..the coiled vs. straight thing, I was thinking that there is something going on WRT either the way the wavefront changes going round and round and/or related to the way that sound when reflected off a surface (don't laugh and point now) as with a speaker like the Blows 901 is somehow altered (often for the better), as compared to a straight shot down the "barrel".
Apparently I am the only one who thinks this is of interest?
_-_-bear[/QUOTE]
No, it is interesting.
If you’re going to bend sound, you have to be aware of where the boundaries are.
At 22Khz, sound will pass through 36 inches of .025” copper capillary tube with a mic at the end just fine. In fact, you can wind the copper tube around a coffee mug and nothing happens at the mic end.
Sound can flow like a pressure driven fluid, can reflect like light off a mirror and most often somewhere in between, it all depends on the dimensions compared to the wavelength involved.
Like sound bending around a horn, consider the inside and outside path, they are different lengths, what happens as the frequency climbs and the differential reaches 180 degrees?.
If you keep the differential small, you can bend sound around corners like a fluid;
.:: VTC Pro Audio - Paraline Element ::.
If one were to apply a rule of thumb for this to an expanding cross section, you find the radius of the bend has to increase with the size of the horn. In the compression driver, sound bends around sharp corners but also where the acoustic size is small.
Ultimately, in the strictest sense, once the acoustic size of the progressing wavefront is acoustically large enough compared to the wavelength, any further change in horn wall angle causes the directivity to vary with frequency and can cause a radiation from that point.
At worst case, you can hear that extra source discreetly, at best, it only draws attention to the speaker. Earl Geddes horn flares are the mathematically least perturbing curved shape (which has superior driver loading).
Consider that for a one inch exit compression driver, that the exit is already large enough to limit the maximum horn angle at 20KHz to 60-90 degrees .
A 2 inch driver is about half that angle. Also, at high frequencies, the wavefront shape emerging from the driver may or may not be a plane wave depending on how the driver is made. For a conical horn like I use at work, a driver with an expanding wavefront like a bms 4550 is a good choice for a tweeter where it is greatly rolled off for CD compensation.
There are three working zones for a horn too, if your well above the low corner, you have the acoustic transformation region which extends form the driver radiator forward to a point in the horn where the area is about 1Wl in circumference (which is also the ideal thumb rule size for a mouth at the low cutoff).
At a higher F, this working region only extends a little way from the driver though. Past that point is where the horn is still bounding the wave edges and can ‘steer” where the sound goes. At some point beyond you reach a dimension where Don Keele’s pattern loss thumb rule applies, the wave front is acoustically large enough to proceed at that angle without horn walls. This may still be well inside the horn up high too, it is why an exponential or tractrix horn has a narrowing dispersion angle as the frequency climbs. At 20KHz, that acoustic transformation zone is well inside the compression driver itself and the exit is already large enough to be partly controlling the directivity at 20KHz.
It’s all important and interesting stuff to play with.
Best,
Tom Danley
Apparently I am the only one who thinks this is of interest?
_-_-bear[/QUOTE]
No, it is interesting.
If you’re going to bend sound, you have to be aware of where the boundaries are.
At 22Khz, sound will pass through 36 inches of .025” copper capillary tube with a mic at the end just fine. In fact, you can wind the copper tube around a coffee mug and nothing happens at the mic end.
Sound can flow like a pressure driven fluid, can reflect like light off a mirror and most often somewhere in between, it all depends on the dimensions compared to the wavelength involved.
Like sound bending around a horn, consider the inside and outside path, they are different lengths, what happens as the frequency climbs and the differential reaches 180 degrees?.
If you keep the differential small, you can bend sound around corners like a fluid;
.:: VTC Pro Audio - Paraline Element ::.
If one were to apply a rule of thumb for this to an expanding cross section, you find the radius of the bend has to increase with the size of the horn. In the compression driver, sound bends around sharp corners but also where the acoustic size is small.
Ultimately, in the strictest sense, once the acoustic size of the progressing wavefront is acoustically large enough compared to the wavelength, any further change in horn wall angle causes the directivity to vary with frequency and can cause a radiation from that point.
At worst case, you can hear that extra source discreetly, at best, it only draws attention to the speaker. Earl Geddes horn flares are the mathematically least perturbing curved shape (which has superior driver loading).
Consider that for a one inch exit compression driver, that the exit is already large enough to limit the maximum horn angle at 20KHz to 60-90 degrees .
A 2 inch driver is about half that angle. Also, at high frequencies, the wavefront shape emerging from the driver may or may not be a plane wave depending on how the driver is made. For a conical horn like I use at work, a driver with an expanding wavefront like a bms 4550 is a good choice for a tweeter where it is greatly rolled off for CD compensation.
There are three working zones for a horn too, if your well above the low corner, you have the acoustic transformation region which extends form the driver radiator forward to a point in the horn where the area is about 1Wl in circumference (which is also the ideal thumb rule size for a mouth at the low cutoff).
At a higher F, this working region only extends a little way from the driver though. Past that point is where the horn is still bounding the wave edges and can ‘steer” where the sound goes. At some point beyond you reach a dimension where Don Keele’s pattern loss thumb rule applies, the wave front is acoustically large enough to proceed at that angle without horn walls. This may still be well inside the horn up high too, it is why an exponential or tractrix horn has a narrowing dispersion angle as the frequency climbs. At 20KHz, that acoustic transformation zone is well inside the compression driver itself and the exit is already large enough to be partly controlling the directivity at 20KHz.
It’s all important and interesting stuff to play with.
Best,
Tom Danley
Yes, they are the simulations of a gentleman of far greater greater nobility and learning than I.
