10 Silly Things Churches Do With Loudspeakers

Transcription

1 10 Silly Things Churches Do With Loudspeakers By Robert Bernecker, President of SEFI Consulting, Inc. Church audio technicians are a special breed of individuals who make things work. Sunday after Sunday, they pull off amazing feats of improbable technical feasibility, only to discover all too often that the miracle they accomplished last week is now expected as the usual the next week. However, sometimes this determined effort to make things work produces a result that is counterintuitive or even dangerous. In this article, which is meant to be both educational and enjoyable at the same time, we will attempt to describe some of the things we have witnessed as we have ministered to churches. For no other reason than the fact that lists of 10 seem to enjoy a high level of popularity in pop culture these days, we have identified our own list of 10 things that we think could provide valuable lessons to church audio technicians everywhere. With no further introduction, then, we shall begin with number 10 and work our way towards the number 1 silly thing we see done incorrectly with speakers in churches. Number 10: Hook em up till it smokes! Or, we might ask, What s an ohm or two between friends? It is too easy, unfortunately, to just add another cord or another adapter and plug in yet another speaker. After all, you re told that there absolutely must be another monitor for the guest instrumentalist on his theremin because the hyperbass flute is drowning him out. 10 Silly Things Churches Do With Loudspeakers Copyright 2015 by Robert Bernecker, All Rights Reserved Page 1

2 The problem with this common practice is that each extra speaker is an additional load to the power amplifier driving that long daisy-chain of speakers, lowering the total impedance that the power amplifier must drive. Lower is not better to put this in perspective, consider that a dead short circuit across the amplifier s output terminals would have a fairly high probability of either damaging the amplifier or making it shut down, and the impedance of such a dead short would approach zero ohms. Church technicians should limit the total impedance (load) to the rating of the amplifier, which is 4 ohms in most cases or 2 ohms in some. Here s the applicable math: Church technicians, however, often have neither the time nor the interest in doing math that involves one over anything. So, if you don t care for math, just remember the simple concept that connecting two identical loads always halves the impedance seen by the power amplifier. This makes is very easy to remember that: Two 16 Ω Speakers = 8 Ohms Two 8 Ω Speakers = 4 Ohms Two 4 Ω Speakers = 2 Ohms Or, going just a bit further, it is likewise true that: Four 16 Ω Speakers = 4 Ohms Four 8 Ω Speakers = 2 Ohms Four 4 Ω Speakers = 1 Ohm 10 Silly Things Churches Do With Loudspeakers Page 2

3 If we mix & match just a bit Two16 Ω speakers & one 8 Ω speaker = 4 Ohms Four 16 Ω speakers & two 8 Ω speakers = 2 Ohms This leads us to a simple-to-remember truism: If you love your amplifier and want to keep your amplifier, then keep the load impedance above its rated minimum. Before we move on to number 9 in the list, we will offer the following extra credit quiz just to make sure that readers have a firm grasp on this important concept: Extra Credit Quiz If we connect three Galaxy HotSpots, two EV SX-100 s, one Radio Shack wall speaker (with volume control), one large brown speaker from parts unknown, Fred s guitar cabinet, two horns in the bell tower, and a feed to Aunt Gertie s ear phones what will be the total load on the 20 watt amplifier on the top shelf in the sound closet? 10 Silly Things Churches Do With Loudspeakers Page 3

4 Number 9: No need to turn em on, we ll just mount em prominently Perception is often everything. The eyes are often stronger than the ears, and we have seen this effect at many places over the years. In the course of an installation, speakers are sometimes installed that are visually prominent but have not yet been hooked up. These recessed front-fill speakers were placed by the contractor before the carpet was installed on the steps, and services were held with the platform looking as shown here: An amazing number of comments were received after these services held during construction, in which people were pleased to report how wonderful it was to be able to hear so much better from the new front-fill speakers (which were not even connected). We have seen this often for under-balcony speakers and new speakers in foyers or lobbies. It seems the reality is that if we can see them, we can hear them! 10 Silly Things Churches Do With Loudspeakers Page 4

