Why not horns?

Weseixas wrote: "I'm sure you are not comparing your bi-polar speaker polar plot to that of a Di-pole with it's figure 8 pattern ?.... I highly doubt the 2 will sound the same, the sonic characteristics are so completely different..."

Duke replies: Indeed I am making precisely that comparison.

My bipolars radiate into a constant 90 degree pattern, both front and back, over most of the spectrum. Few dipoles maintain their initial figure-8 once we get up into the midrange and treble region; their radiation patterns usually change significantly with frequency (SoundLabs being a notable exception).

As we go down in frequency my pattern widens somewhat, but this is of less consequence down where the room dominates anyway. In the bass region I even manage to emulate much of the in-room smoothness of a good dipole system, and here's a recent article in a webzine that describes how I do so.

A few years ago I had a longtime (decades long) SoundLab owner tell me that one of my bipolars came the closest to sounding like his speakers of anything he had ever heard, including monopoles, other dipoles, other bipoles, omnis, or anything eles. In fact he actually placed an order for a pair of my speakers, but later cancelled it after changing the setup of his SoundLabs to be more like how I set up my bipolars - apparently that made a significant improvement in his room. Since my intention was to emulate the radiation pattern characteristics of the SoundLabs (something I did with Roger West's full knowledge), the outcome there was still encouraging to me. I don't mind coming in a close second to SoundLab.

Since you didn't reply to my inquiry about where you heard my speakers, now I'm thinking that maybe you didn't. Is that correct?

Duke

06-20-10: Audiokinesis wrote:

My bipolars radiate into a constant 90 degree pattern, both front and back, over most of the spectrum. Few dipoles maintain their initial figure-8 once we get up into the midrange and treble region; their radiation patterns usually change significantly with frequency (SoundLabs being a notable exception).
-----------------

Sorry , not sure how you are coming by this, everyone i have ever measured exhibits a classic figure 8 or pretty close 2 it, unless there is an reflection issue.

*How were you measuring ? you must have reflections issues.
*Also is your bi-polars reverse phase in the rear ?
* is yours a compression horn or a wave guide ?

I'm not sure about your question Duke, i must have missed it before , anyway i have never insinuated or said that i had listened to your speaker(s) before.

Which of your speaker models are you suggesting for me to take a look at Duke ?

Which one sounds like a dipole ?

.............................. bookmark

Weseixas, the reason few dipoles maintain their figure-8 up into the mid and treble region has to do with beaming and/or driver interference. For example, the horizontal pattern of Maggies changes due to horizontal comb filtering in the crossover region, and then they have a very wide pattern just above the crossover point, narrowing with increasing frequency (beaming) just like any direct radiator. It's still a dipolar pattern, but not the same neat figure-8 we had at low frequencies. Still, overall Maggies are better than most when it comes to radiation pattern uniformity.

Unfortunately electrostats tend to have patterns that narrow significantly at high frequencies because they are relatively wide in the horizontal plane; even the concentric-ring Quad 63 and its descendents beam severely in the top two octaves.

In contrast, my bipolars have a much more uniform pattern over most of the spectrum. From about 500 Hz on up, the faceted-curved SoundLabs are the only bi-directional speakers I can think of that have a more uniform radiation pattern in the horizontal plane than mine (and below 500 Hz is considerably less critical in most rooms, according to Dr. Earl Geddes). I happen to think radiation pattern uniformity matters a lot, because it largely determines what's happening in the reverberant field, and typically most of what we hear is reverberant energy. No doubt other designers place higher priority on other characteristics, which means that the industry offers a pretty good variety of choices for those who go digging for them.

Responding to your questions:

* The measurements I take which reveal radation patterns are always time-gated to exclude reflections.

* The rear-facing drivers on my bipoles are not in reverse polarity; if so, they would be dipoles.

* My first two bipolars use round waveguides based on Earl Geddes' work, but my most recent (and least expensive) one uses what I'd call a waveguide-style horn; that is, constant directivity with no diffraction slots, vanes, or other pattern-widening features.

I wasn't necessarily suggesting you take a look at any of my bipolars, beyond the links I've provided. Any of my three bipolar models would have room-interaction characteristics similar (but not identical) to a dipole. The article linked to in my last post offers a good introduction to the controlled-pattern offset bipole configuration I use.

One other thing you said that I would like to comment on:

"Look I can understand those that have migrated to the sound of horns, they do it for the power."

That wasn't my motivation. As a hard-core amateur in the late 80's (with two SpeakerBuilder Magazine articles to my credit), I came to believe that getting the reverberant field right was critical, but didn't know how to do it well. And I hated horns, so didn't consider them as a possible solution. Well in 2001 I heard a hornspeaker that didn't have the coloration I'd come to take for granted, and suddenly saw a window of opportunity to get the radiation pattern right. I won't bore you with the details of the journey, but it was the possibility of good radiation pattern control that attracted me to horns (or waveguides, to be specific... I'm still leery of most types of horns).

Duke