Tutorial: Center Channel Lobing

[Flint]

Multiple driver lobing is caused by the acoustic output from two different transducers being out of phase due to the distance between each driver and the listener being different. Here's how it works:

When a listener is equidistant from two midrange drivers, their output is summed and the listener hears a combined output 6dB louder than a single driver. However, if the listener is a different distance from one driver than the other, that difference will create an out of phase situation for one very specific frequency and all the harmonics of that frequency. The cancellation frequency's wavelength is 1/2 the difference between the two drivers. If the difference is 24 inches, the cancellation frequency is 1128 Hz. If the difference is 12 inches, the cancellation frequency is 2256 Hz, and so on.

When you have a center speaker with two midrange units side by side the cancellations occurs when you are to the left or right of the dead center, on-axis location. The cancellation is called "lobing".

Here is a diagram with made up values showing the nature of the off-axis cancellations:

(In reality, those frequencies will likely be much higher.)

This effect also happens with vertically oriented dual midrange systems, but since it is difficult to get more than a few degrees off axis (by standing up or lying down), the impact is minimal unless the goal is to reduce reflections off the floor and ceiling as in a D'Appolito array.

This is why I do not approve of center channel speakers with dual midrange drivers side by side. In a HT with such a center speaker, the person sitting in the sweet spot gets excellent sound, while everyone else hears a tone with holes in the response that varies based on where they are located in the room.

I should add that once the frequency wavelength is 1/2 the distance between the two drivers, lobing isno longer an issue. at that point the drivers couple their output and the waveform appears as if it is coming from a single point source. So, if the center of the two midrange drivers is 12 inches apart, then starting at 614Hz going down, there will be no lobing.

Lobing (also called Comb Filtering, thanks Conscious) varies between locations. When you are off the center of the axis one speaker will be slightly closer to you than the other. This difference is the root of the lobing. The cancellation frequencies are specific to one location, once you move further off axis, the cancellations change pitch. The last diagram shows the amplitude of a specific frequency. On axis the drivers are complimentary and the output is full. As you move off axis, the amplitude reduces until you reach the 100% cancellation point. As you mve further off axis the amplitude increases until it is fully complimentary. Moving even further off axis the amplitude of that frequency drops, eventually to zero.

Some frequencies will experiences lobes in and out of phase many times over the 180 degree coverage of the room, others only lobe once. The distance between the drivers and the crossover frequency determine the amount of lobing and intensity of the lobing. If the wavelength of the sound is twice as wide as the distance between the drivers, even at 90 degree to the right (or left) the two drivers are still in phase since the wave is so long. At all frequencies below that which is twice the wavelength of the distance between the drivers will be free from lobing.


Now for the gotcha -

This lobing also occurs at the crossover frequency between the tweeter and the midrange. So, all two way, or three way, or greater loudspeaker systems have lobing issues between the midrange and tweeter. This is why 90% of all home speakers have the tweeter directly above or below the midrange drivers. It is also way almost all speakers use a variant of the Linkwitz-Riley tuning for their crossovers.

Here's a link to a long article about Lintwitz-Riley tuning versus Butterworth tuning:
http://www.rane.com/note160.html

So, even turning speaker with a single midrange driver on its side is a bad idea.

 

[Flint]

Dual midrange lobing will occur at all frequencies starting at the crossover point down to the frequency with a wavelength twice as long as the distance between the center of the two midrange drivers. If the midrange drivers are centered 12 inches apart, the lowest frequency for lobing will be around 570Hz. So, if the crossover is 2.5kHz (pretty common) for a dual midrange system, lobing will occur from 2.5kHz down to 560Hz, which is over 4 octaves of issues. With a single midrange the lobing occurs only about 1/2 octave above and below the crossover frequency for a range of about 1 octave if issues. This is a significant difference.

Below is a top view diagram of my HT as it was setup when I wrote this essay:

The red symbols represent the three locations I chose to use to make frequency response measurements of my center speaker. Also in this diagram, I represent my center speaker not in the usual upright position, but laying on its side in a horizontal arrangement.

Here are a few photos of the speaker placements, both vertical and horizontal:

And with the mic in the camera's view:

If you look at the top diagram, you will see that I made three measurements of my center speaker's frequency response with the microphone placed in three positions. I measured the response at each mic position with the speaker in both the horizontal and vertical arrangement. The mic locations represent:

On Axis = Ideal, perfect, sweet spot
Left Seat = The left most seat on the 3 person sofa, where I usually sit when watching TV with my wife.
Far Left = To the left of the sofa - a location I don't believe has ever been used for actual listening.

