Khorn attenuation/distance less than inverse square

[KhornKerry]

Khorn attenuation/distance less than inverse square law

I measured the attenuation verses distance of my Klipschorn and found it did not fall-off as inverse square law would predict.

The Blue line is measured SPL of one Klipschorn using a Radio Shack digital Sound level meter. The meter only has a 1 dB resolution. The sound source was white noise from my Denon HT receiver, using the speaker level setting. I chose white noise to reduce room effects and standing waves. I set the volume such that it was 100dB at 1 meter. Measurements were taken on the axis line of the midrange (squawker) horn. The room is about 16 X15 feet, sheetrock walls and ceiling, carpeted floor and a vaulted ceiling.

The green line is calculated inverse square with 1 meter distance as the reference.

The red line is calculated inverse square with .25 meter distance as the reference.

I believe the Khorn attenuation was less than predicted because the horn directs or concentrates the sound energy in one direction. Inverse square is based on a point source and a horn doesn't act like a point source.

I tried to convert the Excel to a JPEG .

[mace]

Interesting test!

However, if I'm not mistaken, white noise is just all frequencies together from 20 - 20,000 Hz. So, if your room is prone to standing waves at certain frequencies, say 1000 Hz, then there will indeed be a 1000 Hz boost in the room when white noise is playing. To truly get rid (for the most part) of standing waves, the tone must be constantly changing (warble tones). The wave then doesn't have "time" to "stand around". This type of test is only valid in an anechoic chamber since otherwise there are always room effects, and usually lots of 'em. Even going outdoors isn't good since you get reflections off of the ground at certain frequencies more than others at certain locations.

Plus, the speaker driver itself can't control how much attenuation the sound wave has after the sound wave is launched into the room!

Mace

[mace]

One more thing... what formula were you using for attenuation??

Thanks,

Mace

[JohnA]

The speakers cannot fail to fall off at an inverse square rate. However, you did your test indoors with room reflections added in. That is the cause of your odd measurements. If you can get a sine wave generator and use a single frequency, you should see the proper fall-off with a few bumps as room modes show up.

[KhornKerry]

John

Inverse square only works for a point source, a horn concentrates the energy more or less on one direction. For example a laser beam would not follow inverse square nor would a parabolic radio antenna, or for that matter even a flash light with a parabolic reflector.

As for pure tones room reflection causing standing waves will cause all kinds of measurement problems, ( believe me I have tried it).

I agree that a better test could have been done outdoors.

I would like to see some one do this test with standard cone speakers for comparison but all my speakers are klipsch.

[Poindexter]

Field trip to the anechoic chamber in Indy!!!

Just a general question. . . what are you measuring your dB in

(i.e. base 10,2,e; dBm; dBV) and with what weighting??

Cause you may have to convert to watts (I may be wrong but that's what they taught us in our signal modulation and demodulation class in engineering school)

[WMcD]

I agree one issue is that the speaker may not be acting as a point source. Remember that an infinite line source falls off in direct relation to the distance. An infinite plane source doesn't fall off at all.

We're not quite to infinite anything, but we're not at a point either.

I believe another important issue is that we're by no means in free space. Actually we're pumping the wave into a chamber, i.e. the room. Standing waves are an issue. E.g. the sound will bounce off any and all surfaces. So to some extent the listener is surrounded by large plane sources of echos.

Gil

[Poindexter]

well put. . .

Perhaps we can do a Matlab model with the KHorn as a finite sum of point sources.

I really don't feel like programming that right now though. Any volunteers??

[Poindexter]

On another note. . . I'm such a nerd

[Poindexter]

Plus, the farther away from the source, the reference point begins to "see" the source as a point. and the closer the reference to the source, the source begins to "perform" like an infinite plane. But farther away from the source, the presence of reverberations becomes greater.

[WMcD]

And the above is probably not even half of the variables.

PWK reported on the horizontal interference effect from the K-Horn having two mouths. I expect they become directional in the vertical too. At 400 Hz they're both about a wavelength tall, and the floor reflection might double that. Some theory would predict the output of the K-33 is worse than it turns out at the high end, and this may be why it is not.

He alluded to the fact that the K-400 narrows the vertical pattern above 3 kHz. In my tests most midrange drivers fall off above 3 kHz and the narrowing helps it.

The polar pattern of the K-77 is narrowing at high frequencies based on the polar response published by EV of the T-35.

Put that into room problems and it is a wonder anything works out near correctly.

Gil

[Poindexter]

Simply amazing!!!

Gil: So how did you learn all this?? I'd like to know some of the Klipsch technical history as well.

[KhornKerry]

Mace: the equation for attention I used

dB = dBref - 10 X log ((D2/Dref)squared)

dBref = SPL(a weighted) at 1 meter from speaker

D2 = second measurement point (distance from speaker)

Dref = distance 1 meter from speaker

-----------------

Poindexter:

base 10, A weighted (Radio Shack SPL meter on fast response)

-----------------

Gil:

Standing waves should not be an issue with random noise. Reflected noise should only add a little to this measurement, example if I turned on another speaker with noise at the same volume the measured SPL should only increase 3 dB (twice the power). IMHO

-------

I plant to repeat experiment this using a small standard cone speaker from Radio Shack. Hope the RS speaker can be driven to 100dB SPL at 1 meter.

[mace]

With white noise, nodes will build up in your room at certain frequencies. At one position, you may see higher dB due to a node from frequency X, at another position, frequency Y may show up, at another, maybe weaker nodes. That radio shack meter is just showing the level of the highest SPL frequency.

