Hey, just another crazy idea, what if the horn is bifurcated so that you have two centerlines, which end up at the left and right edges? then you would have two "centers", one close to each mains. Technically that would end up as one big, integrated stereo horn, so entropy gets his stereo after all..... ?
Anybody here go to the INFOCOMM show here in Vegas this year? And if so, did you stop by the Danley room?
They demo'd the new J3-96's. A single 90 degree four foot tall three feet wide horn with six 15's, eight 6.5" mids and four 1.5" highs that absolutely sounded Hi-Fi.
A pair of these would take this project past the holy **** factor. At 28k each they are less than a lot of audiophile speakers.
I really hope this project comes to fruition. Everybody wants a full size 20Hz horn!
Barry.
They demo'd the new J3-96's. A single 90 degree four foot tall three feet wide horn with six 15's, eight 6.5" mids and four 1.5" highs that absolutely sounded Hi-Fi.
A pair of these would take this project past the holy **** factor. At 28k each they are less than a lot of audiophile speakers.
I really hope this project comes to fruition. Everybody wants a full size 20Hz horn!
Barry.
My SynTripP horn using a 3" diaphragm on a 1.4" exit does 90 degrees -7dB at 20kHz, though driver response drops off above 15kHz.
As you can see, the dispersion is wider than 90 degrees at 1900 Hz, narrower at 3500 Hz and 12kHz, but wider than 90 degrees at 15kHz, -4.5dB down at 45 degrees off axis.
It will be interesting to compare polars again when I switch to the B&CDE90TN, some of the upper deviations may be phase plug, rather than horn related.
Horns with narrow "pinched throat" diffraction slots can exhibit near perfect polar response, but generally don't sound as good as conical horns with a secondary horn flare, there are always various compromises to accept with whatever design is chosen.
Art
Attachments
With a single center sub, the stereo mains phase response in the crossover region can only be matched to the sub at a single specific distance on axis.
My preference for stereo outdoor systems with a wide listening area is to have the mains directly over each sub, as a smooth phase transition in the 100 Hz crossover region is needed for good dynamic punch. The lack of a defined "punch" with a single sub in all but one location is a worse option to me than the low frequency lobing issues the separated LF sources cause, which are difficult to notice with the type of music I typically mix or play back.
That said, there are two completely different "right" answers to the single center vs L/R sub debate, if one is listening in the center sweet spot, as I am in my studio right now, (or Entropy Eric's hot tub) the crossover transition is just fine in the seat I have aligned to.
Art
Last edited:
That's really nice on axis but a bit of a roller coaster off axis. Here's the 12 inch SEOS measurements that correspond to the polar I showed previously. (At least I think these are the measurements they used to generate the polar, both this measurement and the polar appear in the same post here - At last, the SEOS12 Measurements - AVS Forum | Home Theater Discussions And Reviews )
The SEOS looks smoother off axis but both have pros and cons.
To compare full system designs (not just the SEOS tweeter horn), I'm not sure which design this is but I'm pretty sure it's one of the ones using the SEOS 12 or 15.
Last edited:
This is the Fusion 15 Sentinal using the SEOS 15 - Fusion-15 Sentinel DIY Sound Group
There's a bit of bunching around 300 - 400 hz but that's well below the crossover point so has nothing to do with the tweeter horn, other than that it looks pretty smooth off axis. The measurement in the last post may also be the same speaker but with a gated measurement, I'm not sure, but it's a SEOS speaker for sure.
An externally hosted image should be here but it was not working when we last tested it.
There's a bit of bunching around 300 - 400 hz but that's well below the crossover point so has nothing to do with the tweeter horn, other than that it looks pretty smooth off axis. The measurement in the last post may also be the same speaker but with a gated measurement, I'm not sure, but it's a SEOS speaker for sure.
Last edited:
Ok, so this bugged me when I first read it, but I was insecure enough to want to confirm from a textbook, which I did.
1/4 wavelength is wrong. 1/2 wavelength is correct. As a function of wavelength, two sources spaced half a wavelength will sum without lobing at that frequency and below.
my source: page 83 in Sound Systems: Design and optimization by Bob McCarthy.
But the rest of your statement is of course valid. The wavelength criterion should be upheld a little bit beyond the crossover frequency until the level offset between sources takes over.
By the way, at one wavelength offset between sources the lobing pattern creates nulls at 45° off axis. Since we are talking about vertical displacement this would basically mean that the nulls would be pointing upwards and downwards. In the forward direction there would be only constructive summation.... or am I missing something?
The interesting question for me is how would this horn behave in the real world. Would it behave A) like JAG has suggested, with the center point of the mouth as an absolute reference point, or B) rather like a row of subs?
I'm beginning to dig this idea........
Would it make the difference and answer my question?
Yes, which calls in the question: what if the sources are only radiating +/- 45 degrees anyway?
Of all the concerns with this, the total radiated power through the crossover is one of the main ones.
