Square Wave Response, Measured Acoustically

Interesting, Mark. What would you guess was the approximate dB level at the measuring position?

Lee

Hi Mark
While it would be a sign of a big problem in electronics, not being able to reproduce the input waveshape of a complex signal (like a square wave) is not an expectation with home loudspeakers, particularly if they have more than one driver in play.

Also, like many subjects there are at least two ways to look at it. For example, the person who lives in theory will say to produce a sq wave requires a bandwidth DC to light and if one considers the math, that is how it looks.
A person who works with electronics will also suggest that on an oscilloscope, that one can get a “perfect” looking square wave with bandwidth extending from 1/10 to 10X the center frequency.
Like looking for an asymptote point, there is the practical and theoretical requirement.

Why is this “hard’ for a loudspeaker?
In the late 70’s Dick Heyser wrote a number of papers, the first to identify a loudspeakers acoustic phase. He suggested that one can consider a loudspeaker’s acoustic phase as something like it’s origin (in time) moving forward and back with frequency. He also developed the first way to measure acoustic phase the TDS or TEF machine.

To preserve / reproduce a square wave (by eye) one needs a large space above and below where the amplitude and time do not change, all the important harmonics have to emerge at the same time.
Loudspeakers on the other hand spread signals out in time, they have an all pass response, when you go to a multi-way system that spread in time from high to low is generally greater still.
It is asserted that this kind of time error can’t be heard (with headphones) and so it is ignored.

In 1980, the company I worked for got a TDS machine and B&K mic for me to use. Being a loudspeaker fan and having just started developing the Servodrive woofers I measured everything I could even non-loudspeaker stuff. Eventually I was presenting the goofey stuff at TDS seminars and was then asked to measure the Great Pyramid in Egypt (if curious here is a story about that)

http://www.livesoundint.com/archives/2000/julyaug/pyramid/pyramid.php

I had been collecting horns and drivers for some years with the thought that someday I would build a large horn system for my home as they would have the peak capacity none of my home built electrostatics had. . The inspiration was the jbl375 and big len’s I had accumulated. These, by measure were essentially “perfect” through the mid band, the problem was I could not see any way to make them part of a system because of the spacing between them and what they mated to.

Working on acoustic levitation sources, it was clear to me there was a fundamental difference between how sources acted when they were less than a quarter wavelength apart and more than that. Closer, the sources added coherently into one source, much greater and the sources do not add coherently and produce an interference pattern.
At wavelength differences, reversing the drive to one source causes NO change in radiated power, only the shape of the interference pattern changes so at best most crossovers are concerned with the on axis lobe only.
I did not see any way to add the ranges in a multi-way that they did not interfere acoustically and over the years I had a good stock of parts but no solution..
So far as a square wave, there had been only three loudspeakers that I knew of first hand that could reproduce a square wave unilaterally. By that I mean that usually one can find a combination of frequency and mic position where one see’s a square wave but if you change anything, it is lost.. This is especially true in multiway systems where the change in paths with position are measurable physically and acoustically.
In your test, unless you had highly directional speakers, one would have a lot of room arriving at the LP.

If one were close enough for the direct to swamp the reflected sound, or outdoors then one could more reliably see what the speaker really does.

The commercial speakers were an 8 inch full range JBL driver above the lower mid, up to a few Khz, the ESL-63’s I repaired for my boss and the most recent a Manger bending wave transducer above a few hundred up to several KHz.
These speakers were much less sensitive to the physical position of the microphone so far as the square wave shape. The ONE thing they had in common was they radiate as a simple spherical segment over a wide range, they measured as having one location in time over that same band.
If you moved the mic, you do not change a significant change in wave shape with position because they radiate as if they were a single point, over a wide band.

Anyway, I had come back from Egypt and a friend teased me about being in the pyramid and the point was “what are you going to use a pyramid to make the drivers sharper” haha big fun.
Being faced with needing a design for a full range loudspeaker for commercial sound, I thought back to an old Synaudcon lecture where Don talked about the up side (constant directivity) and down side (inferior loading at lower frequencies) of conical horns. Funny, being deep into subwoofers and prevented by edict from working on “normal” speakers at Intersonics at the time, I thought, paahhhh, this doesn’t apply to me.
Flash forward, the Shuttle stuff ended, starting a new business on my own constant directivity would be VERY desirable especially for commercial sound, the light went on, I thought the difference in loading might be related to the local expansion rate. A quick calculation suggested this was the case. I made a simple pyramidal horn and mounted some drivers on the side where the expansion rate looked right for mid range. After a number of experiments to get a feel for how the speakers coupled together, the very first Unity horns were born. About 12 years later is the present, the Synergy horns have evolved into what Dick Heyser had suggested was possible when I was a sophomore in High School.

