in Rooms
Authors: Holland, Keith R.; Newell, Philip R.
An externally hosted image should be here but it was not working when we last tested it.
“As can be seen from equation (3), the frequency up to which the mutual coupling occurs is determined by the distance between the two sources; as the propagation distance approaches half a wavelength the phase of the pressure from the second source is no longer in phase with the velocity.”
Instead of physics for idealized field models, a more practical 'rule' can be found in many sources.
Renkus-Heinz
Line-array white paper
“A second requirement is that the sources on the line have to be less than 1/2 wavelength apart. This is the inverse of the first requirement. Olson calculated in the 1940’s that two adjacent sources radiate a spherical polar pattern (i.e. sum coherently) when they are less than 1/4 wavelength apart. Between 1/4 and 1/2-wavelength spacing the pattern narrows, but side lobes (interference patterns, caused by destructive interference) do not appear until the spacing is greater than 1/2 wavelength."
Electro-Voice
In many data sheets of EV you can find the following phrase:
“Mutual coupling occurs when the frequency is such that the center-to-centre distance is less than about one-half wavelength”
L-Acoustics
ARCS (operation manual)
"Double row configurations are typically used to improve low frequency impact due to the enhanced low frequency coupling obtained since all 15'' loudspeakers satisfy WST Condition #2 (acoustic center separation is less than half wavelength over the entire operating bandwidth)."
Genelec
Answers Library | Genelec.com
“For mutual coupling the subwoofers must be place within ½ a wavelength of one another”
Regards,
Djim
Last edited:
Thanks but that text color is really difficult to read on the blue background.
It seems you found exactly what I said in the first place -
The wording is a bit different than the words I used but that's exactly what I said, and there's an endless amount of sources that say the same thing.
All four walls of my bass horns will have a proper exponential contour (no flat planes). The rectangular aspect ratio will be fixed along the horn's center-axis, which will in turn force all four walls of the horn to follow a proper exponential expansion profile. The problem is I just suck at drawing. . . . I apologize for my misleading sketch. I'm not going to cast any flat walls in my horns.
Earlier in this thread I contemplated making the bottom horn wall parallel with the half-space plane (i.e. a flat planar surface), and I talked myself out of doing this.
Also - I am not dead set on keeping the mains outside of the horn mouths. It all depends on how big of mains I need to stuff into the horns. I do not want to choke the horns. . . .
Last edited:
Incase you missed it ben, my first iteration for my bass horns will be to install (qty 1) 12" B&C driver into each throat (with phase plug), and see if the horns will really play clean from 15 Hz to 500 Hz (the old timers say that they will. . ..) I also want to experience the phenomenal impedance matching ability of a full-size exponential horn - i.e. in taking a single 12" driver, and converting it into an equivalent wall of drivers.
Point being - I am NOT dead set on a low crossover point. The folded horns have inherent issues that limit high frequency response. I am building (ideal) straight horns for a reason. Achieving high SPL is not my end-game either. There are a few dozen homes to the South of me, all within a mile from my house (a neighborhood of 5 acre lots). Thus reproducing the bottom octaves at 145 dB would be an occasional thing - so occasional, that there's really not a "need" for it - - - if you know what I mean. . . .
In horn theory, the proper expansion is necessary. In waveguide theory curves such as these produce diffraction, but it is this theory that guides changes to the aspect ratio, arraying, and splitting over a ground plane.
You will be able to adjust the dispersion by making one wall sacrificial. For example making the top wall straight can allow narrowing in the horizontal while maintaining the correct overall expansion. Not that I'm suggesting it without analysing and optimising what the standard horn offers your situation, I'm just hoping this will offer you some direction to make clear decisions on your shape.
Not at all. There were some radial horns in a picture posted earlier and I jumped to that conclusion.
Why? not to say the drivers should be directly on the ground (rain ingress, higher frequencies), but it would make some sense.
Last edited:
Yep. Does "center to center" lose some meaning here, though?
