Vertical vs Horizontal Center Channel Speaker Designs – An Alternate Perspective
About two years ago Chris Seymour of Seymour A/V was kind enough to write us a very informative article (Vertical vs Horizontal Center Channel Speaker Designs) comparing frequency response variations of center channel speakers mounted vertically vs horizontally. This article dealt primarily with the assumption that the listener would be significantly off-axis from the speaker and hence the horizontally mounted MTM speaker stood (no pun intended) at a disadvantage.
Since this articles inception, we’ve noticed it being linked in various forum discussions where people have been discouraging the use of traditional MTM speakers mounted horizontally. Some have even gone so far as to deem common MTM speakers to be severely flawed when placed horizontally under or above a display. While others opted to use a two-way bookshelf speaker in-lieu of an MTM design for these very reasons. Many manufacturers have addressed this issue by designing dedicated WTMW designs that utilize two woofers for bass frequencies, and smaller driver for the midrange vertically mounted below the tweeter. There is little doubt that these designs offer wider off-axis performance than a comparable horizontally mounted MTM. But how far off-axis does one have to get until the MTM design becomes too problematic? There are no free lunches with any loudspeaker designs. Thus compromises exist with both MTM and W(T/M)W designs which are discussed in this article and should be considered when choosing the right center channel for your home theater.
As a refresher let’s review some examples of different center channel designs.
MTM (Mid – Tweeter – Mid)
A MTM (aka. D’Appolito array) is a two-way loudspeaker system that utilizes drivers where the tweeter is centered between two woofers. Ideally the orientation of these drivers is vertical but for center channel usage, we typically turn these designs on their side for horizontal placement to snugly fit under a display.
In theory a D’Appolito array increases the SPL in the forward direction while limiting the off axis response in the axis normal to the largest dimension. For a floor standing speaker, this means the reflections from the ceiling and floor are reduced relative to the direct sound. The problem that can arise however is destructive interference occurring between the two midrange drivers which is more apparent when the speaker is oriented horizontally than the same speaker oriented vertically. This becomes more apparent when the listener moves across the horizontal plane of the horizontally oriented speaker.
W(T/M)W (Woofer – Tweeter – Mid – Woofer)
A W(T/M)W is a three-way loudspeaker design that utilizes a single midrange below the tweeter with one larger woofer on each side of the vertically mounted woofer/tweeter array. This type of approach produces better horizontal off-axis performance than conventional MTMs horizontally mounted but can arguably have other performance trade-offs discussed within this article.
Care must be taken to not compromise system dynamics at the expense of potentially wider off-axis performance, especially if your seats are never that far off-axis enough to be of paramount concern. It is important to counter the argument against horizontally mounted MTM designs and note that a well executed MTM can still offer a level of performance that equals or exceeds a “dedicated” (WTMW) center channel speaker design and almost always beat a single two-way bookshelf alternative using the same driver components with respect to dynamic range and pure SPL output. While it is true that an MTM speaker has more limited dispersion characteristics when oriented horizontally vs vertically, they can still be advantageously used horizontally mounted as a dedicated center channel which we discuss below.
Just how far off-axis from the center channel are the extreme seats really?
While the measurements from our original article include frequency response at up to 40 degrees off-axis, it important to note that most center channels are rarely listened to at even 30 degrees off-axis. In fact, 20 degrees off-axis response is quite often the worst case scenario.
Figure 1. off-axis angles for various seats
As a case in point, take a typical home theater room (see Figure 1) which has a center listening position that is 12 feet away from the center channel and the loudspeakers tweeters are roughly at ear height of the listeners in the seated position. Even with three theater seats across the front row, the listening position in the left and right seats are only 3 feet away from the center point in the center chair. In other words, the total distance we need to be concerned about for listening coverage in the front row is 6 feet wide. If you do the math (3’ off center axis that is 12’ from the center channel) ATAN 3/12= 14 degrees. At 15 degrees off-axis even a mediocre MTM design will work acceptably well. Let’s double the seating in that row and say the listener is now 6 feet off axis (ATAN 6/12). The listener at either far end seat of the theater is still only 27 degrees off-axis from the center channel.
