What’s the deal with speakers with horns?

[Swerd]

I'm rather certain the crossovers were stock. I know the guy who owns the Abbeys, and although he may be called an audio tweak, his tweaking is limited to trying different amps (vacuum tube and solid state) and different cables. He doesn't own a soldering iron, and is generally uncomfortable with the idea of removing screws and opening up any of his audio gear just to have a look inside. He knows that I have done some DIY speaker building in the past, and thought I might be interested in hearing his new toys.

We tried two different amps with the Abbeys, one solid state amp (~125 wpc), and one tube amp (~50 wpc). Both were made by companies with good reputations, and seemed to provide plenty of juice for these rather sensitive speakers. So I doubt if it was inadequate amplifiers.

Back to my original question: what is behind the unpleasant sounding coloration of speakers with horn or waveguide mounted tweeters? For lack of a better description, they sound to me like a musical instrument being played in a large public restroom with tile walls and floor. I could easily hear that sound while a single speaker was playing, regardless of where I was in the room. So I doubt if it is related to sound dispersion, or the lack of it.

It is possible, but I am far from sure, that I hear this with 2-way speakers where the tweeter is horn or waveguide mounted. I don't believe that a 3-way speaker, where the tweeter is horn or waveguide mounted, also does this.

My first guess is that this characteristic sound is limited to a midrange frequency. Earlier in this thread, someone mentioned that horns or waveguides have sometimes been known to resonate at a frequency related to the length of the perimeter of the horn/waveguide bell. An interesting idea, but I have no evidence, such as frequency response, impedance response, or cabinet resonance curves to support that idea.

Dennis – I appreciate your input here. I can only say I wish you had these speakers in your basement, and could run some sweeps.

 

[Steve81]

It is possible, but I am far from sure, that I hear this with 2-way speakers where the tweeter is horn or waveguide mounted. I don't believe that a 3-way speaker, where the tweeter is horn or waveguide mounted, also does this.

I don't really have much more input, but I can offer you some demos that may (or may not) be useful.
KEF Q700: tweeter is coaxial to a midwoofer which acts as a waveguide
KEF Q600c: same as the above, but instead of a 2.5 way, it is actually a 3 way design; I can set it up vertically on a stand if you're so inclined.
Klipsch RF-15s: nothing special, but their little brother, the RB-15s were measured in Stereophile, so you might be able to get a few ideas out of the experience. And if you happen to fall in love with them, they're on sale

 

[randyb]

The speakers I heard were these

Abbey

I assume they are the top of the present GedLee line.

I owned the Abbeys at one time and found them to be very good speakers....effortless dynamics and very clear. I was not enthralled by their imaging. Just for the record, I have owned Salk, and many other speakers. I did not hear any "horn" quality and I did do a quick switch level match between the smaller (Geddes) Harper speaker and the Salk Songtower bookshelf. One thing that stood out was that the Harper had very little bass and all of Earl's speakers need subs. So far the best speaker I have heard that combines pretty much everything for at least a small room is the KEF LS50. Amazing speaker IMO.

 

[Swerd]

I owned the Abbeys at one time and found them to be very good speakers....effortless dynamics and very clear. I was not enthralled by their imaging. Just for the record, I have owned Salk, and many other speakers. I did not hear any "horn" quality and I did do a quick switch level match between the smaller (Geddes) Harper speaker and the Salk Songtower bookshelf. One thing that stood out was that the Harper had very little bass and all of Earl's speakers need subs. So far the best speaker I have heard that combines pretty much everything for at least a small room is the KEF LS50. Amazing speaker IMO.

Interesting, thanks for your observation. Even though I thought the horn coloration was impossible to ignore, it's certainly possible that not everyone hears that. At this point, I can't rule that out.

Maybe that's why, at the audio shows, I always felt like that kid in that "The Emperor Wore No Clothes" story.

 

[randyb]

Interesting, thanks for your observation. Even though I thought the horn coloration was impossible to ignore, it's certainly possible that not everyone hears that. At this point, I can't rule that out.

Maybe that's why, at the audio shows, I always felt like that kid in that "The Emperor Wore No Clothes" story.

Haha. My opinions are just that. Not scientific etc.!! Yours are at least as viable and probably MORE.

 

[DS-21]

I'm familiar with the speakers Swerd heard, and the amplification through which he heard it. And you're wrong on all counts. All parts are stock, the crossover is correctly assembled, the amp is a reputable solid state unit, and they measure in spec, which means pretty good far off axis, not very good on axis. I just don't think the basic theory is correct. If there are problems with the on-axis response, you're going to hear it even if pains have been taken to control directivity, and even if you don't listen on axis. That said, I can see how some people might enjoy the resulting sound, particularly in home theater.

Considering that the OP specifically mentioned the Abbey owner's "amp (vacuum tubes)," your "all wrong" is, well, all wrong.

Regardless, given
-my experience hearing the Nathan in a competent setup (crossover verified, basic AVR powering them, multisubs);
-the result of the GedLee Summa/Gradient Revolution blind test (conducted either by Dr. Geddes or by Duke LeJeune, I don't remember);
-some 1/12-octave smoothed listening position measurements I've seen of a set of Abbeys and Geddes bandpass subs a friend (whose system I unfortunately haven't been able to hear) shared with me (I'll text him to see if I'm allowed to share them); and
-the general incompetent deafness of people who waste time on stupid "tweaks" such as speaker wire swaps,

the inference I made was reasonable, even if it may be wrong.

