Why horn-loaded loudspeakers are subject to design tradeoffs

[Deang]

http://peavey.com/products/index.cfm/item/727/106452/CH%26trade%3B-942QT

http://www.cassellsmusic.com/products/Peavey-CH-942QT-OTC-HORN-131688.html

They need to make a bigger one.

[djk]

"They need to make a bigger one."

http://www.frazierspeakers.com/download/cat59.pdf

Edited July 26, 2014 by djk

[DrWho]

I have heard the same subjective statements about almost every type of speaker.

The problem with plasma speakers is they work by modulating the temperature of the air...which is a very nonlinear process (it heats faster than it cools). You can limit the heating slew rate to the cooling rate, but then it is very quiet. I wanna say around 70dB at 1m or so? We were going to try horn loading, but never got around to it.....the musical Tesla Coils with 15ft sparks were way more entertaining.

[Chris A]

Mark,

By way of example of your comments, if you consider only harmonic distortion, most of the audible distortion (of all types) tends to be present in the bass bin than in midrange and tweeter drivers/horns. Below you'll find a recent measurement of a Belle with its FR EQed mostly flat. This Belle is using a 24 dB/octave L-R filter centered on 600 Hz for the bass bin-midrange transition. It uses the same type of crossover filter at 3500 Hz between the tweeter and midrange horns/drivers. (This Belle is time-aligned.)

It would be nice to also be able to easily plot relative modulation distortion levels (which the ear is much more sensitive to because it's nonharmonic) using REW. Note also that harmonic distortion turns into IMD:

Edited July 27, 2014 by Chris A

[ClaudeJ1]

I might also add that you're putting words into my mouth in regards to what I consider loud. I am not an SPL junky by any stretch of imagination, and take hearing loss very seriously (both for my own listening and when doing live sound). I also take my dynamics and crest factor seriously too, and try to keep them uncompressed wherever possible...crest factor is where the difference is.

"The purpose of an Auditorium is to give people a place to hear."..............PWK

Yes, I get that in terms of reverberation time and spectral balance, it was a "good" room, but you can't ignore being 100 ft. away vs. being 10 ft. away in a home environment. What I suggested is that in a home the electrical power INPUT would be AT LEAST 10 db less, which would lower the IM distortion more significantly with any speaker, whether it's a HORN or not. Don admits that the tweeters in the CBT are too inefficient and is has found one that is simply marvelous, so he can return to a 1 Khz. Xover point vs. 2 Khz. from the one we heard. Also going to a slighly larger woofer will increas the Sd to a 15" woofer equivalence vs. a 12" in it's present configuration. Both of these will reduce the IM distortion significantly, increase clarity, etc. This is the one I will build or buy in kit form. Hopefully by Xmas 2014.

As to the presumption of high decibels in your case, I am going by the video showing how you tweaked your horns in the gym and your posts on the most powerful DIY sub out there, the OTHORN (in "party mode"). Forgive me for taking your ACTIONS into account.

BTW, I now own a pair of Synergy Horns which I put in my photo studio. They are small, very heavy for their size and present an extremely clear and phase coherent image in a highly reverberant environment. I think Tom Danley is not only the undisputed "King of Bass" (see the Matterhorn) when he abandoned the Servo Drive idea for Tapped Horns, but he's also the King of the other 9 Octaves with his Jericho Horns and SH-50 and it's derivatives. The boys in Georgia are busy kicking bootie and takin' more names than the boys in Arkansas or California.

Edited July 28, 2014 by ClaudeJ1

[Schu]

JBL's new Wave guide...

seems they are getting more and more complicated to help with both LF and HF dispersion in the same horn

[ClaudeJ1]

"They need to make a bigger one."

EV did this too with concentric horns. You can make a CBT out of anything, but horns will have narrower dispersion than small DR's. The key is the Legendre shading that does the magic. Even the McIntosh needed this.

[ClaudeJ1]

Forum member JWC had an expensive pair of McIntosh towers with the CBT pricinciple. Like 72 tweeters or something like that. Honestly, they were awesome and did not need subs (by the way). But they weren't some cheap kit with budget drivers and no crossover...........and JC had 1200 watt Macs driving them. Different beast.

The CBTs are just a home audio budget version of what's been around a long time in Pro sound reinforcement. Go to any concert today or back in the 80s and 90s and it's been hanging right in front of you.

Anyhow.............if you read the thread on how to assemble them there is not really anything special about them at all. It is groups of tiny budget drivers positioned along a curved baffle playing at "shaded" volumes relative to each other.

I just don't find anything interesting to me at all about it. Why dioes it amaze everyone if you can walk around the room and get even coverage? Klipsch does that with their speakers too.

