I remember hearing a very good horn loudspeaker for the first time in 1978. It was sheer magic to my ears. I could not believe the incredibly realistic sound that seemed to radiate effortlessly into the room. But when I talk to other audiophiles about horns, the reaction I hear is usually one of disgust. Like a bad hair day, everyone seems to have had a "bad horn" day.
Rich Weiner of "Bound for Sound" recently described his "bad horn" day. "As a neophyte audiophile, I once had the horrifying experience of listening to a friend's horn-based speaker system. "Killer equipment. You've never heard anything like it," he promised. Indeed I hadn't. I found myself sitting about six feet from a pair of Klipsch corner horns driven by Phase Linear 700 amps. My friend was right. I have never heard anything like it, although I understand that standing directly behind a 747 during takeoff is quite similar. Since that time I have avoided horn speakers." (Bound for Sound '99CES Report).
But times are improving. In the same report, Weiner says, "Perhaps it's time to try horns again...Edgar's Titan (Horn) system was quite impressive...detailed and articulate..." Weiner's experience is not unique. Other reviewers and audiophiles are coming around to the concept of an acceptable audiophile horn system.
So you may ask, "What is the difference between the vintage horn systems and the new generation of horn loudspeakers?" I have been researching that question for the last 20 odd years. In the process, I founded Edgarhorn whose main goal is the design and manufacture of audiophile quality horn loudspeakers. But I'm getting ahead of the story. After I had heard my first good horn speaker, I began trying to hear other horn speakers and other pseudo-horns. Most of the ones I listened to didn't come close to that first system that I'd heard.
Eventually, I was able to analyze that good horn system to discover what made it tick. It belonged to a Senior Engineer at the aerospace company where I worked. At the time, I was a research scientist specializing in radio wave propagation and signals analysis. The design came from a group of amateur horn builders on the east coast lead by Ben Drisko in the early 50's. This system used a Drisko folded corner bass horn design, similar to the Klipschorn, and a JBL 375 compression driver on a Western Electric midrange horn. The component that made this system so good was the JBL 375 driver. My engineer friend recently bought a set of my Tractrix horns for his 375 drivers. When we tested them, I was astounded by their ruler flat response from 400 Hz to 10 kHz.
In the early 80's, I set about tracking down all the horn articles and papers and analyzing all of the published and underground designs. I was very disappointed. There was no clear way of horn design. What was even more depressing was the periodic regurgitation of wrong headed ideas of horn design in articles from year to year that has infiltrated into some horn design software that is marketed today. So I embarked on course of experimentation with horn design and construction. I figured that with my background in wave propagation, spectrum analysis, and experimental physics, it should be easy enough to arrive at a satisfactory design. Was I ever wrong. After 20 years of horn building, I'm still finding new ways of designing and building horns. It's the Edison experience of having to go through all the different combinations and permutations of drivers and horns. And my customers keep coming in with different requirements that change my views and shift the proverbial paradigm.
So back to the question, "What made the vintage horn systems sound bad?" There are a variety of problems. I have run into most of the horn problems either in designing my own horns or duplicating somebody else's horn.
The first problem is electronics. Weiner talked about listening to some Klipschorns with high power solid state amps. The amps were one source of the bad sound. Horn systems typically have sensitivities of 100 to 108 dB SPL with one watt input. Even at the loudest sound that you would realistically audition any speaker system (95-100dB), the amplifier is only delivering a watt or so to the horn system at peaks. Most of the time the amp is idling at 100's of milliwatts; yes, I said milliwatts. At this level many high power solid state amps have real problems with crossover distortion. For this reason, I tell people who buy my horn systems to try different amps with them along with their existing amp. The low power requirement of horns means that single ended tube (SET) amps can easily fill a room with sound. However, not all single ended tube amps are created equal. Some SET designs have relatively high levels of distortion that can be easily heard on the horn system. And of course, the horn gets the blame for the distorted sound.
A second problem with horn loudspeakers is that they are easily prone to resonant peaks in their response. I have heard some horn tweeters that made me think that I was being drilled between the eyes. Many horn midranges have a honky megaphone sound that comes from resonant peaks. I have measured many old horn midranges with high resolution spectrum analyzers and have found them loaded with high Q spikes that give each horn speaker its own characteristic coloration.
Bass horns have their own resonance problems. Typically, a well-designed bass horn that can go down to 35 Hz will have a total volume of 20 cu.ft. or more. A company's marketing department will say that it can't sell a big speaker like it. Moreover, the marketers say, "Keep the 35 Hz flare but make it smaller." So the horn length is truncated to make it smaller. The net result is long folded slowly expanding tube that sounds more like a resonant tuba than a wide band bass horn. The response plot looks as a series of harmonically related resonant peaks.
Some resonant problems are caused by structural defects. At the point where the sound comes out of the mouth of the horn, the edge of the mouth can vibrate in a bell mode. If you have seen some old style trumpet PA speakers, they were actually shaped like a round bell. A bell will resonate at a frequency whose wavelength can be wrapped around the circumference of the bell end. No horn is immune to bell modes, but proper damping and tension bracing can eliminate the problems.
A third problem is the use of improper drivers on horns. Good horn drivers require heavy magnets and light weight diaphragms. Most regular speakers used in typical box speaker systems have lower weight magnets and heavy diaphragms. The use of regular speakers on horns will, for the most part, yield restricted bandwidths and irregular responses. Recently, Speaker Builder featured an article touting an easy horn for your dome tweeter. I performed the experiment with a high quality dome tweeter on a 800 Hz Tractrix horn. Without horn loading, the dome tweeter went up to 20 kHz at a sensitivity of 88 dB. With the horn, the efficiency increased several dB, but the response rolled off above 10 kHz. With such a horn loaded tweeter, any listener would ask, "Where's the top end?" My survey of horn construction project articles has yielded many other examples of using the wrong driver on a horn.
As a corollary to the wrong driver syndrome, an associated problem is the use of PA horn systems in home stereo applications. For example, I find many horn enthusiasts over the world using variants of the Altec "Voice of the Theater" (VOT) speaker system. The VOT featured a compression driver on a metal mid horn and a 15" woofer on a front loaded 110 Hz midbass horn. The back of the woofer was loaded by a bass reflex ported enclosure. In general, I have found the Altec compression drivers and woofers to be good quality horn drivers, but the horn design can be improved with my Tractrix horns. And I have done this for numerous customers.
After discussing all the various ills of vintage horn designs, the reader may ask' "Can a good audiophile horn system be designed and manufactured?" The answer is yes, and in my next installment of "Why Horns?", I will discuss the critical aspects of horn design that would lead to a truly worthy horn system. Please email me at firstname.lastname@example.org with your horn questions. I will endeavor to answer them all and work them into future installments.