5 Number 8: You can t get there from here (even if it did say long-throw on the box) It s a long way to those back rows! Especially from here 10 Silly Things Churches Do With Loudspeakers Page 5

6 The best way to understand the problems with this speaker installation (and to develop a better approach) is to construct a three dimensional model of the space in acoustic modeling software such as the industry-leading EASE by AFMG. Here is a view of the same sanctuary, modeled in EASE: It is easy to observe where the speakers are located, and it can be seen that audience planes have been inserted at ear level, floating above the pews and also above the platform. EASE will map the sound from the speaker(s) onto these planes. It is usually easier to study the results of this mapping by simply looking at these audience planes in 2D, and such views are presented on the next page Silly Things Churches Do With Loudspeakers Page 6

7 Let s take a look at how the coverage from these particular speakers maps out in the acoustic modeling software EASE. Of course, we would like to see a fairly even direct sound coverage to all of the seats which the speakers are intended to cover. Here is the EASE direct SPL mapping for the 2000 Hz octave band: The coverage map (above) is showing a hot spot in the center where the predicted SPL is nearly 102dB, as compared to the back of the room, which is 6 to 8 db lower at 96 to 94dB. This isn t great or even good, but it isn t horrible either (we ve seen worse). The problems begin to get much more pronounced when we look at the 500 Hz and 250 Hz octave bands. At these lower frequencies, we do not have the directivity of the HF horn to help us get to the back, and cone drivers simply do not have enough pattern control. On the next page, we will examine the coverage at 250 Hz and 500 Hz Silly Things Churches Do With Loudspeakers Page 7

8 500 Hz Octave 250 Hz Octave 10 Silly Things Churches Do With Loudspeakers Page 8

9 It is easy to see that the speakers are providing almost no pattern control, acting more or less as spherical radiators at these frequencies (and below). Moreover, be sure to notice the location of the pulpit icon in the lower (250 Hz) plot. Since the pulpit is in the hot spot of the coverage of these existing speakers in this frequency range, feedback is obviously going to be a problem. We designed a superior solution for this church, and the EASE mapping for the 250 Hz octave band is presented here: 10 Silly Things Churches Do With Loudspeakers Page 9

10 Now take a look at the old vs. the new coverage for the 250 Hz octave band. Notice we are no longer spilling all over the platform, and notice how much more even the coverage is across the congregation seating area. We have moved the highest SPL level area of the direct coverage from the pulpit area to a much larger area well out into the congregation Silly Things Churches Do With Loudspeakers Page 10

11 How did we accomplish this? What could make such a big improvement? Take a look at this section view from the EASE model: Recall that the speaker in question had little to no pattern control at 250 Hz, and observe that it is only 20 down to the front row seats, while it is over 60 to the back row. By the inverse square law, this would be a 10dB difference. Instead of trying to get there from here, a better approach in this situation is to add a second speaker about half way back. This speaker must be delayed appropriately, so that the sound arriving from it to the listeners in the rear of the space does not arrive before the sound from the front loudspeaker. We were also able to raise the speakers slightly in the new design, and now the difference between the closest seat and the furthest is closer to the difference between 22 and 30, which is only on the order of 2.5 to 3dB in difference in direct SPL. The section shown on the top of the next page illustrates why this is easily a superior solution: 10 Silly Things Churches Do With Loudspeakers Page 11

12 It gets even better, however. There is another huge benefit from this improved design, and this will be seen when we consider what is happening with destructive and distracting reflections from the back wall. An ETC (energy time curve) measurement with the old speakers taken in the congregation seating area (11 rows back) shows a very strong discrete reflection from the back wall that is arriving about 60 to 70ms later than the direct sound from the main speaker, and it is standing out above the beginning of the reverberant decay by at least 15dB. This means it will be a problematic reflection Silly Things Churches Do With Loudspeakers Page 12

