I know, every time I say anything at all I'm trolling, right? But I just got done explaining to you why you are seeing the results you see in your measurements and I already explained to you two years ago why Hornresp can't simulate barn doors.
The really simple explanation is that it's not the right tool for the job. You can't drive to work in your toaster, don't expect to be able to simulate things that the simulator is not set up to calculate.
A simulator calculates the acoustical impedance INSIDE the combination of chambers, ducts and waveguides that you input. It also calculates the acoustical impedance of the driver face that is firing into the outside air (when applicable) and sums the two together and shows the results in reference to the environment (1 pi, 2 pi, etc).
Hornresp DOES NOT concern itself with the stuff outside the horn like the front face of the cab which produces significant diffraction - diffraction includes everything across the whole passband, the ramp up at increasing frequencies, the hump near the baffle step frequency, the rippled response above baffle step and the steady 6 db increase at frequencies well above baffle step. The diffraction curve is also significantly impacted by where you measure it - it will look significantly different if measured on axis at one meter compared to well off axis at 100 meters. Hornresp can't calculate ANY of this stuff if it doesn't know how big the baffle face is or what shape it is - everything is round in Hornresp.
That's why you can't simulate flat barn doors in Hornresp. And the example you just showed, which is not flat, probably can't be accurately simulated either because you've got a huge radiation impedance mismatch between the TH mouth and the extender flare, furthermore the extender flare is only a few inches deep and nowhere near round. This is probably a bit too complex for Hornresp, even if you did have enough segments to attempt to simulate it.
Here's your comments from two years ago.
This was my reply.
After a bit of back and forth banter (in which I clued you in to diffraction) you said this:
Any of this sounding familiar? It's from the trainwreck Danley BC Subs Reverse Engineered thread, where we tackled such ridiculousness as the secret of the BC series being a turbulent vortex at the mouth, the claims that the BC series were stepped tapped horns with a really weird frequency response (they are neither tapped horns nor stepped and don't have a weird frequency response) and several other nutty weird ideas.
But I don't doubt if you don't remember, you certainly didn't learn anything.
I went on to show a sim compared to your barn door measurements that showed that diffraction was definitely at play. It was a quick 5 minute sim and didn't accurately reflect the dimensions or the driver array but it clearly showed a correlation to your measurements.
I had a bit of time so I just went through and recreated the sims with more accurate dimensions. Hopefully even you can't be so stubborn to deny the remarkable correlations to your own measurements this time.
First of all here's you measured stack of speakers and here's the measurement and the key showing what the measurements mean.
An externally hosted image should be here but it was not working when we last tested it.
Look familiar? Now let's sim the diffraction profile of 8PV, 8PVW and 8PVW4 and compare the diffraction profiles to your measurements.
First note that I can't sim the array with this software, I can only sim a single composite driver in the middle, that will throw things off a bit. I might redo this with The Edge if I have time later. Even so, let's see if we can spot any obvious trends.
Here's the diffraction sim of 8PV, just the stack of 8 speakers.
An externally hosted image should be here but it was not working when we last tested it.
Here's 8PVW (the array with two wings, each sized 20 x 109.5 inches). Note that I can't simulate the fact that the wings are a bit higher than the array either, but whatever, let's see if we can spot any trends anyway.
An externally hosted image should be here but it was not working when we last tested it.
Now here's 8PVW4 (the array with 4 wings).
An externally hosted image should be here but it was not working when we last tested it.
First look at your 8PV and 8PVW measurements. You've clearly got about 1 (maybe 1.25) db difference at around 70 - 80 hz. The difference increases to 2 db at 100 hz and then trails off down to 1 db difference at 200 hz. That almost perfectly describes the curve shape of the 8PVW diffraction curve sim across the same described bandwidth.
Now look at your 8PVW and 8PVW4 measurements. You've got about 1 db difference at 70 - 80 hz, about 1/2 db difference at 100 hz and no difference at 200 hz. That pretty accurately describes the 8PVW4 sim, what you did was move the baffle step hump down a few hz so your 1 db gain over 8PVW is at around 75 hz.
Now look at the 8PV sim, the first sim with no wings. See the baffle step hump is way up at 300 hz and you've got a bit of gain from diffraction from 300 - about 800 hz? Now compare that to your own 8PV measurement. Your 8PV measurement WITHOUT WINGS is louder at 300 - 800 hz than either of the measurements WITH wings. Why? It's the diffraction hump that boosts those frequencies.
In all your measurements you get decreasing amounts of gain below about 70 hz, just like the sims suggest.
These sims are VERY far from perfect, your measured array of drivers are simulated as a single composite driver in the middle. Your winged baffle face is notched out in the top at the middle, I can't recreate that.
