Speaker Technology over the last 10 years

04-13-09: Ojgalli
Alright, enough Mr.T bashing.

I wasn't bashing Mr. T. I was using his outdated indictments of dynamic speaker systems to illustrate the very topic of this thread. More cone-based speakers at more price points have reduced or eliminated their glaring disadvantages in the past 10 years.

In an earlier post, I mentioned several developments of the past few years that have catapaulted speaker performance over what was affordably available 10 years ago. I mentioned the API/Mirage developments of the Omniguide and their elliptical rib surround, which totally changes the rules on diaphragm diameter vs. cabinet size, bass extension, and clarity. I mentioned increasing use of neodymium magnets and beryllium tweeters. To that I could add that offshoring manufacturing to China has made intricately built enclosures affordable. Basically China has freed us from the box speaker. More and more speakers at affordable prices have elliptical shapes, curved sides, curved tops, and the attendant decrease in cabinet resonances and standing waves. More are getting away from MDF, using polymers, extruded aluminum columns, multi-ply birch from Europe (the stock used for piano pinblocks), and staved construction a la Sonus Faber and Usher.

Drivers haven't stood still. Witness the emergence of the ring radiator tweeter, which has found a home in many upscale loudspeakers from Sonus Faber, Magico, and AV123, and the ring ribbon as used by Adagio and Genesis among others.

These advances have made their way into entry-level products. Take a look at the averaged anechoic response curve of the $449/pair PSB Image B25 stand-mounted speaker. It is +/- 3dB from 50 to 20KHz, and is about +0, -2dB from 300 to 8 KHz where most of the action is. This level of linearity at this price point IS revolutionary.

There's been a resurgence in open-baffle designs.

Earl Geddes has identified and fixed a major source of horn coloration with his open-cell foam filled oblate spheroidal wave guides.

Tom Danley invented the unity summation aperture (a multi-way point source horn) and tapped horn for bass. He's produced compact bass horns that get a lot of the horn size from the environment.

DSP and digital amplification makes it easier to realize multi-amplification in high output active designs.

There are newer drivers which are linear over higher excursion (for more output, or a MT array with the same output as an MTM but better polar response).

There are newer drivers with less distortion and stored energy, but this is a minor change when you have good polar response and a correctly executed cross-over.

There are more small makers which essentially charge you for parts plus a furniture maker's time. You can get much more attractive veneers and solid hardwood pannels without spending five figures.

OTOH, if you buy a generic speaker with generic drivers it's unlikely to be appreciably better than an older speaker. It'll just cost more due to inflation.

Mrtennis writes:

>i think you are ignoring the obvious. i exchew cone speaker >designs. i realize others like them. i just have a hard time >fathoming why they are so popular, given their faults. i will >never buy one .

You want to listen to a dynamic dipole like the Linkwitz Orion which combines a panel speaker's transparency and natural sound with relative placement insensitivity, a big sweet spot, and second octave bass at realistic output levels.

why is it that whenever i listen to a cone design i hear cabinet and driver colorations. please cite a cone system with no cabinet colorations and i will listen carefully to it when i have the chance.

you can blindfold me and i will recognize a cone speaker. if cone speakers are so realistic, why do i detect a sonic signature ?

i realize it is a matter of personal taste, but i wonder how many who prefer cone designs would not prefer them as much if they heard a panel speaker in blindfold mold.

timbre is most important to me and i feel panels do a better job.

if there are any outstanding cone systems, i will try to appreciate their virtues, the next time i go to las vegas.

>why is it that whenever i listen to a cone design i hear cabinet and driver colorations.

Because you listen to multi-way cone speakers in conventional cabinets which have big polar/power response problems, probably have cabinet resonances, and may have internal reflection problems.

There are not a lot of brands sold out of show rooms which get around that, their dealers are not numerous, and some of the designs are compromised to suit market tastes.

AccentSpeaker's Nola dipoles, Lyngdorf's dipole, the Gradient Revolution cardioid/dipole, and Jamo's R909 dpole use cone drivers in open baffle configurations.

B&W's flagship Nautilus has acoustically small drivers close together in damped transmission lines.

You can demo Linkwitz Orion and Gilmore's dipoles, although that's more likely to require a visit to an enthusiast's home, show, or plane trip somewhere.

The Carver Amazing and Audio Artistry line used conventional cones in dipoles but aren't around any more.

>please cite a cone system with no cabinet colorations and i will listen carefully to it when i have the chance.

Linkwitz Orion - no box, no box sound.

Linkwitz Pluto, where the mid-bass enclosure is

1. A damped transmission line with a 40dB return loss

2. Exceptionally rigid (there's an AES paper out there showing that cylindrical enclosures are as rigid as 4" concrete).

On top of the enclosure (or lack thereof) you've got the room interaction.

Technically speaking, the near field ends at the critical distance where direct sound and reverberant field are equal. With conventional speakers in a domestic environment that point is at 2-4 feet from the speaker. By the time you get a comfortable 8' out the sound you're hearing from your speakers is 6-12dB below the direct sound.

Many of your conventional speakers have a mid-range that's getting acoustically large with narrowing dispersion and the tweeters are uniformly acoustically small so they have wide dispersion. The two are relatively far apart compared to wave lengths about the cross-over region. Combine the two and you have a power response notch there. Conventional MTMs have issues for the same reason. As a tangent I think people prefer first order designs due to the shape (shallower + broader versus deeper and narrower for higher order) of the power response notch since the all-pass response introduced by fourth order filters isn't audible by itself.

When 3/4 to 15/16 of the energy you hear has a big bite taken out of its spectrum the result sounds like a speaker.

Single drivers like a big planar work around the problem with decreasing power output at high frequencies without the cross-over notch although the resulting lobes make the reflections unpredictable and placement finicky.

Dipoles like open baffle cone speakers have at least 4.8dB of directivity so things can't broaden too much at higher frequencies. The angles where cancellation is creating the biggest notch in output about the cross-over region are in a dipole null so the impact on total power response is less severe.

Speakers with acoustically small drivers close together like the Pluto avoid the problem because there isn't a large fraction of wavelength between the drivers at any angle. 4" between mid-bass and tweeter with a 1KHz cross-over is about 1/3 wave length so there shouldn't be cancellation anywhere (I haven't looked up the tweeter time alignment all-pass delay and considered it).

Coaxials can work well for the same reason. I heard a pair of the Seas coaxials in open baffles which didn't sound like speakers.

John Krevosky's NaO for Não em uma caixa "Not in a box" should work like the Orion. The Nomad Audio Ronin which is an open-baffle coaxial at high frequencies might work well too although I think you loose a lot not retaining dipolar radiation down to 40-50Hz.

I'd speculate that the GedLee and Audio Kinesis speakers mating a waveguide to a large mid-woofer matching its radiation pattern at the cross-over point would sound natural for the same reasons cone based dipoles do - more uniform power response because the driver directivity matches and the angles of greatest cancellation are in an area where response is attenuated.