Heck, I can hear the high frequency whine of my TV and I move my head about 2 cm and it's gone, move my head back and it's there. All rooms are just chalk full of nodes (destructive and constructive) at all kinds of frequencies.

Also, I'm convinced the Radio Shack meter isn't linear. Someone posted a "correction" factor spreadsheet to this forum once. I may be able to dig it up. i can't speak for it's accuracy, though.

Mace

[WMcD]

Rich, I didn't learn this from any one publication or theory of study. PWK, or Olson, Keele, or Baraneck (big names in the field) never put it together in a neat package. No one can. We're talking about system analysis which goes into the complexity of the driver, the horn throat, the mouth, the radiation patterns of the horns, the room (big factor), the ability of the ear to sort things out (big factor).

There are some well respected studies on any given subject. People hook up on them and may reach some conclusions or have some questions.

I'm a fan of the amature experimenter. I'm one. It brings us closer to the truth in our mind.

So, we find the inverse square law doesn't hold in our room. Well enough. We've learned something.

It is also instructive to find a test CD with bands of constant frequencies and walk around the room using an RS meter or just our ears to hunt for nodes and anti nodes (peaks). By doing this you'll find, as I did, that rooms and locations have their own frequency response quite independent of the speaker. Also, try adjusting the location of the speaker. Corner, wall, or out in the room.

I do suggest you try this. You'll be moving into the scientist mode rather than just "reading about it" mode.

Hope I'm not too haughty. Run your own expreriments.

Gil

The problem is that some

[djk]

http://db.audioasylum.com/cgi/m.pl?forum=hug&n=40030&highlight=critical+distance+djk&r=&session=

"Outdoors the inverse square law holds up, indoors it only hold up to the point where the direct energy has fallen to the level of the reflected energy."

"But if horns aren't any louder inside than direct radiators, why use them? With enough power even Bose will play loud enough?

Critical distance.

The reflected energy becomes louder than the direct energy beyond a certain distance in any large room. The level of direct sound still continues to drop off following the inverse square law, but the reflected energy remains constant. Once the level of the direct sound has dropped 12dB or so below the reflected sound speech becomes difficult to understand, music isn't as bad. You've all been there before: it was louder than all get out, but you couldn't understand a word being said (think big Catholic church). A horn improves signal to noise ratio, so to speak, by reducing early reflections. Because the horn is more articulate, it sounds better."

[Poindexter]

Thanks Gil.

But sorry, by the details in your post I thought you got a hold of some of PWK's patent applications or something. Yeah I used to do my own tests back when I was in school; since I had free access to millions of dollars of equipment (signal generators, spectrum analyzers, TDRs, etc). That's where I put my AR bookshelves to the test and added some filters to each speaker. Sounds much better now.

I guess I'm just an engineering sucker. I always want to see the well documented evidence. Sure, I guess I can experience it, but I always wanna see it in a controlled environment as well.

[Ray Garrison]

Hi Gil. Long time no chat.

"I'm a fan of the amature experimenter. I'm one. It brings us closer to the truth in our mind."

Hummmm..... that's, uh, actually pretty deep.

Does that mean that by experimenting we bring our minds closer to understanding some external absolute truth, or does mean that by experimenting we bring our understanding closer to the truth that exists (in some sort of Platonic sense) in the mind?

[mace]

Horns aren't just about being loud. With an efficient horn, the driver moves less than a radiator type driver to produce equal SPL. Less motion equals less distortion. So, not only do you get nice SPL levels, it is clean as a whistle! (providing your source material and amp is good!)

Mace

[WMcD]

Hi Ray, good to see you again.

I don't have much of a mind for philosophy and certainly did not mean to say anything "deep". I could not hope to make a judgment on whether abstractions exisit outside the mind, or only in it. Who can tell, and how can we test it? My overall position is that if philosphies can not be tested, they're not science. Of course some allowance must be made for cosmology. We'll never be able to test the big bang theory in a realistic sense.

My point in being an amatur experimenter is that we can gain some first hand experience which may show up that much of what we're exposed to is advertising hype and overoptimistic test reports, or underoptomistic ones. PWK was annoyed with these. Hence the BS button.

Going to the concerns of EE. I have read all the patents. The later ones are a bit more instructive on horn design. But again, not all the info is in one spot.

Basically, the horn is a type of transmission line. You must have taken Field and Waves. At IIT it was basically about transmission lines. And Smith Charts. Most of it assumed one frequency.

PWK published a paper which included an equation for the impedance transfer function for a finite horn. He did this to demonstrate that below cut off Fc, there is still appreciable loading at the input side. So from this we can infer there is transmission of energy down the horn.

PWK also refered to a paper by Thule about the driver parameters which give a good impedance match at the throat. This used voice coil resistance, amplifier output impedance, magnet strenght and voice coil length. But you can't find these data in spec sheets for drivers.

Much later, Don Keele published an AES preprint which showed that T-S parameters of the driver, Fs, Qts, Cms, which are published, can be used to calculate the optimum throat size for a given driver. So, essentially, this allows one to calculate the input end of the transmission line.

The other part of the transmission line question is the impedance at the mouth, the output side. We want to optimize that too. Keele showed that the ancients were correct in suggesting that a mouth that is about 1/3 of the wavelenght in diameter at Fc is pretty good in free space. He also did some calculations about half space and a corner, which is 1/8 of free space.

Now. All that goes to horn design. It does not account for wave mechanics in a room.

A very important paper was, again, by Don Keele, "What's So Sacred About Exponential Horns." In it he describes the theory behind a constant directivity horn and the relation between mouth geometry and directivity.

If you want copies of these papers, please contact me through the forum.

Gil