The DBH218 that Bob used in his model is a folded horn using 2x18" woofers. I don't think that, with a blindfold on, you could play an 80Hz tone through a line of 12 of these boxes, and point to each of the 24 woofers or "centers". I think the whole mouth would be the source.
The giant ye olde multicell vocal range horns didn't have serious lobing till ~7kHz, despite the large size of the individual cells, implying it is more complicated than just putting a ruler between the centres.
His 'mic positions' graph looks good to 200Hz. But anyway, I suggested staying <80Hz. Part of what you snipped out was:
"Run the inner two horns purely as subs <60Hz (ish) [...] in other words: the outermost horns would be ~point sources above about 60Hz, and below 60Hz, the whole horn array would be firing" [because] "the heavy work (where you would want all drivers operating) is in the bottom octave."
He said "12 DBH218 in a row", so I'm going to guess 12.
It would have to be 45' wide to physically fit 12 of those 45' wide boxes.
Also: I gave the sim a go, and a line of only 4 seemed OK.
Yes. I imagine a 45' wide array of large horns or a 32' wide array of very large horns would be similar in that respect.
Yes, you could zoom right in and get ultra detail. Why? If a 5dB null was small, relative to a listener, so that it only hit one ear, I doubt it would matter.
The scale is there in the pics you posted. -20dB would be greenish, which you see in his SimD (two short rows), not in SimC (one big row).
Either one looks pretty good to me. The horizontal array would give stereo, as the OP wants, more control (delay, the ability to run some drivers over a wider range than others) and, as modeled by Bob, has a pretty wide sweet spot (certainly bigger than a hot tub).
The row that Bop posted hadn't been tweaked at all. I imagine that the 'throw' and smoothness of the sweet spot could be optimised with horn shape, angle and delay / DSP tricks. I'm just guessing - maybe one day I'll fiddle with the sim again to see. The pros like Art would know about this.
The sim, as you have lamented, uses Danley products (only). I don't think any of these are 8' across.
Using a large number of the DBH218 horns (each 45" wide) seems like a reasonable approximation of the OP's hypothetical 8' wide horn. Which is what Bob did.
Not sure what you mean, center to center never loses meaning.
OP is now talking about trying to get 500 hz out of his subs - the higher in frequency you go the worse the comb filtering is going to get and the worse it's going to look in the sim.
In the legend 136 db is right in the bright red, dark red is higher in spl and not even on the legend. Yellow is right around 120 db. So the difference between dark red and yellow is about 20 db.
As I mentioned, they aren't far enough apart to provide a stereo image at any distance. And at low frequencies you don't hear stereo anyway because of localization issues.
No, it's not a reasonable approximation and the reason is center to center distances. A reasonable approximation would use those same subs and space them apart so the center to center distance is 8 feet.
What we are looking for in the sim is coverage pattern. This is messed up by introducing comb filtering. Comb filtering is a direct result of center to center distances being too far apart. Therefore if you don't sim the correct center to center distances you are simulating something kind of vaguely similar but it isn't a valid comparison to OP's proposed layout.
Why don't all of us have power response problems at 500Hz? Do you put your stereo speakers side by side in the middle of the room? Is your sim telling the whole story? Whatever shall we do at 10kHz?
Don't worry, I am going to address the stereo mains comb filtering soon. I'm really busy (the weather is finally cooperating and I get to work most days now) and don't have a lot of time and there's lot of information to generate but I'm going to provide a fairly in depth analysis of a few things both in text descriptions and in simulated images that tell the story really well. One of the most intriguing is how bad the stereo mains comb filter with each other. But there's really nothing that can be done about that - whereas you can do something about it with the parts of the passband that don't have to be stereo - the subs.
I'm sure we've all noticed that when listening to stereo or measuring stereo speakers the sound changes wherever you place your head or your mic. There is brutal comb filtering happening with stereo mains. This is something you don't have to put up with in the subwoofer frequencies.