For most multi row theaters containing multiple seats per row, you are typically further away than 12 feet from the front three speakers. Thus the further away you are from those speakers, the less the angle of incident is to the listeners. It is important to note that the further distance you place the listener from the loudspeaker, the more the soundfield becomes dominated by reflections than direct sound. Since those reflections are often as much dependent upon the off axis sound of the speaker as the on-axis sound, the loudspeakers off-axis behavior can become an even more important determinant in the sound that reaches your ears. As the listening room becomes more reflective, and/or the distance from the sound source increases, the angle of incidence to the loudspeaker becomes less important as the direct sound plays a smaller and smaller role in the sound that reaches the listener.
Let’s rework the example of being 3 feet off axis but this time from the second row which in our case is 15 feet away from the center channel when on-axis. In this case the listener on either the far left or right seat in the second row is now only arctangent (3/15) = 11 degrees off axis. If we put two more additional seats in that back row and sit the listeners at the extreme far left or right seats, they will now be 6 feet off axis from the center channel. The listener is now only 21 degrees off-axis.
I encourage you to work this out in your own theater room. All you have to do is measure the distance from your central seat which is on axis from your center channel. Then measure the distance from that position to your desired seat in that row. Next take the arctangent of the ratio of the off center divided by the other axial distance. (See example in the prior paragraph). The bottom line is in most cases you are not as far off-axis from the center channel as you might think. Try measuring your own theater and see what you come up with. If your angle exceeds 30 degrees, you may wish to consider a WTMW or two-way alternative.
What happens when you are significantly off-axis from the center channel?
A horizontally mounted MTM design which has decent 15-20 degree off-axis response could have poor coverage by the time you get to 50 degrees off-axis or more. Of course much of this depends on the distance between acoustic centers of the drivers, the distance to the listener, and the frequency range of interest. But let’s consider worst case here. Imagine our theater room again where we there is 12 feet of distance from the center channel. To get 40 degrees off axis, the extreme seat would have to be 10 feet off-axis from the center position! If the room is only 15 feet wide, that means your extreme seats are only about 3 feet from either sidewall. If you get that far off-axis from the center or sweet spot of the theater room sidewall reflections will start dominating the soundfield. Not to mention you will likely be too close to a surround channel which will dominate the soundfield. An MTM design placed horizontally can be quite beneficial with regards to limiting these reflections coming off the sidewalls in the critical 500 – 2 kHz range. We also realize a reduction of SPL in the physical space that should be dominated by the left or right speaker. So in this case, one could argue a horizontally mounted MTM would be a preferred choice despite the fact that its horizontal off-axis performance is narrower than it’s vertically mounted equal.
Technical Notes about dispersion of sound from loudspeaker systems by Paul Apollonio
Since the speed of sound does not vary with frequency, as the frequency of sound decreases, the distance between the air pressure maximums (or minimums) increases. This relationship is simply defined as C/F = W ; where C = Speed of Sound, F = frequency of sound, and W = wavelength (distance in space occupied by a complete cycle of sound; ambient pressure to max, to ambient, to minimum to ambient again; 0, 90, 180, 270, 360 degrees = 1 cycle). Remember, the speed of sound is constant, and independent of the frequency. So, when you excite the air at half the frequency, the distance between the pressure peaks doubles! That distance is your wavelength. Since the speaker producing this wavelength does NOT change in size, its size relative to the wavelength MUST change with frequency. This relationship is a primary cause of directivity in loudspeakers and is very important in multiple speaker arrays, as the distance between speakers is relatively small at low frequencies, and large enough at high frequencies to stop the speakers from acoustically coupling and acting as if they were a single speaker.