BTW, they are half-space sims, as if the mouth of the the horn is set in an infinite baffle. So in the unbaffled real world, the mouth diffraction would likely make a bigger appearance.
Bill, I think Pano edited your post 640 instead of quoting it. If I am right, this is your original post?
By definition, a true point source is omnidirectional throughout its bandwidth, and therefor its size must range between infinitely small and perhaps a quarter wavelength at upper cutoff, which would be about 4mm @ 20kHz.
Picture a single lentil suspended from a length of your favorite brand of speaker cable--that's your point source.
So the giant mouth of a WE 15A is about as far from a point source as anything I can think of... ...Could call it a rectangular plane-source with widely varying degrees of wavefront coherence and curvature, I suppose...
Just sayin'.
By definition, a true point source is omnidirectional throughout its bandwidth, and therefor its size must range between infinitely small and perhaps a quarter wavelength at upper cutoff, which would be about 4mm @ 20kHz.
Picture a single lentil suspended from a length of your favorite brand of speaker cable--that's your point source.
So the giant mouth of a WE 15A is about as far from a point source as anything I can think of... ...Could call it a rectangular plane-source with widely varying degrees of wavefront coherence and curvature, I suppose...
Just sayin'.
Yeah, that one had me scratching my head, too.
But maybe it's a slice of the 360 deg sphere generated by a point source. No a perfect slice, it's got to beam and diffract some, but a good approximation of what that section of the omni-sphere would be.
Um... Cal,
That's not my post at all...
Mine involved a lentil. Seriously.
Bill, I think Pano edited your post 640 instead of quoting it. If I am right, this is your original post?
By definition, a true point source is omnidirectional throughout its bandwidth, and therefor its size must range between infinitely small and perhaps a quarter wavelength at upper cutoff, which would be about 4mm @ 20kHz.
Picture a single lentil suspended from a length of your favorite brand of speaker cable--that's your point source.
So the giant mouth of a WE 15A is about as far from a point source as anything I can think of... ...Could call it a rectangular plane-source with widely varying degrees of wavefront coherence and curvature, I suppose...
Just sayin'.
Bingo! There's the lentil!
Tom,
One of the things I love about your posts is your references to some of the most wacky, fascinating audio experiments I've ever heard of. I envy your life experiences.
No reason at all.
Well, maybe a few:
1) reflected spectral balance matches direct (also enhances "they are here" illusion).
2) EQ can be used with complete impunity, because it corrects the same problems everywhere in the listening area.
3) Absence of self-cancellation, lobing (before room reflections)
4) More I'm not thinking of....
Horn folks, of course, will cite room-related potential pitfalls...
Well, the point-source ideal is just that--an unattainable ideal--at treble frequencies. At 7kHz, even just the diaphragm of a 1" dome tweeter is greater than half a wavelength in diameter and no longer can be considered a point source, never mind the width of the bezel and rest of the box.
I can't think of any speakers that actually are treble point sources. The only exception I can think of is the corona discharge on the tip of the spike in a plasma tweeter. Maybe some esoteric BMR drivers can approximate half-space radiators...
So most speakers that claim point-source really aren't at all.
So many variables--no telling what's going wrong in your room with a compromised point-source approach.
But horns/waveguides seem to do a better job of the "you are there" illusion--taking you into the performance space instead of bringing the performers to you.
I can't think of any speakers that actually are treble point sources. The only exception I can think of is the corona discharge on the tip of the spike in a plasma tweeter. Maybe some esoteric BMR drivers can approximate half-space radiators...
So most speakers that claim point-source really aren't at all.
So many variables--no telling what's going wrong in your room with a compromised point-source approach.
But horns/waveguides seem to do a better job of the "you are there" illusion--taking you into the performance space instead of bringing the performers to you.
Last edited:
Tom,
One of the things I love about your posts is your references to some of the most wacky, fascinating audio experiments I've ever heard of. I envy your life experiences.
Hey thanks, experimenting and measurements often hint at the way to new solutions.
In that case, I had to find a way to measures sound that was at 1600 degrees C, so getting it out of the furnace to where a microphone could be placed was the problem.
Fwiw a point source is one that has a single origin and the horizontal and vertical planes radiate equally and the SPL falls at the inverse square law.
In the simple case, it is omni directional but if one places a source well behind a wall with a large opening, the sound that radiates through is only a portion of a sphere.
A small source in the ground, against a wall or in a corner are cases of fractional space spherical radiation, a conical horn is of that family as well, just a smaller fraction of space.
These can have an advantage in that they can radiate as a simple source as opposed to an interference pattern comprised of lobes and nulls like a larger or multiple source.
In larger speaker systems it is the interference pattern that makes them sound funny when the wind blows, it is why the frequency response changes with distance and position. For those reasons a powerful full range point source can sound much more musical than the traditional array type, sounds the same everywhere in pattern and the wind has hardly any effect at all on the sound.
In the living room, how a speaker radiates governs how easy it is to localize it’s depth location when your eyes are closed. Having the source location be easily identifiable is not good for stereo imaging.
Here too simple is better. Try a pair of little fostex full range drivers on a large flat baffle, these radiate as a simple source over much of their range and can produce a wonderful stereo image as a result. While limited at each end and in loudness, they work well up close.
Best,
Tom
That was my point exactly, that it gets closer to being a point source than other designs in the most crucial bandwidth. It makes other comprises like perhaps some deviation from on axis response and less than ideal dispersion compared to modern multiway designs. But the folks here are simply stating their preference for WE's performance parameters over those that are thought to be related to performance and subjective listener preference.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.