Here in a post from some years ago is an SH-50 at a meter in a living room, a square wave at 250, 1K and 2.5KHz, square wave in, mic sig on scope. These also radiate as a single source, the sources sum coherently and do not produce an interference pattern (lobes and nulls).

http://www.diyaudio.com/forums/multi-way/71824-making-square-waves-2.html

The driving need for our business was to make all the sources combine acoustically in a coherent way is so that one can radiate as if it were one source with a much larger acoustic power. When you do that, the sound spectrum is not only constant with left or right position but in depth as well (sans hf absorption at large distances) because one has satisfied the distances needed to add coherently into a single spherical segment of radiation over a wide band.
If you do not have constant directivity, then the spectral balance changes strongly with location (think line array outdoors) and inhomogeneous radiation causes the sound to change moving side to side or when the wind blows as well.

Hey, you like big sound, I have not done the square wave test with these, but they should reproduce one over an even broader band.
Two will do a 40,000 seat stadium nicely, four in a large stadium like the second link. With headphones, you can get sort of an idea what it sounds like to have one uniform source though. They went back and measured later, they were 700 feet at the far point.

http://www.facebook.com/pages/Danley-Sound-Labs-Inc/126113687424773?ref=ts

Big stereo, I kept thinking, they really need to show movies here.

http://blog.mixonline.com/briefingr...ty-to-brigham-young-university-football-fans/
Best,
Tom Danley

Mark (Basspig) Weiss said:

So here is what I got:

Click to expand...

Interesting, and not too shabby considering. Maybe try again, but only 2-3 feet in front of your speakers? That way you'll exclude much of the room's influence.

--Ethan

Ethan Winer said:

Interesting, and not too shabby considering.
--Ethan

Click to expand...

That was my thought. Especially now that we know it was at 85 dB. I wouldn't expect that from most systems.

Lee

Mark (Basspig) Weiss said:

I was shocked at just how hard it was to find a source of a clean square wave to start with!

Click to expand...

You don't have Sound Forge or equivalent? Or am I missing something? I create all wave types in SF. It takes 5 seconds. You pick the frequency, duration, level, and wave type. Boom, you save it, all done.

--Ethan

Hi Mark
“Tom, that was a fascinating read, but could you post the scope photos here? I could not see them as the forum required login and I don't have an account there.”

Sure, or at least I will try to attach them. I took them on third octave centers but will only post a few.

“My room is actually pretty dead at the front end, and even at the front row of seating, I'd guess that reflections are pretty far below direct radiation.”
Cool that makes a big difference, close reflections are bad.
“Of course, an anechoic chamber would be the best way to measure speaker performance. Placing a mic 8' away in an indoor venue adds a nearly infinite array of secondary reflections and even small amounts affect the waveshape radically.”
Yeah, outdoors (weather and circumstances permitting) is as close as you can come to “the speaker by itself”. Ones best hope is to have close up absorption and or significant directivity, the farther away the reflections occur from the speaker, generally the lower in SPL they are relative to the direct.
Best,
Tom

Hi Mark
“I've been under the impression, for a number of years, that the better the squarewave reproduction, the more accurate the sound reproduction. With good room treatments, so that you hear mainly the speakers and not the room, it should be possible to recreate the concert hall experience with palpable accuracy.”
Well to do what you did requires more than one thing to be right. Every link in the chain has to preserve that signal shape, the biggest errors at the last two steps..

The signal makes a good visual because it is a repetitive signal with a simple geometric shape and is comprised of a fundamental and many harmonics.
If the amount of OR phase of any of them are perturbed relative to the original, what comes out doesn’t look like a square wave.

So, first your speakers have to have flat response pretty far above and below the fundamental F.
Next it also means the various harmonics are arriving at the right time. Below the hf, most speaker begin to spread out in time and cannot preserve the input waveshape like that.

Since you measured that at the LP, (assuming it isn’t a spatial coincidence and gone 6 inches away), you would have a low level of reflected sound. You mentioned an LEDE room, I suspect you have loudspeakers which have significant directivity as well.
You mention you have tons of headroom, I would reply (in my best C. Montgomery Burns voice) EXCELENT!
So far as a speakers stereo image in a room, being “in time” over a broad band and directivity /low levels of reflected sound are both very good things .
Best,
Tom