13 The geometry causing this reflection from the back wall is no mystery, as the path is easy to follow: It should be no surprise, then, that the musicians on the platform were also complaining about a distracting and annoying echo. Here is what we measured on the platform, right at the keyboard player s position, revealing the strong reflection coming in over 100 milliseconds late (notice it is much stronger than the arrival from the main speakers): 10 Silly Things Churches Do With Loudspeakers Page 13

14 We could treat the rear wall with absorptive material, and that may well be the most reasonable approach in some cases. In other cases, however, adding such absorption may make a room too dead for its intended purpose. In this case the bite back from the rear wall can also be tamed by a better loudspeaker design. By increasing the down angle on both the main speaker and the fill speaker, the amount of sound reaching the rear wall is dramatically reduced, much of what does hit the rear wall is now going down into absorptive pews, and the inverse square law is now working in our favor. Compare these two scenarios: 10 Silly Things Churches Do With Loudspeakers Page 14

15 The modeled ETC of the new speaker design for a listener near the same location about 11 rows back reveals a very well-behaved space. Following the arrival of the strong direct sound are some early reflections around 10dB down, followed by a nice initial delay gap, and a nice decay of the reverberant field without any of the specular reflections that were measured previously. The EASE model allows us to see that this design will work well without the addition of costly absorption to the back wall, and, more importantly, we know the design will work well before the church spends the money to install it. The lesson to be learned from this case study is that churches should not expect to just hang a good speaker at the front of the sanctuary and automatically expect good results. Sometimes, the throw required for the trim height available makes good performance for the entire space from that location impossible. As usual, there is no substitute for careful design augmented by good modeling techniques Silly Things Churches Do With Loudspeakers Page 15

16 Number 7: Homemade Speakers (Why buy them when you can build them yourself?) We have often seen homemade speakers at various churches we visit. Presumably, the church thought that these speakers could be built better or cheaper than those from a real manufacturer. While there are certainly high-fidelity hobbyists that build some fantastic sounding speakers for their home hi-fi systems, we have never seen, heard, or measured any homemade speakers in churches that were even in the same class as good professional loudspeakers. One particular company that shall go nameless in a city that shall remain unnamed made a practice of preying on unknowing congregations all over that city. We took down his homemade speakers from way too many spaces. If he was bidding against $1100 professional speakers, his speakers (always the exact same model everywhere) were sold for $900. If he was bidding against $700 professional speakers, his speakers were sold for $500. However, when we took these speakers apart, we discovered they were built from less than $50 worth of parts. Inside was the cheapest of cheap stamped out cone driver, a five dollar piezo tweeter, and a single capacitor wired in series as the crossover network. Some churches had been sold 8 to 10 of these speakers in clusters. Don t let your church be a victim. Buy reputable loudspeakers from reputable manufacturers that were skillfully selected with careful modeling by a qualified design professional as an appropriate choice for your sanctuary. Another consideration is safety. All of the homemade speakers that we have seen have been sorely deficient in rigging integrity. Most often, inexpensive eyes or brackets are observed simply screwed into the usually particle board cabinet with wood screws or drywall screws. Simply stated, speakers flown without the proper internal bracing and support for rated hardware are a danger to those under them and a liability for the church that put them there Silly Things Churches Do With Loudspeakers Page 16

17 Things are not always what they seem! Here s a loudspeaker cabinet that was deployed at a church that was clearly marked as a 2-Way loudspeaker. We have no idea who made it for them or how much they paid. We do know, however, that removing the grill cloth tells a different story. The drivers are quite literally painted on, there is no tweeter, and the little cone is painted to make it look bigger through the grill Silly Things Churches Do With Loudspeakers Page 17

18 Number 6: One size fits all! Many church sound experts don t realize that what s good (or even great) in one place may be totally wrong in another. A good speaker in the wrong place can be just as poor as a bad speaker in the right place! One size does not fit all. Sometimes this is immediately obvious, such as this speaker that is clearly the wrong speaker for the room, although someone clearly thought it was a great speaker. The low-q nature of this loudspeaker is spraying sound all over the parallel walls in this long and narrow sanctuary. A quick look at the havoc caused by this deployment is shown on the top of the next page Silly Things Churches Do With Loudspeakers Page 18