The speaker being measured is custom made with two Audax 5.25" Carbon Fiber cone mid-woofers and one Dynaudio 3/4" silk tweeter. The woofers are spaced in an ideal D'Appolito MTM arrangement to take advantage of the lobing/cancellations to reduce the floor and ceiling reflections in the midrange (when positioned vertically). For these tests I used a modified Rane AC-22 electronic crossover with a perfect phase correct Linkwitz-Riley 24dB/octave slope centered at 3.4kHz. I had a single Parasound HCA-800 amplifier with one channel driving the woofers and the other channel driving the tweeter.

Here is the on axis (sweet spot) response of the center from 500Hz to 5kHz with the speaker in the horizontal position as compared to the vertical position:

Here is a chart comparing the response of the Vertical versus Horizontal positions from just the left seat:

The purple curve is for the vertically positioned speaker and the blue curve is for the horizontal positioned speaker

Without question the difference is staggering and really doesn't need much explaining. Note the output of the horizontal speaker is over 12dB lower at 1.4kHz.


Here are the response curves of the center speaker in the vertical position from all three mic positions:

The green line is the On Axis curve, the purple line is the left seat, and the yellow line is the far left spot.

Notice these are all in a nice tight window, especially the on axis and left seat curves which remain within 2dB of each other in this entire range, aside from a single fluctuation in the 2kHz to 2.5kHz range. This shows the sound one hears both on and off axis with a vertically aligned speaker will be nearly identical.


Here are the three measurements made with the speaker in the horizontal position:

The orange line is the On Axis curve, the blue line is the left seat curve, and the green line is the far left curve.

It is clear the differences here are huge. The left position is like a different speaker (compared to the on axis curve) from just above 1kHz to nearly 5kHz with differences of over 10dB in places and at least 6dB from 1.2kHz to 3.5kHz. The far left curve is not as bad in the 1kHz to 2.5kHz range, but it is way off starting at 800Hz then mimmics the left position from 3kHz up.

It should be very clear that sitting just a few degrees off axis will make a HUGE difference in the most critical range when it comes to intelligibility and detail. If you calibrated you center with a standard test tone in the sweet spot, it would be at least 6dB off in the left seat.



I chose to narrow the curves to the range of 500Hz to 5kHz because that is the range most affected by the lobing effect with two way systems having a crossover in the 2kHz to 4kHz range. Below 500Hz the room acoustics play a major role in the output response and above 5kHz, the curves are nearly identical since only the tweeter is operating that high.

I also chose that range because it is the range humans are the most sensitive to sound. There's a reason the original telephone speakers and mics were engineered specifically to operate from 200Hz to 7kHz - and the entire analog telephone network was tuned to operate in that range. The midrange is where we recognize sounds. It is the range we need to have most accurately reproduced, especially with center channel speakers which are usually used for nearly all the dialog in a movie.

There are some higher end speakers which won't have this issue, like the top end Paradigm, top end B&W and certain KEF's which use coaxial mid/tweeters. This issue is mostly limited to speakers with two midrange drivers aligned horizontally - or side by side. Center speakers with the tweeter above or below a single midrange, or two midranges aligned vertically, are not going to have this problem. Center speakers with a single midrange place to the side of the tweeter will still have a lobing issue, but only in the range right at, above and below the crossover frequency where the tweeter and midrange are working together.

There are lots of centers out there which do not have this lobing issue (as long as they are aimed down at the listener). They aren''t very common, though. Also, solving the lobing issue is only one way to improve the sound coming from the front of the room. You also need to match the sound of the center to the main left and right speakers, and that is a different thread - one we have created about a hundred times in the past 6 months.

 

[Flint]

Many have written at length about the audible and appreciable benefits to having a matched set of speakers in a 5.1 or 7.1 surround sound system, especially when it comes to the front three speakers. Since our brains are most sensitive to sound coming from in front of us, and since the soundtracks in many TV shows, movies, and concert videos rely heavily on sound being spread across the front speakers, often moving from the left or right speaker to the center then to the opposite left or right speaker, it should be easy to understand why matching the sound from the front three speakers will provide a more engrossing and realistic representation of the sounds played back by our systems. As often mentioned, one way to get there is to just use the same speaker model for all three front speakers, or at least match the midrange drivers, tweeters and crossovers. And, I have gone further to suggest that the ideal would be to use a vertically oriented center speaker instead of a horizontally arranged center.

