I was thinking of making a front horn design. Couple of questions from a newbie..
(please don't say don't bother etc)
Will a 180 hz spherical horn accoustically go up to 20 khz?
Is there any way of reducing the diameter of horns without losing low end(I want to go down to 180 hz with the flare but its bloody wide), ie lengthening the throat.
(please don't say don't bother etc)
Will a 180 hz spherical horn accoustically go up to 20 khz?
Is there any way of reducing the diameter of horns without losing low end(I want to go down to 180 hz with the flare but its bloody wide), ie lengthening the throat.
Last edited:
As asked, not in this physical universe. A horn + driver's beaming HF response can do it on axis though.
A typical horn is no different from any other box alignment, it's gain BW is a function of its net volume (Vb), so as you shrink it for a given axial length, F3 rises.
Tapped horns OTOH utilize both sides of the driver's radiation to shrink it, but the trade-off is that it has no usable HF BW, so strictly a 2-3 octave band-pass (BP) alignment depending on the driver's specs and flare factor.
GM
A typical horn is no different from any other box alignment, it's gain BW is a function of its net volume (Vb), so as you shrink it for a given axial length, F3 rises.
Tapped horns OTOH utilize both sides of the driver's radiation to shrink it, but the trade-off is that it has no usable HF BW, so strictly a 2-3 octave band-pass (BP) alignment depending on the driver's specs and flare factor.
GM
Thanks very much for answering that - I'm guessing an elliptical horn could be used to reduce the width of the speaker cabinet- is there any trade-off with using an elliptical shape?
You're welcome!
Well, 'sound is round', so any reshaping of the waveform distorts it, though according Dr. Geddes an oblate spheroid expansion is the most benign. Other than his speakers though, the only way to get them AFAIK is to DIY.
That said, In another lifetime me and a few others compared round, square and several different aspect ratio rectangular WGs and came to the same conclusion as the pioneers of audio, i.e. for a given mouth cutoff, a square or rectangular one with the same area as a round one (perimeter > 1 WL) is 'close enough' to same-same with the caveat that the rectangular one's aspect ratio doesn't exceed ~ 1.0:1.273 where the short side is 1/4 WL of the mouth cutoff.
This by itself is grossly misleading though and doesn't really tell us much about what it takes to make a truly low distortion horn/WG as it's the throat and mouth transition that matters by far the most IME.
Anyway, the only elliptical looking horns I've auditioned sounded audibly worse than my square and low aspect ratio rectangular DIY units, but I blame it more on their abrupt mouth transition than its shape.
GM
Well, 'sound is round', so any reshaping of the waveform distorts it, though according Dr. Geddes an oblate spheroid expansion is the most benign. Other than his speakers though, the only way to get them AFAIK is to DIY.
That said, In another lifetime me and a few others compared round, square and several different aspect ratio rectangular WGs and came to the same conclusion as the pioneers of audio, i.e. for a given mouth cutoff, a square or rectangular one with the same area as a round one (perimeter > 1 WL) is 'close enough' to same-same with the caveat that the rectangular one's aspect ratio doesn't exceed ~ 1.0:1.273 where the short side is 1/4 WL of the mouth cutoff.
This by itself is grossly misleading though and doesn't really tell us much about what it takes to make a truly low distortion horn/WG as it's the throat and mouth transition that matters by far the most IME.
Anyway, the only elliptical looking horns I've auditioned sounded audibly worse than my square and low aspect ratio rectangular DIY units, but I blame it more on their abrupt mouth transition than its shape.
GM
I think I get that- so basically if elliptical, round or square, the 'circumference area' of the mouth (for want of a better term) should be the same. If rectangular or elliptical the ratio should be within the one given in your post.
Now thinking of an upward firing square/rectangular horn (alas, a cylindrical shape would take up too much room), using a deflector of some sort. Possibly a one driver compound horn ('rear' horn floor firing, 2.5 metre length).
edit: ah.. but will the HF be flat ???
Now thinking of an upward firing square/rectangular horn (alas, a cylindrical shape would take up too much room), using a deflector of some sort. Possibly a one driver compound horn ('rear' horn floor firing, 2.5 metre length).
edit: ah.. but will the HF be flat ???