19 We have absolutely no doubt that someone had heard that particular speaker somewhere else and decided that it sounded good to them. Therefore, the reasoning must have gone, it was a fantastic speaker, and it would certainly be a good choice for their sanctuary. Hopefully, no reputable audio contractor would have hung that speaker in that position in that room. We will return to this speaker a little later, and we will take a closer look at the problems with this silly installation. Let s look even closer at why one size does not fit all. We will refer to this as A Tale of Two Churches. Both of the churches are real projects, and both were very pleased with the results of the systems we designed for them. But, after we considered using the same high-quality professional speaker in the 2 nd project that we had already used in the 1 st project, good modeling revealed that it was not the right choice, and completely different loudspeakers were deployed Silly Things Churches Do With Loudspeakers Page 19

20 A Tale of Two Churches 10 Silly Things Churches Do With Loudspeakers Page 20

21 A great speaker with great coverage in one church Rectangular Converted Theater Doesn t cut it in the other Octagon Sanctuary but not because it doesn t cover the seating well. The coverage of the seating area is great the problem is what is going elsewhere! 10 Silly Things Churches Do With Loudspeakers Page 21

22 Here s the real problem: The model revealed that the same speakers that worked great in the first church spilled too much sound onto the side walls, the back wall, and the ceiling of the second. And here s what we actually used: These speakers eventually selected and deployed used a tighter horn pattern and a bipole LF design to minimize spill onto the side walls and rear wall. Notice that the sound booth is still covered well. One size definitely does NOT fit all! 10 Silly Things Churches Do With Loudspeakers Page 22

23 Number 5: Push or Pull? Or, why can t the speakers all just work together? If a person is on one side of a door attempting to push it open with 20 lbs. of force, and another person is on the other side attempting to push it closed with 20 lbs. of force, we know intuitively that the door is not going to move. The force pushing it open is cancelled by an opposite force pushing it closed. Unfortunately, too often church audio technicians do not keep the same principal in mind when it comes to the polarity of the loudspeakers in their systems (or any other device in the systems, for that matter). Polarity is too often either ignored altogether or confused for phase (which is a timerelated issue, not a wiring issue). While the absolute polarity of a single-driver loudspeaker is probably inaudible in most cases, improper relative polarity between multiple loudspeakers or drivers is a crucial and foundational issue. To better understand this basic concept, consider the simple diagram below: 10 Silly Things Churches Do With Loudspeakers Page 23

24 When several loudspeakers, or drivers within a single loudspeaker, are unintentionally connected with incorrect polarity, one driver will be creating positive pressure during the same point in time that the second out-of-polarity driver is creating negative pressure, and this difference will produce cancellations at the listener s position. This cancellation will not be complete because the path lengths involved will be slightly different, the sound does not emanate from an infinitely small point on the drivers, and there will be reflections in any real situation. However, because the much longer wavelengths of the lower frequencies minimize the effects of these differences and the shorter wavelengths of the higher frequencies are more sensitive to such cancellation-reducing differences, the cancellation will always be much greater in the low frequency range than in the high frequency range. These facts lead us to two simple suggestions: 1. If the system is producing thin sound with no bass and lots of treble, think of polarity issues first. 2. Check the system from end to end for polarity, using a polarity checking device if possible. This is critical for multiple devices within an array or for multiple subwoofers stacked together. There are many ways to check polarity, ranging from simple to sophisticated. We often use an NTI XL2 for this purpose when commissioning an audio system one of the very first steps in this process is to check the polarity of every device. One unfortunate but crucial fact is that even when a system is installed with great care and considerable skill, loudspeakers can come from the manufacturer with polarity issues inside the cabinet. In fact, they often do. As a result, even if the wiring has been checked and re-checked, the final output of each loudspeaker must also be verified Silly Things Churches Do With Loudspeakers Page 24