The on axis response of the horn upwards of 100 Hz will have a rising response on axis due to the deep, narrow throat angle, just like trumpets or trombones do, but octaves lower. That single source point will become larger as the frequency is lower, when the wave length is longer than the mouth boundary, it wraps around the horn,The interesting question for me is how would this horn behave in the real world. Would it behave A) like JAG has suggested, with the center point of the mouth as an absolute reference point, or B) rather like a row of subs?
A row of individual subs creates an interference pattern dependent on frequency.
Hi bob4,
In forward direction you have "constructive summation" of signals from left and right only if both sides have their own sub. Than all signals arrive at the same time and result in a summed signal of + 6dB
In situation with a single sub the transients from all three sources (left, sub, right), around the crossover frequency, only arrive at the same time in a relative small sweet spot. Outside the sweet spot (even if it is on-axis) they arrive at different times and they can't sum any longer. Meanwhile the rest of the signals still sum. This is experienced as the "The lack of a defined "punch"" and I agree with Art that it is a worse option compared to the 'lobing issues' of separate subs.
Regards,
Djim
Maybe you did miss something
In forward direction you have "constructive summation" of signals from left and right only if both sides have their own sub. Than all signals arrive at the same time and result in a summed signal of + 6dB
In situation with a single sub the transients from all three sources (left, sub, right), around the crossover frequency, only arrive at the same time in a relative small sweet spot. Outside the sweet spot (even if it is on-axis) they arrive at different times and they can't sum any longer. Meanwhile the rest of the signals still sum. This is experienced as the "The lack of a defined "punch"" and I agree with Art that it is a worse option compared to the 'lobing issues' of separate subs.
Regards,
Djim
I can't hear over 17 kHz. I have some hearing loss between 2 kHz and 4 kHz in both ears - but not nearly enough to warrant hearing aids. My hearing is still pretty good - considering my years of submarine service.
The "room" is outside (half-space). I hadn't thought about chairs. The treble should sound good. I'm not overly concerned with sound above 17 kHz - and not because I can't hear it, but rather because there's not much up there anyway. I've got a pair of Radian horns on the wife's stereo, powered by a DC-300A2 (1300 Hz high-pass active crossover) - and I really like the way they sound. Signal source will primarily be CD recordings. I don't own records, or tapes. I will be investing in a digital delay processor, to time-align the drivers (still shopping for brand - Yamaha has a nice professional sound processer). As I said before, this is NOT a "reference" system. I fully intend to turn up the bottom end, and also the "kick" frequencies, so it sounds like a live show. I already know that you do not approve. . . . .
But down the road, I'll probably build a much smaller indoor "reference" type system - so that a recording of a cello, actually sounds like someone is in the room playing a cello.
Last edited:
Bob4, AllenB, Djim - originally I was planning to build a pair of 20-Hz bass horns (a true stereo system).
Some folks were able to talk me out of it - citing the fact that human ears cannot hear stereo below 80 Hz, and also because of potential lobing issues between horns.
What is your opinion. Should I build stereo bass horns, or one mono horn??? My intuition (which is admittedly quite novice in this realm) is that I should be building stereo bass horns.
Please advise?
Some folks were able to talk me out of it - citing the fact that human ears cannot hear stereo below 80 Hz, and also because of potential lobing issues between horns.
What is your opinion. Should I build stereo bass horns, or one mono horn??? My intuition (which is admittedly quite novice in this realm) is that I should be building stereo bass horns.
Please advise?
Ok, several notes here.
First, I didn't say lobing occurs at 1/4 wavelength distance between sound sources. I said ideally the sound sources should be 1/4 wavelength apart or less. Big difference. And this is why.
2. Two sound sources that are within 1/4 wavelength of each other will provide full mutual constructive summation (+6 db) up to the frequency of the 1/4 wavelength. This is ideal.
Here's my source - https://books.google.ca/books?id=xo...v=onepage&q=wavelength 6 db summation&f=false
"only within 1/4 wavelength can true summing be expected"
Not sure if that link is going to work right or not, but this is fact.
Lobing occurs at a higher frequency. Two separate issues.
3. The cause of lobing is the triangulated summation when the sources are more than 1/4 wave apart doesn't add to +6db because the relative phase at the listening position is different, so instead of summation you can get very large notches in response depending on where you are in relation to the sound sources.
4. Getting textbook definitions isn't going to show you much about what the issue looks like, it should be simulated or measured. So here's some sims.
This is a random .frd file input into Bagby's Passive Crossover designer twice (two sound sources). The virtual mic location is 2 meters back perpendicular to the centerpoint of the two sound sources (in between). I've set the horizontal offset to 0.5 meters and -0.5 meters, so the virtual mic is centered between them. I chose to space them 1 meter apart to make the numbers easy, 1/4 wavelength of 86 hz is 1 meter and 1/2 wavelength of 172 hz is also 1 meter. And the wavelength of 344 hz is 1 meter.
