Critic's Corner

Posted by morricab (R) on February 01, 2008 at 02:16:33

In Reply to: RE: Sean Olive posted by KlausR. on January 31, 2008 at 08:08:27

Why is Constant Directivity (CD) important? What effect does it have on listening quality?
Both questions are equally important. Loudspeakers without horns radiate in a narrowing pattern as they transverse higher frequency ranges—this is a function of their effective radiating width. In most multiways, and planars in particular, each driver is much larger in diameter than the wavelength of the top frequency it must reproduce. As drivers approach this cutoff, response suffers from an undesirable phenomenon called “roughness” and their directivity narrows to the point that, once their width/diameter is larger than the wavelength of the upper frequencies, they more or less radiate in a straight forward beam, like a headlight. This reduces the “listening window” to on-axis and makes it difficult for listeners sitting away from the “sweet spot” to hear all of the music.

The problem worsens in the crossover regions of a multiway where a large diameter driver and the smaller one above it differ in phase and dispersion causing frequency response protrusions and suckouts collectively known as “lobing”. While this effect can be reduced with increasingly higher-order crossover filter slopes, there are practical limits to implementing passive high-order networks due to poor transient response, group delay, and reduced dynamics caused by losses through the filter elements. Speakers like ours with first order filters (chosen for their optimum listening quality, not for smooth dispersion) suffer from the “lobing” phenomenon more than their steep-sloped cousins.

What is needed for CD is for all drivers in a multiway to be the same width, and that width needs to be smaller than the wavelength of the frequency you want to hear with good or constant directivity. If you want to hear 20kHz well off axis, that means a speaker no more than 2/3” wide, down to as low in frequency as possible, where it can mate with larger diameter woofers which are already working into 2 pi (180 degree) space and are about to transition to omnidirectional or 4 pi space. In the past speakers have been made with long and narrow drivers. These tend to be tweeters, however, and virtually all exceed 1” in width, which means they will beam at around 13 kHz or lower.

All our planar speakers use a 2.5” wide midrange panel working from about 280Hz up to 7 kHz, handing off to a 7/16” wide tweeter operating to above 20 kHz. While their horizontal dispersion is very good, there is lobing in the crossover region and some wave interference in the vertical plane (vertical dispersion of long ribbons is poor, usually not much beyond the length of the driver itself). If we want Constant Directivity type dispersion then whole speaker would have to be made drastically skinnier, impossible to do without sacrificing LF extension and sensitivity. For this reason older designs have not attempted to achieve CD coverage without horn loading.

In a REAL room, Klaus your on-axis FR means very little.