Horn Honk and Towels

[Pano]

Fans of the Hitchhiker's Guide to the Galaxy will know to always bring along a towel, for many of the reasons given in the Guide.

Here is another.

A few months back I was conversing with Jack Bouska about horns and "Horn Honk". Jack has a clever experiment outlined here: Factors Affecting Sonic Quality of Mid & HF Horns & Waveguides (Go down to the 3rd post)

The idea is to better terminate the exit of a horn or waveguide. A towel is wrapped around the horn mouth in this experiment. It works.

I had been meaning to do this for months, but hadn't gotten to it. Last week I taped some rolled up dish towels to the flares of my Altec 811 horns. It worked, as advertised. Now the 811 is not a truncated horn, like so many smaller pro designs, or the megaphone in the example - so I didn't really expect much. And I didn't get "much."

The results were a bit surprising. The sound was different, a subtle but real difference. It sounded a bit flat and dull. Not what I had expected. But since I wasn't covering the mouth of the horn at all, that difference must mean that something had been removed, and not something that was supposed to have been there. I guessed that was good.

Then I went back and read Jack Bouska's posts again. He says:

While talking through the magazine horn with the towel, it will sound flat or dead at first, but if you quickly remove the towel cowling, the honk returns instantly.

After a few trials, it will become apparent that the sound with the towel is neither dead nor flat, but rather neutral and tonally smooth, especially compared to the naked magazine megaphone.

Yep, that's it. The way the horn mouth terminates sure does make a difference.

The first horns I heard in a good Hi-Fi rig where the wooden Iwata horns. They look similar to an Altec 511, in wood, but have diffraction slots at the lips. IIRC, the slots start narrow inside the flare and get wider as they approach the rim. This is done to smooth the impedance transition from horn mouth to free air. Unfortunately, I can't fin any online documentation of the Iwata slots. Maybe someone knows where, perhaps a Japanese site?

I encourage you to try this simple experiment. With an abrupt flare termination, like a megaphone, it makes a big difference. With better mouth flares, the difference is smaller, but real enough.

Some distortion removed. Good thing.

Open for discussion.

[Sheldon]

Thanks for posting that. It's an very rich source of information on horn behavior, backed up with actual measurements (what a concept). One passage struck me:

"Midrange Cone transducers vs. horns and waveguides:
I want to remind readers that compression driver distortion (2nd and 3rd harmonic) through the upper midrange (500 Hz to 2 kHz) is typically much higher compared to modern cone midrange transducers. As an example, see figures 9 and 10 in the jbl tech note V1#8: http://www.audioheritage.org/vbullet...ead.php?t=4410 (http://www.audioheritage.org/vbullet...ead.php?t=4410)
And compare this to the much lower distortion from the jbl 2012:
http://www.jblpro.com/pages/pub/components/2012h.pdf (http://www.jblpro.com/pages/pub/components/2012h.pdf)


Compression drivers with high compression ratio’s generate strong harmonic distortion when high acoustic amplitude levels exceed the capability of the air (trapped between the diaphragm and the phase plug) to act as a linear spring. High compression ratios do improve efficiency; however the penalty comes in the form of increased distortion. "

I have a Unity horn system, wherein the horn consists of a TAD 2002 and 4, 5" midrange drivers. The midranges cover from about 350 Hz to about 1200 Hz. This is crossed over to an open baffle 15" woofer. Since my open baffle woofer can easily go up to at least 2kHz, and the directivity match should be better at crossover frequencies higher than 350, I tried various crossover points between 800 Hz and 2kHz, using only the compression driver. The benefit is one less crossover (active or passive). But after a short while I've always gone back to using the midrange drivers, because it just sounded smoother and more solid. Maybe I'm not imagining things after all.

Sheldon

[pdan]

I don't know if this is anything or not: what I've in mind is the difference in the tone between a "Bach" trumpet and its modern version.