Also the diffraction profile changes depending on where you measure it, when the virtual mic is moved back to 30 feet all the sims ride the +6db line above the baffle step frequency and the curve shape changes a bit, the peaks and dips get a bit more pronounced and rise a bit in frequency.
Despite the imperfections in the sim you would have to be blind to not see the trends when comparing the measurements to the sims. The curve shape of the simulated diffraction profile explains the measured gains when adding wings reasonably well.
Ideally this should be done in Akabak, a MUCH more detailed and accurate sim could be done but my Akabak computer is broken and I don't have time to confirm things that should be obvious, things I told you two years ago already.
Last edited:
Someone correct me if I'm wrong, but as I understand it, with conventional "bass horns" for PA, there is usually a compromise in bandwidth. If the horn is supposed to play low it has to be long, and thus folded, which will restrict upper cutoff frequency (e.g. lab horn). On the other hand a horn for upper bass/low mid can be short enough to be built straight, but will not go low.
Wouldn't it be possible to "have the cake and eat it" (sub + kick from the same horn) with a long, straight horn such as OP intends to build?
I'd have thought that if he wanted kick frequencies from the sub horn he would have to go back to stereo subwoofers.. (above ~100Hz you can start to localise the sound) There is no such thing as the 'no compromise' system he wants afaik - all choices have good and bad elements. You just need to pick your own flavour of poison...
Last edited:
The higher the crossover frequency the closer the mains should be to the sub to mitigate lobing issues around the crossover frequency, as all three speakers (the sub and the two mains) are playing the same frequencies over a small bandwidth and will destructively interfere with each other. The lower the crossover point the farther apart the speakers can be. Ideally all speakers playing the same frequencies should be within 1/4 wavelength of each other at the highest common frequencies they are both playing. OP seems to want a VERY wide spacing of the mains, probably well away from the sub; since the ideal spacing would be 1/4 wavelength apart something has to give, either lobing or mains separation.
At the same time, OP hasn't considered the fact that if the mains are horn loaded and have 90 degree horizontal coverage, he could make them triangular and place them right directly in front of the horn mouth back to back with each other and cover the entire 180 degrees in front of the horn and the two back to back mains wouldn't interfere with each other (due to the 90 degree dispersion) or the sub. There are more ways than one to cover an audience, the speakers don't have to be 30 feet apart, they can literally be back to back and touching each other and cover 180 degrees without overlapping each other's coverage zone. It's all about dispersion angles, coverage area and speaker placement.
At the same time, OP hasn't considered the fact that if the mains are horn loaded and have 90 degree horizontal coverage, he could make them triangular and place them right directly in front of the horn mouth back to back with each other and cover the entire 180 degrees in front of the horn and the two back to back mains wouldn't interfere with each other (due to the 90 degree dispersion) or the sub. There are more ways than one to cover an audience, the speakers don't have to be 30 feet apart, they can literally be back to back and touching each other and cover 180 degrees without overlapping each other's coverage zone. It's all about dispersion angles, coverage area and speaker placement.
Last edited:
Here's a super high quality image to show the concept. This is a top down view, the bass horn mouth is obvious enough. The two triangles back to back are two triangle shaped horn mains with approximately 90 degree horizontal coverage (each). The green part of the hemisphere is the the lower frequency 90 degree radiation pattern from the left speaker, the red is the radiation pattern from the right speaker. They don't overlap so they don't destructively interfere with each other, and clearly they can be placed very close together and right on the sub horn exit axis if desired to avoid interference issues with the sub. Now you can cross over as high or as low as you like and have no lobing issues.
The only problem with this is that obviously the horn loaded mains will become directional with higher frequency so if they beam too much the center position of the audience doesn't get good high frequency coverage. But there's ways to deal with that, this is just a concept showing close mains placement as a way to avoid lobing and allow for a high crossover frequency. There are many different ways to apply this concept (with differing radiation angles and different placements. It doesn't have to be two mains spaced 30 feet apart and toed in to point directly at the middle of the audience as the OP is suggesting.
The only problem with this is that obviously the horn loaded mains will become directional with higher frequency so if they beam too much the center position of the audience doesn't get good high frequency coverage. But there's ways to deal with that, this is just a concept showing close mains placement as a way to avoid lobing and allow for a high crossover frequency. There are many different ways to apply this concept (with differing radiation angles and different placements. It doesn't have to be two mains spaced 30 feet apart and toed in to point directly at the middle of the audience as the OP is suggesting.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
More Doodles for Post #438.
Hi again,
Just to follow up on Post #438, here is what happens if you divide the horn by 4, and then use the Hornresp petal output (4-sided) to give you individual horns, which then can be combined (hm, that looks familiar). One petal horn could be build from 3/4" form wood for testing? If four are not enough eight? Concrete horns w/ earth fill for damping, I must be dreaming.