25 Here is our XL2 being used to measure loudspeaker polarity. It is shown set in a simple go or no-go mode. And here the XL2 is again measuring polarity of the same loudspeaker, but this time it is set to provide more frequencyspecific information than just the go or no-go indication. This sort of information allows us to see which driver may (or may not) have issues, as well as to evaluate the effects of such things as tuned ports Silly Things Churches Do With Loudspeakers Page 25

26 Other instruments that can make checking polarity quicker and easier are the Galaxy Cricket or the CAB-DRIVER and SC48RJ checkers from Whirlwind. A church might also consider the excellent Studio Six Digital professional audio tools for ios that run on the iphone and ipad. Potential Sources of Polarity Issues 1. Connections reversed at the power amplifier(s), or a swapped polarity in the line level feed to the amplifier (pins 2 & 3 of an XLR connector soldered incorrectly). If banana plugs are used at the power amplifier, these are often inadvertently inverted during setup or troubleshooting. 2. Mis-wired jacks on the platform, snake, or console. Color-code errors in the installed wiring. 3. Improper settings in system processor or DSP. 4. Mis-wired speaker cables. 5. Mis-wired speaker cabinets. 6. The phase buttons on the mix console, which are in reality polarity-swapping buttons Silly Things Churches Do With Loudspeakers Page 26

27 Number 4: Hang em high! (or why speakers and ceilings don t always play nice) Remember this room from our silly problem number 6 - One Size Fits All? We are now going to pick on this sad-but-true installation just a little bit more. It isn t the height of the speaker above the congregation that is the problem in fact, such height is often a good thing. Rather, the problem here is that we have a very low-q loudspeaker mounted up against a reflective ceiling. By low-q, we mean a loudspeaker that has very little directivity, one that tends to spray sound in every direction Silly Things Churches Do With Loudspeakers Page 27

28 Directly below is a closer look at the mess that this situation causes. The ceiling reflections are strong and early, and the side wall reflections are exhibiting a fluttering as the sound bounces back and forth between the side walls. The problem we are going to look at more closely here, however, is the comb filtering that is caused by the ceiling reflections. This will reveal to us why this was such a poor mounting location for what was already a poor loudspeaker choice for the space. The reflections from the ceiling, as shown in the TEF time measurement above, arrive to the listener later in time than the direct sound from the speaker. Whenever we have a situation in audio where a signal is summed with a copy of itself that is delayed in time (a time domain event), we will always see cancellations in the frequency response measurement (a frequency domain measurement). It is worth noting that these cancellations cannot be fixed with an equalizer because they are a time domain problem (late arriving reflections) that can t be fixed by manipulating the frequency response of a system (the function of an equalizer). The frequency response measurements from this speaker that are shown on the top of the next page exhibit the telltale comb-shaped response that shows up when early reflections are present Silly Things Churches Do With Loudspeakers Page 28

29 We have had some people suggest that all real-life systems measure similarly to this mess. However, this is certainly not the case. Compare the train wreck above with the measurement of a well-tuned system below that was designed and installed to be free of the comb filtering problems caused by reflective surfaces near the loudspeaker. The measurement below is displayed with the same minimal amount of smoothing (10%) as the measurement above, and it was taken with a significantly larger time window, which means that the measurement below actually has better frequency resolution than the measurement above (250 Hz vs. 67 Hz). This means the measurement below would reveal even more detail in anomalies in the frequency response than the measurement above if they were present. Which do you think would sound better? 10 Silly Things Churches Do With Loudspeakers Page 29