The modern sounds much brighter and more aggressive giving full throat to its 3rd harmonics. The "Bach" on the other hand is more reticent.

The explaination given is that the "Bach" trumpet is hand made, with the metal hammered over a former; and that it is the random patterns of small indentations caused by the hammer work that disperses and diffuses the 3rd harmonics. The modern by contrast is smooth and regular.

Don't know tho' if a front "Bach" horn would be any good - perhaps it would create other problems ?


Cilla

[Pano]

@ Sheldon. I've seen your work on the unity horn, very nice! Complicated to build, though, isn't it?

I do love the Unity Horn idea. My guess is that it would suffer the same mouth diffraction problems as a single driver horn, but maybe not the internal problems since some of the sound is introduced radially.


@ pdan. The trumpet thing is interesting because it jibes with what I learned about audio horns years ago. Often wool felt or even sand paper was placed in in the horn, usually near the mouth, to create a sort of turbulence near the inside surface of the horn. The layer of turbulent air was meant to reduce reflections off the walls, IIRC.

John in CR says he gets the best results by lining the horn walls with damping material.

In the is bit of EAW Advertising, they talk about "Horn Honk" and horn "Splashiness" - a term I didn't know.

splashiness is usually heard at the highest frequencies and obscures the fine detail in instruments, such as cymbals

The towel on the edge of the horn mouth sounds to me to be reducing "splashiness", but I couldn't swear to it. Further test are needed.

[Magnetar]

You may want to look at the Peavey Quadratic Throat Waveguide - Foam is used on the mouth

[gedlee]

Originally posted by panomaniac
@

I do love the Unity Horn idea. My guess is that it would suffer the same mouth diffraction problems as a single driver horn, but maybe not the internal problems since some of the sound is introduced radially.

To me the problem with the Unity is exactly that the sound is injected radially which will tend to set up strong Higher Order Modes (HOM). To me the idea is to reduce the HOM and the mouth reflections as much as possible through design. The unity horn does not fit this criteria.

[Sheldon]

Thanks for the kind words panomaniac. They would be somewhat challenging, but doable with a good saw and a program to compute the angles. I bought mine from Nic about the time he shut down Lambda.

After reading Earl's work on HOM's I bought some foam (maybe shouldn't say as I don't know that the provider wants to be in that business). Some slight "honk" before the foam. None that I can detect now, though I haven't much experience so take it for what it's worth.

Sheldon

[Pano]

Does anyone have any photos of foam in horns?

Are we talking about a sort of phase plug, or foam on the lips like the Peavey horns linked to in Magnatar's post?

I had seen the Peavey document before, it was the 1st thing that came to mind with the towel trick. Has anyone heard them? Maganatar?

Wonder if screen or mesh would do well as an impedance adaptor or HOM buster?

[gedlee]

Originally posted by panomaniac
Does anyone have any photos of foam in horns?

Are we talking about a sort of phase plug, or foam on the lips like the Peavey horns linked to in Magnatar's post?

Wonder if screen or mesh would do well as an impedance adaptor or HOM buster?

I have attached a photo of the Summa. The foam is quite evident in the waveguide.

I tried about five densities of foam and at least that many shapes, took complete polars, etc. The density is a critical factor, too dense, too much loss, too open not enough effect, but the shape played a small role - works like an acoustic lense shaping the waveform.

Small features like screens don't have much of an effect. You can see that my foam is about 7 inches thick from front to back. This has a major effect. I even tried wall coatings, but again, not enough effect to be worthwhile. Make no mistake there is a price to be paid for the foam, it takes about 3 dB out of the high end that has to be EQ'd back. Fortunately with a awaveguide and compression driver we usually have a lot of output to spare. But believe it or not once the response is CD and the foam is used the top end of the waveguide has about the same output as the 15 woofer - no more. Its the limiting factor of efficiency.

[Sheldon]

Shown in the Unity. Goes all the back into the the throat of the compression driver, about 12" worth.

Sheldon