At $60,000 a pair, JWC's McIntoshes were a different beast alright. AND he sold them off in favor of the MCM top end, similar to yours and mine, with the Tractrix midbass of his own design. At a measly $2,000 for the CBT-36 kit, you need to compare it to other systems of the same price range, which, would be mostly direct radiators, not horns. They are all "different beasts" as you aptly put it. So the only fair comparison would be to Direct Radiator systems in the same price range, so I firmly believe they would win over the listening audience in that scenario.

I disagree about "not anything special" about the CBT-36. The radius, angle, economical passive Legendre shading, and amazing CD performance, and uncanny refudiation of the inverse square law makes them quite unique.

If you want to talk "nothing special" then that would apply stacking horns of different bandwidths into a corner, with compression drivers. What makes our stacks special is the optimization by DSP's, which can apply to any speaker.

Besides, it's all magnets and moving coils moving air anyhow, not too special after 100 years.

Edited July 27, 2014 by ClaudeJ1

[ClaudeJ1]

To woof or to tweet? It seems that most of the emphasis on using horn loaded drivers is with tweeters. Horn loaded woofers are rarer due to the size. But, relative to musical importance, that's a backwards. With nearly the whole of the human voice range falling to the woofers, that's the place to start. And, what has to be right is the tone. Have woofer horns advanced the way treble horns have? Such as constant directivity? Don't we still fight resonances and unequal distribution because of how difficult it is to implement large bass horns?

THIS is the main reason I designed a smaller Klipsch MWM-derived woofer horn that performs well, to my taste anyhow.

Edited July 29, 2014 by ClaudeJ1

[ClaudeJ1]

So, thinking back to the original question about tradeoff of horn loudspeakers, I'd say the #1 tradeoff is just size. And this becomes dramatically true in a full range horn system.

I agree with this statement, size is the only downfall.

Not if you step out of the Klipsch/JBL world and get into Tom Danley's Synergy horns. The best of them get BASS down to 50 Hz., (as low as my Quarter Pie horn, which is much larger) in a very small size conical horn. Plus they are totally phase coherent (they can pass a square wave over a full decade from 259-2500 Hz.)and Constant Directivity, but boy are they heavy.

Yes, they need a sub, but all speakers do IMHO, unless they build one in.

Edited July 30, 2014 by ClaudeJ1

[Chris A]

Mark,

I agree that the curves are stacked together too closely - but I'll blame it entirely on REW, which selects the default colors for the curves by itself. Within REW, each curve will stand out on the graph using point-to-focus on each of the harmonics within the legend so that each is easier to see.

And you are correct: the point that I was trying to emphasize (that you made above) was that there is a rich amount of distortion in the bass bin, almost all of which is audible since even higher order harmonics (and modulation distortion side bands) are all within the human hearing range. Harmonic distortion in bass bins has been shown to be almost the same in horn-loaded, ported, and I assume also for acoustic suspension bass bins.

It's interesting that Modulation distortion isn't is offered as a measurement by the program REW (which now offers a variety of tone generators, including linear sweep sine, log sine, square wave, dual tone [iMD], CEA-2010 Burst, Pink Noise, "Pink PN", & "White PN" in the beta 20 version of REW V5) which is where horn-loaded designs outperform direct radiating designs by the largest margin (and my ears tell me so, too). Too bad. I think that the world of hi-fi loudspeaker makers and buyers typically measure/pay too much attention to harmonic distortion, and really pay no attention whatsoever to modulation distortion, which is much more objectionable. I believe PWK spent a lot of time on bass bin design because he realized and also heard the distortion most clearly in the low frequency drivers/horns, too.

Edited August 4, 2014 by Chris A

[kg4guy]

So, thinking back to the original question about tradeoff of horn loudspeakers, I'd say the #1 tradeoff is just size. And this becomes dramatically true in a full range horn system.

I agree with this statement, size is the only downfall.

Not if you step out of the Klipsch/JBL world and get into Tom Danley's Synergy horns. The best of them get down to 60 Hz. in a very small size horn. Plus they are totally phase coherent (they can pass a square wave over a full decade from 259-2500 Hz.)and Constant Directivity, but boy are they heavy.

Yes, they need a sub, but all speakers do IMHO, unless they build one in.

YES I must agree

[DrWho]

It's interesting that modulation distortion isn't even offered as a measurement by the program REW, which is where horn-loaded designs outperform direct radiating designs by the largest margin (and my ears tell me so, too). Too bad. I think that the world of hi-fi loudspeaker makers and buyers typically measure/pay too much attention to harmonic distortion, and really pay no attention whatsoever to modulation distortion, which is much more objectionable. I believe PWK spent a lot of time on bass bin design because he realized and also heard the distortion most clearly in the low frequency drivers/horns, too.