30 In order to demonstrate how this audibly destructive effect is also both measureable and predictable, let s take a look at one more situation. This is a measurement of a speaker mounted directly to a side wall near the front of a sanctuary. This is a mounting position seen all too frequently in smaller churches. On the ETC below, it is easy to see both the direct sound from the speaker AND the early strong reflection from the wall that it is mounted against. In the ETC measurement shown above, the reflection from the side wall arrives to the measurement microphone (the listener position) about.53 milliseconds behind the direct arrival from the loudspeaker itself, which translates to just a bit over 7 inches longer in path length. We can take this.53ms difference in arrival time and calculate the frequencies at which we should expect to see cancellations, as well as where the first peak (summation) will occur. These calculations are shown here: 10 Silly Things Churches Do With Loudspeakers Page 30

31 Does it measure as expected? We see that the comb filtering indeed takes place in the actual frequency response measurement exactly where we predicted it should based on the problems we saw in the time domain measurement. We also see the predicted peak in the 1800 Hz to 2100 Hz area. You will recall that we have already mentioned that these cancellations cannot in any way be fixed with EQ. Even if these cancellations could be fixed with an EQ (they can t), the problem would still be unworkable because we have to keep in mind that if we move just a few feet in any direction, the path length differences will change and all of the cancellations move to different frequencies. That is the joy of a sanctuary with excessive comb filtering no two seats sound alike! 10 Silly Things Churches Do With Loudspeakers Page 31

32 The lesson for churches is an easy one. Churches should avoid mounting a loudspeaker near a reflective surface. If a speaker absolutely must be mounted against a ceiling, for example, then measures should be taken to minimize the issues. Often something as simple as hanging a 2-way loudspeaker inverted (with the HF horn below the LF driver) instead of the usual horn side up orientation can make a large improvement. In the photo below, a low ceiling and difficult geometry forced us to design speakers installed tight to a ceiling. However, we carefully selected high-quality speakers (by Danley here) with a big enough and well-designed horn that would provide a good amount of pattern control and thereby reduce the amount of mids and highs hitting the ceiling surface. In this case, we did not mind coupling to the ceiling in the lower frequencies (as a technical matter, this would be where the reflections arrive to the listener within a quarter of a wavelength of the frequency being considered; such reflections are usually constructive rather than destructive). Finally, note that we specified absorptive material to be added to the ceiling in the problematic area. If it has to go tight to the ceiling 10 Silly Things Churches Do With Loudspeakers Page 32

33 Number 3: Left and Right works great at home so why not at church too? We have all seen it. In fact, we have all seen it so often that most of us don t even think about why it might be a problem. Like many things in church audio, the issues are not always obvious or intuitive. What is so ubiquitous of a problem that it rises to number 3 on our list of silly things done with loudspeakers? We are talking about a pair of speakers split left and right in a small to medium size church. Left and right speakers are presumed to be better just like the stereo speakers at home or the ear buds from a smart phone. Most people don t realize the major difference. Real stereo sound requires that the listener be able to hear both channels equally well and implies that the program material in one speaker (or ear bud) is different than the material in the other. Aside from those in a fairly narrow area right down the middle of a sanctuary, most of the listeners in places where a pair of speakers is deployed split hard left and hard right can only hear one of the speakers well. A good illustration is this pair of custom made (homemade) cabinets mounted to each of the side walls at about ear height. Be sure to duck as you walk down the aisle! 10 Silly Things Churches Do With Loudspeakers Page 33

34 Here is a closer view at one of these cabinets: This one qualifies for an award under our previous discussion of speakers near reflective surfaces (silly practice number 4), because it is mounted adjacent to both a reflective wall and a reflective ceiling! Let s first list the many problems with this installation: Speakers against not one, but two reflective boundary surfaces. Homemade speakers. The coverage patterns of the two speakers overlap significantly in the middle seating area (comb filtering). The speakers are 6 feet from the front row with no elevation, and 45 feet from the rear row Silly Things Churches Do With Loudspeakers Page 34