I'm sure we've all noticed that when listening to stereo or measuring stereo speakers the sound changes wherever you place your head or your mic. There is brutal comb filtering happening with stereo mains. This is something you don't have to put up with in the subwoofer frequencies.
The 500 Hz comment is not a design goal, but rather a test - as the old engineering texts say that my horns "should" play well from 15 Hz to 500 Hz, using a single 12" driver (high BL, low Qts, low MMS, and high Xmax).
My gut feeling (which I openly admit is a wild-a$$ guess), is that I'll be crossing the bass horns somewhere between 120 and 160 Hz. Nonetheless, in the interest of full disclosure - should my horns sound simply amazing from 15 Hz to 350 Hz, there'd be no need to cross them at 160 Hz. . . . Right?
Back to system design.
If there is no problem of stereo image, and only OP can decide how important that is in their setting, then crossover can be anywhere. Or up to 130 if you do fuss about stereo image.
But for good reasons, you must be canny about it. Speakers can handle only so many octaves well. Bass is rather specialized, perhaps less so with horns. Some folks like to keep crossover points away from sensitive hearing ranges, etc.
That's why I like a mid-range than goes from an easy-to-play 130 to an easy-to-play 2,000 Hz, just to pick numbers out of the air. That makes your sub and your tweeter icing on the cake.
If you read some of the comments about comb filtering and other fantasies critically, you'll find them really funny.... except they are meant as serious advice to you.
With all the irrelevant comments about lobing, odd that nobody has told you about Doppler distortion?
Ben
Last edited:
Exactly the point I made many posts ago. It is a fallacy to believe that comb filtering would be an issue between two stereo speakers at any frequency. The whole thing wouldn't work if it were.
Think about it this way. Each loudspeaker in a stereo setup produces a (semi)spherical wave front, and where they cross each other a complex 3D interference pattern is created. Much like when you throw two stones some distance apart in a motionless pond. The wave fronts don't really combine, they go right through each other. If two planar out of phase sound waves cross each other at 90 degrees, there is no destructive interference (ideal gas presumed).
Last edited:
Yeah, it's amazing how fast those concrete subs can get up and run 😀
BTW, may I ask what audio uses you've found for an o-scope?
Good picture at showing it is nonsense. But your image relates to orderly pure tones (water ripple physics).
Imagine the situation with never-standing-still "music ripples", varying every way instantaneously and in stereo... and also recorded that way to double the "comb" effect.
There's one poor fellow who hears the sound change whenever he moves his head. Poor guy. Do you suppose psychiatry would help?
Ben
Last edited:
It's measurable and you can hear it. Move your mic 6 inches and you get different frequency response. What's so hard to understand about this?
As I am about to show (sometime later today) at higher frequencies the comb filters are theoretically very deep and narrow. Theoretically very deep and narrow notches don't show up much in real life, they get smoothed out to a gentle ripple, so at higher frequencies (like midrange and treble frequencies) it's not a huge issue, definitely not a deal breaker or no one would listen in stereo, much less 7.1.
Perhaps the papers Ben has read point to the fact that comb filtering in stereo is fairly benign at mid and treble frequencies - this much is true. But at bass frequencies it's a whole different story.
As I am also about to show (at some point later today), at the lower frequencies, these comb filters are deep and very very wide, like 2 octaves wide. That does not smooth out into a gentle ripple, it presents as a massive suck out in the bass at certain locations for certain frequencies.
I'm going to build a pair of side-by-side "stereo" 15 Hz straight exponential bass horns. My current plan (through actual experimentation) will be to adjust the driver/throat configuration (I.E. number of drivers, and compression ratio), until I get the sound I like, at a reasonable SPL.
If it takes Qty 4 (or more) 18" B&C drivers per horn, that's fine. If it takes just a single 12" driver, that's fine too. It could be somewhere between (a few 15" drivers perhaps?) We shall see what happens. I am intentionally building the horns with a steel flange-joint in the throat, so that I may swap entire prebuilt throat/rear-chamber/driver combinations - using just a socket wrench & flange gasket. I'm going to build a small gantry crane over the rear of the horns (1-ton capacity - inside the out-building), because a wooden rear-chamber/throat assembly, with (qty 4) 18" drivers, in a box constructed from 1-1/8" plywood, is going to be too heavy for me to lift. Even a light throat assembly is too heavy - considering I'll need both hands to install the flange fasteners.