Last edited:
For a true omni, only round will work.
It will if it has the right amount of rising on axis response and the phase plug is an acoustic mirror of it. No trivial pursuit, especially if the driver isn't designed for the app.
GM
It will if it has the right amount of rising on axis response and the phase plug is an acoustic mirror of it. No trivial pursuit, especially if the driver isn't designed for the app.
GM
Ok thanks, you know your stuff and explain in a way that's understandable to a novice! (rare)
If I do an omni it will end up looking too much like the Duevels so will now do a front firer in a completely rectangular cabinet/wide baffle/ black lacquer and floor firing. If possible single driver compound horn, if 2-way the bass will be reflex
FINAL question (he says); back folded horn loading- if the line is long enough to reach 30 -35 hz or so, is there a stipulation on how wide/area the mouth needs to be?
If I do an omni it will end up looking too much like the Duevels so will now do a front firer in a completely rectangular cabinet/wide baffle/ black lacquer and floor firing. If possible single driver compound horn, if 2-way the bass will be reflex
FINAL question (he says); back folded horn loading- if the line is long enough to reach 30 -35 hz or so, is there a stipulation on how wide/area the mouth needs to be?
Last edited:
in free space the circumference of the circular mouth should be greater than the wavelength of the lowest frequency. 30Hzwl=11.4m, mouth diameter >= 3.65m, mouth area = 10.5sqm
For half space loading you can halve the area. (horn sitting of the floor.)
For quarter space loading you can quarter the area. (horn on the floor and backed up hard against the back wall)
For corner loading divide the area by 8. This is the Pi/2 seen for eighth space.
For half space loading you can halve the area. (horn sitting of the floor.)
For quarter space loading you can quarter the area. (horn on the floor and backed up hard against the back wall)
For corner loading divide the area by 8. This is the Pi/2 seen for eighth space.
Last edited:
yeah, quarter wave so I'm guessing the line has to be around 2.5-3 metres long.
I didnt realise that the cabinet has to be flat against the wall either, thanks.
I didnt realise that the cabinet has to be flat against the wall either, thanks.
Hmm, last things first: I believe you mean the effective circumference, ergo less than a WL of the lowest desired horn loaded frequency.......... 😉
GM
You're welcome!
How does a novice know when someone 'knows their stuff', he asks rhetorically? 😉
Exactly!
Yes, it depends on what in-room F3 you want, i.e at what point do you want horn loading (1/2 WL) to roll off into TL loading (1/4 WL).
GM
OK, just so you understand I'm anything but infallible. 😉
For the purposes of this conversation, horns are 1/2 WL resonators with a 1/4 WL fundamental whereas a TL is strictly a 1/4 WL resonator: Resonances of open air columns
From this we see that as the horn unloads due to not being big enough in net box volume (Vb) to support 1/2 WL resonances it decays into a 1/4 WL one down to its pipe cut-off (Fp), causing ripples in its HF BW. Anyway, I imagine this is all be explained in-depth in one or more of MJK's various technical papers.
For a given FLH mouth cut-off, the only way to make it shorter is to truncate the throat with the attendant rolling off of LF gain BW, so to get it back, attach a bigger driver with the right specs to it or in the case of this '40's era Westrex cinema horn, four of the same size/type:
GM
For the purposes of this conversation, horns are 1/2 WL resonators with a 1/4 WL fundamental whereas a TL is strictly a 1/4 WL resonator: Resonances of open air columns
From this we see that as the horn unloads due to not being big enough in net box volume (Vb) to support 1/2 WL resonances it decays into a 1/4 WL one down to its pipe cut-off (Fp), causing ripples in its HF BW. Anyway, I imagine this is all be explained in-depth in one or more of MJK's various technical papers.
For a given FLH mouth cut-off, the only way to make it shorter is to truncate the throat with the attendant rolling off of LF gain BW, so to get it back, attach a bigger driver with the right specs to it or in the case of this '40's era Westrex cinema horn, four of the same size/type:
GM
Attachments
Greets!