Aside from the benefits laid out in the posts above on this subject, there is also the acoustics to take into account.

A center speaker should, in a perfect world, not just sound the same as the left & right speakers, it should also present the same ambient sound in the room as the left & right speaker. This can be difficult to achieve when the center is mounted higher or lower than the left & right speaker, and it can be even more complicated when there is a television directly above or below the center speaker. However, the reflections off the side walls, ceiling and floor can easily be made to appear similar to the listener''s ear simply by using a vertically oriented MTM center instead of a horizontally arranged MTM center. And, these reflections will be audible by every listener, not just those sitting off axis as in my thread about lobing.

Here is a diagram of the reflections a traditional horizontal MTM would provide off the side walls as compared to traditional left & right speakers which are vertically arranged:

You''ll notice that while the left & right speaker are causing side reflections off the side walls, the center speaker has no side wall reflections and thus will present a much different ambient sound than the others.


If you flip the center on its edge to make it a vertical MTM, the side reflections look like this:

See how similar these dispersion patterns are?

Since our heads have ears on the side of them, our ability to hear a very wide horizontal soundstage is very, very good while our ability to recognize anything in the vertical plane is rather poor. So, this aspect of the sound can be very important to how we hear a speaker system.


Looking from the side, the horizontal MTM center has an inverse effect in regards to floor and ceiling reflections:

This time the left & right speaker have no reflections off the floor and ceiling will the horizontal MTM has lots of reflections off those surfaces.


Flip the center vertically and the reflections are matched:

So, there are many benefits to using a vertically arranged MTM center speaker, or getting a center speaker that has no issues with off axis lobing.

Of course, these reflection issues are not applicable to rooms where the side walls, ceiling and floor are all treated at the reflection points with midrange absorbing acoustic panels. In fact, I never bothered to talk about this issue because I am such a strong advocate of acoustic treatments at these early reflection points. But in rooms where the side walls are reflective and the room is shaped similar to my diagrams, there are clear acoustical benefits to staying away from the traditional horizontal MTM center speaker.

 

[PaulyT]

Big thanks! I'm very glad you're taking the time to post this and other tutorials. At some point we should probably make an index sticky post for all of these...

BTW, sweet looking MTM's there, my friend!

 

[Alien]

Is the dual midrange center channel lobing issue exclusive to those designs that have each midrange driver producing identical frequencies in phase? Or does it even matter.

 

[Flint]

Lobing occurs with any two drivers producing the same frequency range in parallel (identical signal) where the distance between the drivers is greater than 1/4 wavelength of any frequency they are reproducing.

 

[Orbison]

Thanks Flint. Even though this has been discussed many times, now my pea brain has a better understanding of one of the reasons why my system needs updating. C'mon tax refund!

 

[yromj]

Good job, as always. It's things like this that set this forum apart from the rest, IMHO.

John

 

[Doghart]

Excellent use of RTA and diagrams to simplify your point.

I "get it" now.

D

 

[jomari]

read this article 3 or 4 times in the previous home, prior to 'getting' it.

Thanks for the explanation COF.

I am curious tho, what brought about the discussion of MTM once again?

Id love to see your older postings, hold em here for us to learn more from, and understand.

 

[Haywood]

I completely understand the theory, but I just haven't noticed it as an audible issue when sitting as much as a few feet off from center on my couch, which is about 10.5 feet away. It may very well be there on paper, but I have been unable to hear it moving around on the couch while watching movies.

 

[mzpro5]

I completely understand the theory, but I just haven't noticed it as an audible issue when sitting as much as a few feet off from center on my couch, which is about 10.5 feet away. It may very well be there on paper, but I have been unable to hear it moving around on the couch while watching movies.

I would have to agree in my setup. About the same distance (10.5 ft) as you are.

 

[Flint]

As wih anything having to do with audio, one's perception is all that really matters. I am really sensitive to sound quality and subtle changes between two supposedly identical sounds, but that could be a byproduct of my ear training to be a recording engineer. The principles are solid science with scores of blind tests to support the conclusions that it can make a difference. However, no one should choose to spend money or make drastic changes purely on something they cannot perceive or care about. This is a fun hobby, these tutorials are definitely NOT intended to call anyone out if they don't choose to go along with the conclusions. I am merely attempting to expand the minds of those who love learning about this great hobby.

