Lobing Accoustic Interference ~ How The Dual DRC Onken Center Channel Performs 01Jun24

Directivity... in audio describes the way a speaker’s frequency response changes at off axis (off straight ahead) angles. A wide directivity speaker is one from which sound amplitude (sound pressure level, SPL, often expressed in dB) is maintained on and off axis and a narrow directivity speaker is one where off axis SPL is different from on axis, often lower. As frequency increases, a driver's directity is very likely to increase/narrow. These effects are often rendered in polar plots like below, where dB are plotted versus frequency or wavelengh (1st example is pretty wide directivity; 2nd shows narrowing directivity). Driver manufacturers also often provide off-axis performance in a normal superimposed SPL sweeps (0°, 15°, 30°, 60°, etc) graph.

Fostex FE108EΣ Directivity... It's better if you can measure the off-axis performance yourself, or access data from a lab who has done that... as from HiFi Compass below for the Fostex FE108EΣ. You can create a polar plot from that data ~ for me it is a more visual representation of the directivity... these graphs are with on-axis performance normalized to 0dB; also these are 2D while the sound really is emitted 3D from a speaker. For a full representation one really should use a 3D sphere "graph;" the polar plot is accepted for my use here and round drivers typically work nearly equivalently above and below as they do left and right (though of course there are baffle effects and wall reflections in reality).  While I completely enjoy the Fostex FE108EΣs, one can clearly see the directivity of my drivers increasing/narrowing as frequency increases. However also, notice the widening from 8KHz to 10KHz; weird. The directivity behavior is also called beaming.

Beaming Affects Audio Performance...  Not just in speakers but in microphones as well. Polar plots of responsiveness/sensitivity are commonly included with many mics. In speaker-created soundfields, often, listening to music off-axis is not as good as doing so on-axis ~ this is kindof why every image of a key listener in home stereo is shown directly in the center of the soundfield, often on-axis. With this general attribute in most speakers, one knows why if you sit dead center on-axis, you often get full signal amplitude across the audible spectrum and it just sounds better.

Theater/club horns are less directed... the audience is known to be widely spread out and the sound produced better be darn good wide across the stage/theater/club. So, why not just use horns every time? Oh baby, were we able... the best horns need to be very large and cleverly placed in a room; they also cover only a couple octaves each as their throats & mouths are finely tuned. A gigantic 4 way horn system is spectacular. Anyway... suffice it to say you can't easily get there in "your" room... argh, digression.

If a speaker is big it will start to beam sooner as it goes up in frequency. The frequency at which a speaker will start to beam can be determined by f = (2 * c) / (π * d) where c = speed of sound (343 m/s) and d = effective diameter of speaker (m). For my Fostez FE108EΣs that's f = (2*343)/(3.14*.1) = 2184Hz. That's born out in the graphs above where 1KHz and 2KHz are wide but 3KHz substantially narrows. The dispersion of tweeters is important as they render the highest audible frequencies where directivity is most noticeable ~ this is why many tweeters are domes, which disperse sound better.

Comb Filtering & Acoustic Interference... are audio terms to describe how multiple sound sources interact with each other, and combine, when playing the same signal. Comb filtering occurs when a delayed signal combines with another equivalent signal. These can be two drivers emitting the same signal (separated in space) or a single driver's signal interacting with a reflection (from ceiling, wall, etc). Sounds are oscillations and can be thought of as having + and - phases and are added linearly in time in space (and in math and in our ears); the result of combining can be reinforcing, + with +, or cancelling, + with -. Combing causes lobing (C, constructive or D, destructive) and is measurable acoustic interference, the degree to which changes in space and distance from the driver and the distance off-axis.

The amount of combing and acoustic interference in a soundfield increases with the frequency of source material, which often increases directivity of the driver, which is the source or more frequent interference off axis. There's plenty of debate whether this effect is audible in many situations, but we know it exists and is measurable and many speaker and sound designers must deal with it, especially in multi-driver systems, especially at crossover points, and especially when listeners will be known to be off-axis.

Light Lobbing Math... For the multi-driver case, if the distance between centers of the drivers is less than than 1/4 wavelength then lobing is not an issue. Plus, I sit dead center, so not an issue for me, ever. Yeah, I have my guests sit abreast me in the NestMT so it might affect them. If speakers are allowed to operate over a frequency range above where the physical distance between the sources is large compared to the wavelengths of sound, then acoustical interference may become a problem. That said, realize in home audio it's not like a pebble in a still lake, it's like a fistful in a toilet bowl ~ there's tons of reflections and wave interaction. I'll try to isolate for the specific direct lobing effects but it is a challenge without an anechoic chamber for a room.

NestMT Center Channel... I use two DRC Onkens for my center channel and drive them with the same signal. The drivers are 20" apart center to center. Why is this design performant when the driver has been shown to beam and the same signal is applied to both while delivering frequencies above 2.2KHz? First off, there is lobing (as in nearly all audio systems) but I cannot really discern it. We sit at a critical distance from sources and much of the sound is nearfield but a good mix of reflections are in there too; we have two ears and integrate sound across that space; from psychoacoustics, in humans loud supersedes quiet and there are 9 other drivers delivering content simultaneously in my music theater; and I am doing some DSP to widen the sweet spot from the mains. However, I have now placed the center channel drivers much closer together and mathematically they have a much larger sweet spot (green “rays” represent “destructive” lobing).

Empirical Measurements...  I made multiple SPL sweeps of just the DRC Onkens CC driven alike; I did so at all five listening positions and at three dual CC center to center distances of 20", 11" and 9" separation. The target center is listening and is excellent always, but empirically the flanking seats are great too. At 20" I do measure lobing at 4.6KHz and 8KHz, but of just 3dB at the flanking seats (~3.5' abreast); at another 3.5' abreast of even that I see it again at 8KHz especially but the rest of the curve is largely coincident, but just down a skosh overall. Fine for dancers.

At 11" c-to-c separation there was improvement in lobing-based acoustical interference observed from on axis to flank seating as the curves are more proximate ~ but not much. The curves were closer together but the lobing was still present at 4.5KHz, 7.2KHz, and even 10.2KHz, again by just several dB. Not surprising as in the NestMT sweet is 8'2" and flank is 8'7" so not a great difference ~ but axis relative to driver 0° is important. The dual DRC Onkens CC is lobing but powerful and full to sweet spot and not high interference for flanks or even wide of flank. Yet, be invited to center if you can.

Even @ 9" driver separation... lobing interference was noticed at 4.4KHz and 9KHz, again of about 3-5dB. This really is not much. And, wider yet is similar. This is insufficient for me to re-orient the center channels from vertical dispersion to horizontal... as my ears are ~2ft above the drivers and I like the vertical dispersion aspect.

Most importantly is whether the sweet spot is affected much by these driver separations. A smidge... the flipped DRC Onkens with closer center-to-center drivers measure and sound better. The 9" vertical arrangement may be modestly better, but is more problematic in my music theater. I will flip the speakers.

In the midst of our returning to the Nest for these measurements, K crushed a super dinner of Asian chix thigh and asparagus... so good with sriracha cashews, scallions, and cilantro mated with marinate and sauce. I've been cookin' a bit recently with K failing to retire and earning $ for us, she's too good and back on the task.

This entire shit storm, and improvement, is sourced in a question from my young & clever Kiwi friend, Curtis Gray. Rock on Gray.