35 We have already seen the problems caused by mounting speakers against reflective boundaries, and we have also discussed the pitfalls of homemade speakers. The other issues here are a result of the coverage from these speakers, so let s take a look at an EASE coverage map (since these are homemade speakers, representative data from 15 two-way boxes with a nominal 60 x 40 coverage pattern was used in the model). A quick examination of this coverage plot shows that the modeled direct SPL is around 103dB for those listeners up front and below 89dB for those listeners in the rear, so there is obviously a huge difference in direct SPL. Of greater concern, however, is the area where the coverage patterns of the two speakers overlap, which in this case will be the majority of the room. The problem with this excessive overlap is that the listener hears the direct sound from each of the two speakers (not just one), and because the sound has to travel a longer 10 Silly Things Churches Do With Loudspeakers Page 35

36 distance from one loudspeaker or the other (dependent upon where the listener is sitting), the sound does not arrive at the listener at the same time. This is best illustrated with the sketch below: The orange arrows represent the nominal 60 degree horizontal coverage of the HF horns, so the excess overlap is very obvious. For those in the congregation seated exactly on the center line (the dashed gray line), as at position A, they will hear the sound from the left speaker and the right speaker at the exact same time (hypothetically, since both ears cannot actually be on the center line at the same time, and we have already seen how path length differences of only a few inches cause severe comb filtering). However, for a person in the congregation seated at position B, they will receive the sound from the left speaker (the red line) sooner than the sound from the right speaker (the green line), because the red path is shorter than the green path. As we have seen previously in this article, the difference in path length can be stated in either distance or milliseconds, and the resulting comb filtering is predictable and measurable. Of course, this comb filtering is also audible producing timbre shift and deep notches in the frequency response. The perceived image (source) also shifts to the speaker with the earliest arrival time Silly Things Churches Do With Loudspeakers Page 36

37 A person seated at position C first hears the sound from the left speaker (the purple line) and then the sound from the right speaker (the cyan line) later. However, because the purple line is shorter than the red line, and the cyan line is longer than the green line, the person at position C will experience a completely different set of detrimental comb filtering effects than the person at position B. How can the sound engineer mix effectively when some in his or her audience have a deep notch at 250 Hz, while others have a deep notch at 630 Hz and no notch at 250 Hz, and still others have a deep notch at 140 but no notches at 630 Hz or 250 Hz? The takeaway is that speakers installed in configurations such as this produce an infinite variety of comb filtering effects across the congregation, the specific effect at any given seat being a product of the various differences in path lengths at that exact location. Here is a measurement taken in this sanctuary at a position approximately 2 feet left of the center line (position B above): 10 Silly Things Churches Do With Loudspeakers Page 37

38 The two arrivals, one from each of the loudspeakers, are easy to see, as is the difference in arrival time due to the difference in path length. As a side note, because the speakers are positioned roughly at ear level and firing horizontally at the rear (reflective) wall, the late reflections from the rear wall are quite strong, only down about 6dB from the direct sound of the right speaker. In order to illustrate the disparate comb filtering at every position, we have here overlaid the frequency response measurements taken at each of the three positions in the diagram above (positions A, B, and C ). Even if these problems could be fixed with EQ (they can t), where would one begin to EQ such a situation? 10 Silly Things Churches Do With Loudspeakers Page 38

39 A better approach would be to install speakers in a manner that minimizes overlap, as well as negative interaction between the speakers. As a simplified example, consider the alternative below (but please note that it is not possible to just pack two ordinary speakers together and accomplish this result!) The beauty of this approach is that in the area down the center line, the path length difference is minimized, while seats to the left and right are increasingly out of the pattern of the speaker which would otherwise produce a strong, late arrival. In summary, here is a good general rule: For most sanctuaries, it is a bad thing for two or more speakers or speaker arrays which are being fed the same signal (mono) to be covering the same seat. For better results, minimize overlap between speakers and array elements Silly Things Churches Do With Loudspeakers Page 39