I interpreted this as meaning that the physical circumference of a horn in 4pi (free) space must be greater than 1 WL in circumference of the lowest frequency to support it, a commonly held theory at one time and often repeated as irrefutable fact over the ensuing decades, but the reality is that it need only be large enough to support it acoustically, so its circumference (area actually) can be somewhat less due to the rather elastic boundary conditions (for lack of a better term) near/at the terminus (mouth).
Not being formally schooled on the subject or having made the kind of exact measurements available today to determine either a theoretically correct or real world exact circumference, about the only way I know how to explain it is to visualize a sound 'bubble' as being basically analogous to the creation of a bubble by blowing against a thin film of soap suspended in a circular wand (AKA 'bubble ring' or whatever the technical name is of what is supplied with a child's bubble blowing kit).
A soap bubble at rest then can be construed as a sound frequency at its 180 deg zero crossing, so as the bubble is expanded to half its size equates to its compression (positive/+) mode and if blown from the other side, its rarefaction (negative/-) mode only a sound 'bubble' is flip-flopping at ~1130 ft/sec (~344.42 m/sec).
In the real world though, the resonant membrane is just exciting air molecules 'stuck' to it, so obviously a sound bubble somewhat larger than it will break loose same as our soap bubble does and it's this difference between the ring's smaller circumference referenced to the soap bubble's that sets the horn's circumference, not the WL's circumference.
I originally worked out a 'close enough' rule-of-thumb to satisfy me from this and by comparing Altec's and others various horn designs to whatever measurements I could find, but sometime after joining the Sound Practices mailing list (AKA 'joelist'), someone posted a D.B. Keele? AES paper reference on the subject that has the math solution, but haven't ever gotten around to finding/reading it, so don't know how mine compares to theory.
So, do you still believe we said fundamentally the same thing? If not, then please elaborate on what point you meant to convey that I misunderstood you to say.
TIA,
GM
I interpreted this as meaning that the physical circumference of a horn in 4pi (free) space must be greater than 1 WL in circumference of the lowest frequency to support it, a commonly held theory at one time and often repeated as irrefutable fact over the ensuing decades, but the reality is that it need only be large enough to support it acoustically, so its circumference (area actually) can be somewhat less due to the rather elastic boundary conditions (for lack of a better term) near/at the terminus (mouth).
Not being formally schooled on the subject or having made the kind of exact measurements available today to determine either a theoretically correct or real world exact circumference, about the only way I know how to explain it is to visualize a sound 'bubble' as being basically analogous to the creation of a bubble by blowing against a thin film of soap suspended in a circular wand (AKA 'bubble ring' or whatever the technical name is of what is supplied with a child's bubble blowing kit).
A soap bubble at rest then can be construed as a sound frequency at its 180 deg zero crossing, so as the bubble is expanded to half its size equates to its compression (positive/+) mode and if blown from the other side, its rarefaction (negative/-) mode only a sound 'bubble' is flip-flopping at ~1130 ft/sec (~344.42 m/sec).
In the real world though, the resonant membrane is just exciting air molecules 'stuck' to it, so obviously a sound bubble somewhat larger than it will break loose same as our soap bubble does and it's this difference between the ring's smaller circumference referenced to the soap bubble's that sets the horn's circumference, not the WL's circumference.
I originally worked out a 'close enough' rule-of-thumb to satisfy me from this and by comparing Altec's and others various horn designs to whatever measurements I could find, but sometime after joining the Sound Practices mailing list (AKA 'joelist'), someone posted a D.B. Keele? AES paper reference on the subject that has the math solution, but haven't ever gotten around to finding/reading it, so don't know how mine compares to theory.
So, do you still believe we said fundamentally the same thing? If not, then please elaborate on what point you meant to convey that I misunderstood you to say.
TIA,
GM
I suspect this is the Keele paper you were refering to Greg? http://www.xlrtechs.com/dbkeele.com/PDF/Keele (1973-09 AES Preprint) - Optimum Horn Mouth Size.pdf Not sure how he's managed it, but it looks like he's got all, or the majority, of his AES papers up on his website now. 🙂
Last edited:
I guess, I can't seem to find the old joelist message. Anyway, good to know about his site, thanks! I'll DL/read the paper a bit later. The rainy weather's let up for the moment, so got to go back to getting the shop functioning again.
GM
GM
- Status
- Not open for further replies.