Source size is a very important determinant of the radiation of the sound coming from the speaker. As the wavelength radiated from your speaker shrinks to a size approximating half the circumference, the radiation from the speaker beams in an ever tighter pattern. Eventually the pattern becomes scalloped and uneven, with a polar pattern that looks like multiple small balloons pressed closely together. At low frequencies, the radiation from a speaker looks like a complete round balloon when that speaker is suspended high in the air away from reflecting surfaces. On the floor, a large wavelength comes off the speaker like a large bubble cut in half with the flat cut against the floor. At high enough frequencies, the speaker cone itself acts like a baffle for the radiated sound.
Like a professional line array, an MTM configuration is an array that trades off the width of the dispersion for a greater on axis SPL. The symmetry of the array, placing the high frequency driver in the center allows the effective diameter of the radiator to be equal to the diameter of the speaker plus the distance between the acoustic centers. As the frequencies get lower, the wavelengths get longer and larger, and the coupling between drivers becomes more complete. So, for an MTM array with two 6 inch drivers separated by a 1 inch tweeter on a 4 inch diameter plate, the distance between the acoustic centers of the "M's" in this array is 10 inches. The size across the long dimension is 16", while across the smaller dimension is only 6". This causes the radiation pattern (size and shape of the balloon so to speak) is going to be different across the 16" dimension than it is across the axis which is only 6". At the larger dimension, the speaker becomes directional at a lower frequency. For a horizontal mounted MTM as a center channel, the reflections off the side walls will be less than they would compared to a MT speaker, since the MT speaker is going to spread the acoustic power over a wider angle through the middle frequencies.
What about dynamic range?
Dedicated center channels (WTMW) are typically 3 way designs which have their own issues. If extreme care isn’t executed in the design of both the box and crossover, they can be less phase coherent than a simpler 2-way design. W(T/M)W designs can have less axial sensitivity and dynamic range in the upper mid frequencies than a comparable MTM design. Think about it, most W(T/M)W designs utilize a single 4-5” midrange driver as opposed to an MTM using dual 6 ½” drivers. The MTM in this case has over 2.8 times more driver surface area than the W(T/M)W design to produce the critical midrange frequencies below 2kHz. As long as the density per unit area of speaker is constant, and the radiated frequency is not so high to preclude the speaker from a uniform radiation pattern, for every doubling of surface area, you get a 3 dB increase of sensitivity. So theoretically, we can pick up a fair amount of output and sensitivity at certain frequencies when utilizing the increased cone surface area of the dual woofer MTM design over a single midrange W(T/M)W design. A good W(T/M)W can compensate for this somewhat by significantly bandwidth limiting the midrange driver and manipulating the speaker parameters to optimize it for a narrow range of frequencies. The only way of knowing which design is inherently superior in this respect is to do a side by side comparison.
As an experiment, the home theater enthusiast is encouraged to try putting on a movie that has a gunshot coming through the center channel and see how it differs between a dedicated W(T/M)W center speaker and a MTM design. The reader is also encouraged to review the related articles section below on loudspeaker power ratings.
Our hearing may be more geared for 1/4 octave resolution but if a note's fundamental or first harmonic is at 450Hz, it will be noticed. Phase cancellations are easily seen, even at 1/3 octave resolution and if it's not caused by two drivers that produce the same sound having different distance to the mic or listener's ears, it's either a crossover or first reflection issue.
Perhaps not so distracting when watching a movie, as one's position will be stationary, so the variance won't be obvious. But it will be there.
Our brains are pretty good at filtering/ignoring things when it isn't essential to hear them, just like it's good at adding things that aren't really there .
The best part was where you showed just how far off-axis you would be while sitting in a typical 3 cushion-wide sofa 12 feet away from the speakers. It isn't far enough off-axis to be real trouble.
Again, reality trumps conventional wisdom.
I would find it helpful if you could show those measured SPL levels in a polar plot such as below. Would that be possible?