I can drink any number of red wines, and while I can mostly tell the difference between a Merlot and a Pinot Noir, I cannot tell a $50 bottle of wine apart from a $400 bottle of wine. My friends who are expects think I am a moron for that reason.

 

[Batman]

As wih anything having to do with audio, one's perception is all that really matters. I am really sensitive to sound quality and subtle changes between two supposedly identical sounds, but that could be a byproduct of my ear training to be a recording engineer. The principles are solid science with scores of blind tests to support the conclusions that it can make a difference. However, no one should choose to spend money or make drastic changes purely on something they cannot perceive or care about. This is a fun hobby, these tutorials are definitely NOT intended to call anyone out if they don't choose to go along with the conclusions. I am merely attempting to expand the minds of those who love learning about this great hobby.

I can drink any number of red wines, and while I can mostly tell the difference between a Merlot and a Pinot Noir, I cannot tell a $50 bottle of wine apart from a $400 bottle of wine. My friends who are expects think I am a moron for that reason.

:text-goodpost:

 

[Botch]

I can mostly tell the difference between a Merlot and a Pinot Noir, I cannot tell a $50 bottle of wine apart from a $400 bottle of wine. My friends who are expects think I am a moron for that reason.

If you come down to my range, it's a lot easier to tell the difference between an $8 bottle and a $50 bottle...
Above you've talked about the importance of matched front and surround speakers. I just recently got the SACD Surround copy of Roxy Music's Avalon. Most of it was mixed from the original multitracks, but there's one instrumental where all they had was the stereo master. The engineers just started that mix in the LF and RF, then slowly panned to RF/RR, then to LR/RR, LF/LR, repeat.
It really highlighted just how much my surround speakers sound different from my fronts, even though they were sold as a "matched" set!! (Monitor Audio Silvers)
I have to point out: My surround speakers have smaller woofers than the fronts (don't know if the tweeters are identical, but there's two of them per speaker, not one as in the fronts), their box volume is much smaller, and they're mounted with tweeters at about 5' off the floor, as opposed to about 3' off the floor for my fronts.
I'm learning so much on this forum, and I hope I remember it all when I get my next speaker system, when I'm 80 (I got thirty years out of my Advents, dammit I'm getting thirty years out of these speakers too!)

 

[PaulyT]

If you loan me your Roxy SACD, I'll try it on my 100%-perfectly-Flint-guaranteed-matched system... :angelic-green:

 

[Botch]

If you loan me your Roxy SACD, I'll try it on my 100%-perfectly-Flint-guaranteed-matched system... :angelic-green:

I wouldn't even carry that damn disk thru TSA, in my explosive underwear, coming to visit you. bscene-birdiedoublered:

 

[heeman]

Someday outside of Dreamworld, if I ever have a room dedicated to Home Theater/Audio.

I WILL DEFINIETLY HAVE 5 OF THE SAME SPEAKERS IN THE SYSTEM!!

DO YOU HEAR ME!!!! bscene-buttred:

 

[Haywood]

I am definitely not arguing with your science and I'm sure the effect is there, I'm just admitting that I usually can't hear it or at least don't notice it.

 

[soundhound]

Someday outside of Dreamworld, if I ever have a room dedicated to Home Theater/Audio.

I WILL DEFINIETLY HAVE 5 OF THE SAME SPEAKERS IN THE SYSTEM!!

DO YOU HEAR ME!!!! bscene-buttred:

Not so fast. I worked in the feature film industry in Hollywood for many years, spending much of my time on the dubbing stages where films are mixed. I never encountered an instance where identical speakers were used in the surrounds and the front. In every instance, the surrounds are very much smaller direct radiators; and there are more of them in order to create a "blanket" of diffuse soundfield throughout the audience. The fronts on the other hand were always large horns, way, way larger than the surrounds.

If somebody truly wants "to hear film mixes as the film's director did", then trying to use identically large speakers in both the front and surround positions is futile.

For multi-channel music, then identical speakers might make sense, but even then the surrounds take a more subordinate role than the fronts, so the speakers can reflect that by being smaller. Besides, multi-channel music is really a dying format.