Overall, what you seem to be saying is that the response slightly off the design axis (the term "on axis" doesn't make much sense here) is important. I agree, but it depends on the nature of those problems. David Smith (JBL, KEF, McIntosh, Snell, PSB, etc.) is fond of noting that the subjective preference research we have is pretty clear that holes slightly off the design axis are not that big a deal. Peaks/swells (such the midrange mushroom cloud that's an inevitable result of the 7" woofer + 1" tweeter on 180 waveguide combination) are audibly deleterious, which is why most "High End" speakers make music sound like the can it came in.

Now, I don't know what response aberrations you measured. Were there notable off-the-design-axis peaks, or mostly holes such as the expected treble cancellation notch?

***We tried two different amps with the Abbeys, one solid state amp (~125 wpc), and one tube amp (~50 wpc). Both were made by companies with good reputations, and seemed to provide plenty of juice for these rather sensitive speakers. So I doubt if it was inadequate amplifiers.***

"Adequate juice" isn't the issue. Any chip amp would be satisfactory to drive them to rock concert levels in a small room.

Rather, the issue I raised is midrange FR aberrations caused by high source impedance. Some people call such errors "tube sound."

Back to my original question: what is behind the unpleasant sounding coloration of speakers with horn or waveguide mounted tweeters?

What's your reference for reproduction again? Don't need to mention a brand name or model. Drive-unit configuration will suffice for the purposes of my question.

 

[randyb]

Just for the record, I took Toole's recent CEDIA courses and his opinion was that there was nothing wrong with horns and that implemented correctly they could compete with other competing more traditional approaches. "Knowing" a horn speaker is at play kind of makes me think "bias" just as it does with "knowing" a ribbon tweeter is playing so one thinks high "distortion". Whether it is audible or not is another question. Again, just an observation and opinion. For the record, I sold my Abbey's to someone who was replacing his Linkwitz Orions and he thought the Abbey's were much better speakers to his ears. Again, anecdotal so probably not worth much.

I will say that I am inclined to agree that a "horn" speaker will have better dynamics than a more traditional approach. The recent love affair with JTR speakers by some very enthusiastic KC adherents attests to that. Jonathan (Achaea) over on AVS is a big believer that there is a lot of bias out there amongst the various brand enthusiasts. I try to avoid that by buying and then selling them all Well, that was true until retirement...now I just have to settle with the end result of my musical chairs approach to speakers (and related components).

 

[DukeL]

When I heard these GedLee Abbeys... It sounded like a clarinet being played in a large public restroom with tile walls and floors... I talked about how good his amp (vacuum tubes) and speaker cables seemed...[emphasis Duke's]

Earl Geddes designed the Abbeys for use with voltage-source (solid state) amps, and when a voltage-source-optimized speaker is driven by a tube amp, the frequency response can be severely degraded.

I don't know what the Abbey's impedance curve looks like, but will take an educated guess: A single peak in the bass region, and another peak associated with the crossover region, about 1.2 kHz. The nominal impedance is 8 ohms, the minimum (probably a broad minimum across the lower midrange region) perhaps 6 ohms, and for the sake of illustration let's assume the peak in the crossover region is 24 ohms... a 4-to-1 spread in the impedance curve is not unusual.

So we have our solid state amp putting out 2.83 volts into this load. That translates to 1 watt into 8 ohms, 1.33 watts into 6 ohms, and only 1/4 watt into that 24 ohm impedance peak. But that's okay, the speaker was designed for that kind of amp.

Now, what happens when we hook it up to a tube amp? A tube amp tends to put out approximately the same wattage into the load regardless of the impedance (within reason). So into 8 ohms, all is well. Into the 6 ohm region, our tube amp is a little anemic, but not much. But into that 24 ohm peak, our tube amp puts out FOUR TIMES AS MUCH WATTAGE AS THE SPEAKER WAS VOICED FOR, so we have a SCREAMING 6 dB PEAK! In the real world it might only end up being a 4 dB peak, but given the region (1.2 kHz ballpark), that peak will stick out like bold with the caps-lock on.

(It is possible that Earl's crossover topology results in the impedance peak being as much as an octave above or below the crossover frequency, but I'm willing to bet it's in there somewhere).

SO... if the most obvious feature of the speaker is a "horn" or "waveguide", and if the peak is in a region that sounds like "horn coloration", then of course the horn or waveguide gets blamed. But, it's not the real culprit.

The real culprit is a mis-match between what kind of amp the speaker was designed for, and what kind of amp was used. This is not at all obvious to most observers, but as a designer of speakers that aspire to qualifying as "tube-friendly", it's something I deal with regularly.

-the result of the GedLee Summa/Gradient Revolution blind test (conducted either by Dr. Geddes or by Duke LeJeune, I don't remember)...

Earl designed and conducted the test, I brought the Gradients and assisted, and was given the distinguished honor of crouching down and shuffling the speakers around behind a curtain because we didn't have Harmon's budget. There were four speakers in the test, and an early version of the Summa tied with the Gradient for first place. The speaker I was "betting on" came in behind those two, followed by the decades-old JBL Model 4430 studio monitor, a truly landmark loudspeaker in my opinion, but that would be another post for another day.