Conveniently we end up w/ 8'-or so-square individual mouth (close enough).
Regards,
Hi again,
Just to follow up on Post #438, here is what happens if you divide the horn by 4, and then use the Hornresp petal output (4-sided) to give you individual horns, which then can be combined (hm, that looks familiar). One petal horn could be build from 3/4" form wood for testing? If four are not enough eight? Concrete horns w/ earth fill for damping, I must be dreaming.
Conveniently we end up w/ 8'-or so-square individual mouth (close enough).
Regards,
yep you are right, a lot of choices involved in this system design.
I doubt playing the horn up to somewhere between 120 and 150 Hz will yield any problems with localization.
https://en.wikipedia.org/wiki/Sound_localization#Evaluation_for_low_frequencies
Anyway I suppose a ballpark figure for the crossover point will be anywhere between 80 - 160 Hz?
One thing many horn enthusiasts report to love about horns is the way they reproduce kick frequencies (and events
). It would seem a bit like a waste to have such a marvelous horn and cut away all the kick. I think it wouldn't hurt to design the horn with sufficient bandwidth to play upper bass as well. If it is not needed, simply lower the crossover point.
Entropy, do you have another thread to discuss the rest/the whole system, or do you want to keep everything in this thread? It would be interesting to know what kind of top end you've been thinking about to go with your behemoth?
Note that the midrange horns would have to be VERY large for this to work with zero overlap, even with a 200 hz crossover the mid horns would have to be very large to limit their dispersion to 90 degrees, but OP wants big horns for the entire frequency band so ...
Some separation of the mains is ok, some overlap is ok, this is just a concept to show the exact opposite of what OP is planning can work too.
So listening at 'x' he wouldn't really be getting the left channel ?
I think it would need to be mono to work ?
Attachments
Don't want to sound negative, but I doubt OP wants to listen to the Peppers in mono??
(although I also pointed him to the narrow sweet spot problem with the bob mc carthy link I posted.....)
Hi bob4,
Post #432: "...correct me if I'm wrong, but as I understand it, with conventional "bass horns" for PA, there is usually a compromise in bandwidth..."
You are correct. Right now we are just trying to come up w/ system parameters. I previously suggested a 4-way horn system w/ app. 3 octaves coverage per horn. Depending on how high you can actually use the sub horn you can then define the next stage. This will require a full size experiment.
I have always liked multi-cellular horns, just never thought about them for bass application. But, why not? This would give an additional stage of control over the distribution.
Would a L-C-R system be more appropriate than a very wide stereo L-R?
Regards,
Post #432: "...correct me if I'm wrong, but as I understand it, with conventional "bass horns" for PA, there is usually a compromise in bandwidth..."
You are correct. Right now we are just trying to come up w/ system parameters. I previously suggested a 4-way horn system w/ app. 3 octaves coverage per horn. Depending on how high you can actually use the sub horn you can then define the next stage. This will require a full size experiment.
I have always liked multi-cellular horns, just never thought about them for bass application. But, why not? This would give an additional stage of control over the distribution.
Would a L-C-R system be more appropriate than a very wide stereo L-R?
Regards,
Yes Indeed!
OP could also go bonkers with a "center" channel and experiment with MidSide/orthoperspective techniques to get awesome 3D effects, which admittedly screw up the stereo soundstage localization, but can sound very immersive. I've done some small scale experiments.
Yeah, I probably should have explained this a lot better. There has to be some overlap, it has to overlap everywhere that a listener would be standing but it doesn't have to overlap before the audience area. If the main listening position (the hot tub) is 30 feet out you don't need overlap until you get 30 feet out unless there are going to be people closer than that on a regular basis.
This is just a concept drawing to show that coverage patterns can be controlled with directivity, and that the mains can be very close together instead of the proposed 30 feet apart. With a close mains placement you can avoid lobing issues and cross over very high if desired. But yes, the coverage does have to overlap the audience position.
This is the kind of problem a close centralized mains location would fix and the reason I brought this up and totally messed up describing the concept.
Hi Rob Wells,
Post #457: "...the crossover point will be anywhere between 80 - 160 Hz?"
That would be my guess. The 18SW115 will start rolling off around 160. I also assume that a single horn of the general proposed dimensions will already start beaming in the 160 Hz range, but a multi-cellular solution would provide greater dispersion. These are guesses on my part, I've never been involved in setting up something of this size.
Regards,
Post #457: "...the crossover point will be anywhere between 80 - 160 Hz?"
That would be my guess. The 18SW115 will start rolling off around 160. I also assume that a single horn of the general proposed dimensions will already start beaming in the 160 Hz range, but a multi-cellular solution would provide greater dispersion. These are guesses on my part, I've never been involved in setting up something of this size.
Regards,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.