I'm estimating a final crossover frequency around 120 to 160 Hz. I'm going to build mid-bass/midrange horns. My highs are going to be off-the-shelf 2" Radian drivers. I might run some bullet tweeters (but probably not, because I can't hear over 17 kHz anyway. . . ) Please understand that I am not going to cast mid-bass horns in concrete, until I know my final crossover frequency. I.E. building 80 Hz mid-bass horns is a complete waste if I'm crossing over at 160 Hz. That being said, I'll be quite disappointed if I build 160 Hz mid-bass horns, if I end up needing an 80 Hz crossover point. . . .
Regarding lobing. I've been to the symphony orchestra - which is essentially a large field of multiple individual point-sources. Using sound-field FEA-type software, you'd probably expect the kitchen sink garage disposal to sound better. . . . But it doesn't. The symphony orchestra sounds pretty darn good. Point being - I want to make reasonable efforts to control lobing, but it is not a primary design concern of mine. . . .
Last edited:
Sounds like the system design is jelling. Great. Funny all the little pieces that matter like the weight of the box.
I'd just stick the mids in the horn mouth, no need to be concrete. Simplifies the time-alignment. Gawd, that will make the sim guys go nuts.
Speaking of symphonies, sounds like a swell place to play one of my favourite composers, native of Seattle, Alan Hovhaness.
AUDIOPHILE ALERT: Eric, you'll certainly want to get his symphony #50 named "Mount St. Helens". Even simple guys who only listen to pure tones as their heads move, will understand the audiophile glory of the volcanic eruption. For sure you'll have the wonderful recording by Gerard Schwarz and the Seattle band.
https://www.youtube.com/watch?v=CLs7rT8AGe4
Ummm, maybe a good test of your horn(s) would be the piece by Hovhaness called, "And God Created Great Whales" that has whale recordings. Or any Hovhaness (usually have exuberant brass playing... Armenian).
Ben
I'd just stick the mids in the horn mouth, no need to be concrete. Simplifies the time-alignment. Gawd, that will make the sim guys go nuts.
Speaking of symphonies, sounds like a swell place to play one of my favourite composers, native of Seattle, Alan Hovhaness.
AUDIOPHILE ALERT: Eric, you'll certainly want to get his symphony #50 named "Mount St. Helens". Even simple guys who only listen to pure tones as their heads move, will understand the audiophile glory of the volcanic eruption. For sure you'll have the wonderful recording by Gerard Schwarz and the Seattle band.
https://www.youtube.com/watch?v=CLs7rT8AGe4
Ummm, maybe a good test of your horn(s) would be the piece by Hovhaness called, "And God Created Great Whales" that has whale recordings. Or any Hovhaness (usually have exuberant brass playing... Armenian).
Ben
Last edited:
I've been advocating the mains placement as close to the centerline as possible since the beginning. You need some space to provide stereo separation of course, but I've been saying stick it in the mouth from the beginning, you can still get probably over 14 feet separation. Still wondering when you are going to start paying attention to what's going on. It seems your only purpose here is to be a general purpose cheerleader for big horns and to take pokes at me - doesn't matter much as I continually show you technical evidence proving how wrong you are.
Ok, I'm going to address a number of issues now and I've got a ton of information to present so I'm going to break it up into a number of posts. Although OP has seemingly moved past the design issues that these points address, Ben and weltersys are still arguing them so it should help to see these issues played out in full technicolor. I know this is going to be very boring to read but if you don't read it you won't understand what I'm showing and there will be endless debate about things that should be very clear, so please read what I put here. I put a considerable amount of time into this, the least the critics could do is take 10 minutes to actually read it.
Last edited: