A horn that's undersized becomes a sort of horn/transmission-line hybrid. That in and of itself isn't bad, and in fact, with the right tuning can be very good. But there's a difference in implementation that is significant. Transmission lines are damped to reduce unwanted standing-wave modes. They are designed to encourage some modes and attenuate others, using careful driver placement and damping. Horns, on the other hand, are usually made without damping and with the diaphragm at or near the apex of the horn. They are so-called wide-bandwidth resonators and so resonant modes aren't damped. The designer intends for the horn to load the diaphragm resistively, not reactively, so it is expected to act as if it weren't a resonator. But the undersized horn does act reactively at low frequencies. That's why I've characterized it as a horn/transmission-line hybrid.

A horn decreases diaphragm motion in its passband, so distortion is reduced. It increases diaphragm load and that improves efficiency. Between increased efficiency and increased directivity, a horn amplifies sound volume in its passband. All that is good news. But when a basshorn is undersized, its loading decreases as frequency drops. Worse than that, it fluxuates loading instead of being uniform. This tends to create ripples in response.

Another thing that can hurt performance of an undersized basshorn is using it below the horn's passband, in a direct radiator mode. This is tempting to do in order to increase bass extension. It may be a worthwhile compromise in some cases, but it should be noted that this mode increases distortion more than a direct-radiating front-loaded design. The reason is pretty simple - The horn amplifies harmonics, so it amplifies distortion. Fundamentals in direct-radiator mode are not amplified by the horn, but the harmonic distortion is. This can result in harmonic distortion that is actually louder than than the fundamental.

A well-executed basshorn provides high-output and low-distortion. If made large enough, it can provide deep response. Sometimes smaller horns can be used in groups to achive the necessary size. But when a basshorn is used too low in frequency, it can actually sound worse than a direct radiator. It is good to look at the matter realistically when deciding what kind of bass system is best for any particular installation.

The thing that Bill Fitz always harps on in his articles is "look at the output in the passband". That's all well and good in a club but the horns he builds are almost always best for kick drum on up - that 60-70 Hz region. For music reproduction you need good response at 40 Hz and that just makes it impractical listening at 10 or 12 feet away. Edgar has his fancy horn designs but he always has to have a sub for realistic reproduction. ALE horns go all the way down, but with a mouth at 5 feet X 5 feet they're pretty hard to accomodate in a ranch house.

Every design I end up admiring looks like the 4Pi - which looks like Lansing's original Iconic - which looks like the JBL K-series speakers - which looks like every Westlake plus a heavy doese of steroids. For audio in a domestic environment it's simply the best way to go.

Don't get me wrong, I love horns. And I've made some really great basshorns that I'm proud of. But the problems that present themselves when designing a basshorn are challenging. Seems like people tend to want to ignore some of the facts about horn loading when it comes to basshorns. This is true especially when they want to make their basshorns portable. The easier a speaker is to carry, the more popular it will be, especially if you can push the idea that it works as well as a real horn of adequate dimensions.

Fitzmaurice's basshorn designs are usually pleasantly small and easy to carry. But they also tend to suffer in terms of quality, having response curves with lots of peaks and dips. Further, his designs depend on the woofer as a direct radiator for the lowest frequecies. In my opinion, the way they're tuned pushes the driver way too hard, way beyond its mechanical limits. The rear chamber is large and Fitzmaurice generally suggests light-cone, low-excursion drivers. That approach is acceptable for a light-duty system, but I wouldn't suggest it for high-power basshorns.

His approach makes a couple very unattractive things happen. One is driver over-excursion. I expect that's probably the main failure mode in Fitzmaurice designs. The other is distortion. Since he often suggests that his basshorns be used well below the frequency where the horn is useful, it basically becomes a direct radiator in a large box. So driver excursion goes way up, distortion skyrockets and the harmonics produced are amplified by the horn. In a design like this, distortion is louder than the fundamental for a good deal of its range. It would be interesting to see distortion measurements of Fitzmaurice designs, but he doesn't have any good measurements of his horns, just graphs made using primitive sound meters.

I've also seen the opposite argument made just as forcefully. Tom Danley was on a crusade pushing low-fs drivers a while back on Audio Asylum. I think his basic deal was that his Servodrives had high-mass drivers because of the fact the motors were pushing belts and pulleys in addition to cones. So he was forced into finding horn solutions that worked with high-mass drivers. Of course, on a bass system, mass isn't your enemy like it is on higher-frequency drivers. So he was able to design with it pretty successfully, and when using linear motor drivers, he tends to choose drivers with similar specs.

Danley pushes the idea that basshorns should have high-mass, low fs and loose suspensions. His opinion is that the rear chamber volume should be used as the main suspension force and that the driver's mechanical suspension should therefore be loose. Danley suggests that the driver's suspension isn't uniformly compliant, and that's why he likes using the rear chamber volume as the main suspension instead. But to tell the truth, I don't think acoustic suspension is the real reason Danley likes this tuning method. I think Danley likes high-mass drivers because that's what the Servodrive motors have, and it is what he is comfortable with. All of Danley's newer designs tend towards this same approach, with a twist like using boundary reinforcement to augment mouth area or making a tapered quarter-wave pipe having similar proportions and characteristics.

Edgar is a different breed, since he doesn't ask as much of his basshorns. He does't ever expect them to see more than a handful of watts, so excursion is never a problem. He tends towards the same idea as Fitzmaurice, which is to use a light cone and to push the driver well below horn cutoff, using EQ to extend bass response. In Edgar's case, this is an acceptable compromise since the horn will never be pushed very hard. But I do wonder what kinds of distortion are generated by his basshorn designs. I don't think I've ever seen distortion measurements of an Edgar basshorn.

Edgar tends to use high quality drivers, so that helps a lot. I've always liked Edgar's choice of drivers. But his choice to increase output below horn cutoff with EQ is something that even the best driver can't help. Shorting rings are ineffective below 100Hz. So distortion must rise pretty steeply at subwoofer frequencies in his basshorn sub design. When a basshorn is used below cutoff, the harmonic distortion is amplified by the horn but the fundamental isn't. So distortion may be louder than the signal sent to the driver.

For all these reasons, I've generally preferred direct radiating subs. A good horn sub can be made, but it must be large. And I think it wise to employ other distortion reduction mechanisms like push-pull drive. That's how I've made basshorns that worked well, using a combination of each of these things. My thinking is this: If a basshorn is going to be a compromise, why build the basshorn at all?

The thing is, distortion at bass frequencies manifests itself as harmonic components in the midbass and midrange. That's why bass distortion is so noticeable. Look at the Fletcher-Munson curve above and you'll see that it falls rapidly below 100Hz, indicating that our ears become more sensitive to sounds as frequency rises from the deepest bass range. Harmonics above about 100Hz are troublesome, and these are generated as second-harmonic distortion from fundamentals above 50Hz and/or third-harmonic distortion from 33Hz signals and above. So you can see that bass systems aren't exempt from the requirement that low distortion be needed for good quality sound.

I watch all of the horn designers/manufacturers as to their general practices and especially what they are doing within their targeted niche to give their design an edge compared to the competition. It seems that the design goals for PA are actually quite different than the goals of Hi-Fi. For now, I'll throw HT into the Hi-Fi group.

One thing I derived from reading about PWK is that he tended to stay with what worked - he didn't tend to take risks, developed some preferred characteristics and used them repeatedly, so his designs seem related to each other even at a cursory glance. I also like the fact that he tended to publish and in those patents and AES papers, he attempted to teach others about what he was doing and why - anyway, I don't regard it as mere advertising, too much usable information for that IMO. I suppose it lent credibility to his designs, but since the particular designs were being sold, the actual proof was in the performance that anyone could hear.

So with that in mind, I find I also like Edgar, for the same reasons. He is pretty conservative in his designs, too.

The interesting thing is that Klipsch is heading for the HT market with their horns. Halfway between PA and Hi-Fi, I think that is the future of the horn market, too. That's not to say that the PA market is going to go away, but it probably won't grow for horns either.

The dedicated home theatre room is one place where perhaps a large speaker footprint IS not only acceptable, but even expected, or even actually DESIRED! We can only hope...

DM

My take on using undersized basshorns for PA is that it is a dubious compromise. I mean, on one hand, if you use several of them, it's a reasonable compromise. But then again, if you're going to compromise, why even go to the complexity of using a horn? A direct radiator subwoofer is exceptionally flat and easy to use. I think bass-reflex subwoofers get maligned too much by us "horn guys."

I've measured some pretty bad basshorns. Some of them have become pretty popular with enthusiasts that boast proudly about their one note wonders. I guess it's builders pride. Then again, like I said, when you get enough of them going, they smooth out a bit. Funny to watch a guy use two of them though. Yuck!

The thing that intrigues me is the hybrids. The transmission lines with additinal resonators for higher mode reactance annulling. I think that's a pretty good idea, a way to make a horn with a small footprint. Whatever label you put on them, transmission lines, horns, hybrids, half horns, stubs, resonating lines, tapped horns - the idea still has merit. Use the rear chamber for annulling the reactance of the first standing wave node, and then use additional resonators to annul the higher wave nodes. That's a clever approach.

Aaah, but then again, if you have the room for a full size basshorn...

Since the reflex porting eliminates the possibility of annulling the driver, he has come up with a unique approach which uses tuned cavity filters along the horn pathway to reduce the reactance peaks and tuned ports along the channels to increase the response in the troughs caused by the short pathway and unannulled driver. The back chamber is also ported to the front of the cabinet. It's well worth investigating.

DM

• ELENA horn pdf document

DM

• ELENA cross-section plan view