In order to create modulation distortion, you need a two tone source. I have created several different two tone tests, but you will quickly find that it is almost impossible to come up with an apples to apples comparison between different designs. This is because the specific frequencies you choose can have a huge impact on one speaker, and very little on another due entirely to its design....

You are then left trying to sweep every possible frequency combination and then there is too much data to be compared readily.

The short story is you can't get a nice number like "%THD" to make quick comparisons, and there isn't enough data to create an appropriate audible weighting factor to an array of tests.

And all that is ignoring the time it would take to generate the data in the first place.

I personally find it easier to understand the excursion requirements of the various drive units, and then use that in light of the bandwidth of the system to get a rough understanding of what level of impact that the Doppler Effect might have on the system. Once you rule out the Doppler Effect, the THD and IMD performance will track very nicely. I think I read that in an AES paper somewhere, but I don't remember the title.

All that to say, if you have some ideas for better IMD testing I'm all ears and probably have the equipment to explore those ideas.

[Chris A]

If you take a look at the Klippel procedures found toward the bottom of this page under "application notes", there apparently is a "IEC 60268-5 Sound System Equipment, Part 5: Loudspeakers" standard that covers these tests. The references that I'm referring to are:

AN 8 3D Intermodulation Distortion Measurement ,

AN 10 AM and FM Distortion in Speakers, and

AN 16 Multi-tone Distortion Measurement

The problem is the buy-in cost of the Klippel R&D or QA test setups. But I have reason to believe that the test procedures may be standardized now.

I got the reference to the new IEC standard from Floyd Toole. He's a great source for this kind of stuff.

Chris

Edited August 18, 2014 by Chris A

[ClaudeJ1]

I got the reference to the new IEC standard from Floyd Toole. He's a great source for this kind of stuff. Chris

Is that in his book, or did you find a download from somewhere?

[Chris A]

LinkedIn - look at my profile there for his name (...you and I are already connected there).

Toole writes into a group there called "Electro Acoustics" from time to time,, and that was one of the discussion items.

[DrWho]

So if a system doesn't even have a 4:1 bandwidth (like 20 - 80 Hz), then these tests don't apply....

Also, even with the 3D plot, you're not getting the full story. These tests are very much driven by finding fixable driver design or manufacturing issues.....not so much a holistic representation of the audible impact of a design. You really need to test every combination of two tones, which includes relative frequency and relative amplitude across the entire spectrum.

Btw, do you agree with the Klippel chart showing that IMD is a low concern for narrow bandwidth systems? That chart is also showing that amplitude distortions are much more important....

[Chris A]

My feelings on IMD are perhaps not one that I care to share in full here, but yes, I believe that the most bothersome effects are AMD effects, as we have discussed before.

It seems that all electromagnetically driven acoustic drivers (of solid and not gaseous state) have AMD effects at some level of acoustic output. I believe that it's much more prevalent in all loudspeakers than what most want to acknowledge--much more so for direct radiating types. That in itself is saying something about a potential shift in the loudspeaker technology market. Klipsch would do well to press on this subject and advertise their various loudspeaker models' IMD performance against "brand x" competitor loudspeakers.

But since we presently never get to see the output of the Klippel/lEC modulation distortion tests when looking at manufacturer's driver performance data, this fact alone is a good reason to standardize the tests--even "good enough" tests--since some information on this important subject is better than none at all. My feelings are that the IEC and Klippel labs probably feel the same way about the subject, so they proposed standard tests and got them approved. I applaud their efforts for they have brought forward a suppressed subject that PWK battled for a great deal of his adult life, almost alone.

YMMV.

Edited August 18, 2014 by Chris A

[DrWho]

Just to be clear.....this is the chart I'm referring to (was posting from my phone before):

It indicates Doppler DIstortion is not a significant factor for narrow bandwidth systems....you seem to be constantly trying to make a different point.

As far as amplitude related distortions - these always show up in normal THD measurements. Doppler is the only thing that requires a multi-tone signal to quantify. Sure, our ears may be more sensitive to the IMD products generated from the AMD, but a system with lower THD will also have lower IMD. In other words, THD is almost always a sufficient metric to compare the linearity of two different systems. I think your desire to see IMD measurements is a bit naive/idealistic, and ignores what can be interpreted from a normal THD Plot.

There are direct radiators out there with some insane linearity over a very wide range of frequencies and excursions. And in some cases, the THD is actually lower than a comparably priced / sized horn system.

[Chris A]

AMD isn't "Doppler distortion" unfortunately, but rather Doppler distortion is FM distortion. Which type of distortion are we talking about?

Why do you want a "narrow bandwidth system"? I don't (except for perhaps the subwoofers--which are always a compromise design due to space limitations, IMHO).

Edited July 31, 2014 by Chris A