This is directly on axis, 0 degrees. The black line is the summed output of the two sound sources. On axis there are no problems.
An externally hosted image should be here but it was not working when we last tested it.
This is 14.04 degrees off axis. I chose to show this as an intermediate step between 0 and 45 degrees.
An externally hosted image should be here but it was not working when we last tested it.
Now this is 45 degrees off axis.
An externally hosted image should be here but it was not working when we last tested it.
According to your book learnin' the one meter separation between the sound sources should have given us a lobe at 344 hz but we see the strong comb filter null much lower, at around 250 hz. Furthermore, you theory says we should have perfect summation (+6 db) at 172 hz (1/2 wavelength of 172 hz is 1 meter). Clearly this is not the case.
And if we move out to 90 degrees the comb filter notch moves down to around 172 hz.
So I'm not sure if your book is wrong or if you interpreted it wrong. This spreadsheet I'm using is not perfect but I have no reason to doubt that it can pretty accurately predict comb filter notches.
5. Moving on, if we look at the 45 degree sim, we can see that we start to lose the 6 db perfect summation at the wavelength the 1/4 wave theory predicts, around 86 hz. It's not a steep rolloff until around 150 hz, but anything less than +6 db summation is lost spl, that's why I repeatedly said "IDEALLY" the sound sources should be within 1/4 wavelength.
6. You seem to think that vertical lobing is going to be an issue but not horizontal. There's no real reason to lift the mains higher than the centerpoint of the sub horn height (which I'm guessing will be 6 - 8 ft), OP's audience is small enough and close enough that it's not a necessity and will cause more issues than it attempts to solve. On the other hand, you've got a proposed mains spacing of 30 feet, 30 feet from each other and 15 feet each from the horn center. As I just showed, with a 1 meter spacing you get comb filter notches down at 250 hz at 45 hz, with a 15 foot spacing these notches are going to be MUCH lower in frequency, there are going to be a lot more of them and they are going to be spaced more closely together.
The interesting question for me is how would this horn behave in the real world. Would it behave A) like JAG has suggested, with the center point of the mouth as an absolute reference point, or B) rather like a row of subs?
Without going into the differences between a single large horn, a multicell horn or a row of modular individual horns I can tell you this - it doesn't matter, there's always going to be lobing.
Consider this - instead of one big horn mouth you've got 8 individual horn mouths side by side that are the same combined size and shape as the one big horn mouth. Now place your mains at each end of your 8 horn array. The main speaker on the left is going to be very close to the sub horn right beside it, but it's going to be very far from the horn sub at the other end of the sub array. You need to consider the center to center distance of EVERY sound source that 's playing the same frequency, not just the one closest to it.
Line arrays are a similar example. They have benefits and do provide a somewhat coherent wavefront but the summation of each driver relative to EVERY other driver in the array is complex. There's a lot of positive summation (because there's a lot of drivers in a close space) but there's also a bit of destructive cancellation because if the array is long enough, some of the drivers are not going to be within 1/4 wavelength of each other.
Hi bob4,
Maybe you did miss something
In forward direction you have "constructive summation" of signals from left and right only if both sides have their own sub. Than all signals arrive at the same time and result in a summed signal of + 6dB
But at the low frequencies the left and right separated subs interfere with each other destructively, resulting in lobing.
These are not separate issues, it's all the same thing. The triangulated summation of two separate sound sources and a listening position. At some frequencies there's going to be summing, at others nulls, it depends on the the frequency and the distance and angle.
You can't avoid lobing by having two stereo subs, the subs will interfere with each other and cause lobing. And the "lack of punch" caused by separated sound sources is caused by the exact same issue that causes lobing - distances, angles, arrival times.
Just a guy, can you please run the same simulation for a full-range system - spaced 60 feet apart, standing roughly 80 feet back? I am interested in seeing how much lobing will actually happen from 120 Hz and down - and how far away from the ideal listening position one must stand for it to happen. . . .
"Outside" is an acoustic environment. Walls either existing or in the works? Berms? Noise sources? Neighbours?
For sure, a lot of the heartfelt agonizing about lobes and comb filters don't have much relevance to your goals (or anywhere else pretty much). But producing a coherent (rather than an incoherent) stereo image matters. Can't say as a reproduction without a room makes sense unless you are playing the instrument, not reproducing it. Like those "minus one" recordings (with no singer), where will you find recording suitable for playing out of doors? Anybody?
Open air eats sound something fierce, esp. the good parts. So I'd devote a lot of effort to getting speakers (horn excepted) up close. Which leads to headaches with your big horn. In public assembly occupancies*, almost always what's called a "distributed system" which means a lot of little speakers close to everybody's head. A circle of speakers around your hot tub.
No disrespect, but like half the elderly men on this forum, you'd cry if you saw your hearing plotted by a legitimate professional. No other way to know.
Ben
*but within halls and churches with a stage you've got a whole other kettle of fish
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.