40 Number 2: Would you want your family under that thing? Over the years, we have seen many methods employed for flying loudspeaker cabinets, many of which we can only charitably call resourceful. If we can only make one point in this entire article, it would be this: Flying loudspeakers over people s heads is a lifesafety issue; hire a professional who has the knowledge and skill to do it properly. Unfortunately, too many churches do not observe this rule, and there are too many dangerous situations thereby created. We can only look at a few here. Our first example is all too common. People somehow mistakenly assume that plastic carrying handles are an appropriate attachment point for flown speakers. Speakers are not dogs. This would seem to be an obvious fact, but too many people have no problem using dog chain wrapped through a plastic handle to suspend a hundred pounds or more over the heads of their families. Don t do it! It should also be noted that the s-hook used to double this dangerous chain back on itself is also wrong and dangerous. If the speaker does not have fly points installed by the manufacturer, it should not be flown Silly Things Churches Do With Loudspeakers Page 40

41 You would not want to be under this speaker. It is obviously not a speaker designed to be flown because there are no fly points. The installer has simply used screw eyes threaded into what is most likely particle board. The chain used is inappropriate, and the three chains all go to a single point of failure. Stay clear below! This photo showed up on an internet forum a few years back, ostensibly posted by someone who was proud of their install. There are really too many problems to list here. See how many you can spot. Be sure to note the plumbing strap holding a pulley to the bottom of the array, presumable for hoisting additional load. Danger! Generic hardware store s-hooks and cable is another dangerous practice that is seen too often. Danger! 10 Silly Things Churches Do With Loudspeakers Page 41

42 These dangerous homemade beam clamps, along with the hardware store pulley, were actually removed from a job by a pro contractor before they killed someone. Even loudspeakers that appear to have fly points from the manufacturer can be dangerous. When the horns were removed from these cabinets to be rotated, it was observed that the fly points were no more than standard tee nuts which were also installed without any appropriate internal support or bracing Silly Things Churches Do With Loudspeakers Page 42

43 Can the rigging handle 150 MPH winds? This cabinet was one we flew many years ago, and although a hurricane took off the roof of the building and destroyed the CBS wall at the front of the platform, the loudspeaker did not come down! A professional rigger will use hardware more like this a rated, forged eye, a rated safety shackle with cotter pin to secure the nut, and Grade 80 or better chain rated for overhead lifting Silly Things Churches Do With Loudspeakers Page 43

44 5 Rules to Follow 1. The big hardware store down the street does not carry any hardware appropriate for hanging loudspeakers. 2. The big hardware store down the street does not carry any hardware appropriate for hanging loudspeakers (just in case Rule #1 was not clear). 3. If the manufacturer did not provide fly points on the loudspeaker box, the box should not be flown. 4. A structural engineer should certify that the building structure can handle the weight of a speaker cluster or array, and a mechanical engineer should certify the rigging to be used. 5. If your business isn t rigging, then you shouldn t fly heavy objects over the heads of your congregation. Hire a professional! 10 Silly Things Churches Do With Loudspeakers Page 44

45 The Number 1 Silly Thing Churches Do with Loudspeakers: Tight Packed Arrays Or at least the poorly designed ones which are most of them. It looks easy enough. It even seems to make good sense to build a cluster and hang it right there over the pulpit. But in fact, homemade tight packed arrays do not perform the way churches expect them to. Nor do many professionally-designed tight packed arrays. The reality of what the array is doing is often very different from what the church thinks they are getting, and they wonder why they are fighting feedback problems and experiencing poor coverage. It will take some space to fully explain what happens with deployments of this sort, but an understanding of this subject will serve one well for years to come. Let s take a close look at a great example of a tight packed array that was constructed from trapezoidal loudspeakers (also known by many as trap cabinets ) made by a reputable manufacturer. But, as we shall see, the design was a poor one, and these speakers should not have been arrayed in this fashion. It is not at all that the loudspeakers were shabby speakers; in fact they were good speakers. The mistake is building tight packs from speakers that don t array well, this without giving proper attention to how the low frequency drivers will interact with each other in such an array. This ubiquitous tight pack array seen here flown over the center of the front lip of the platform consists of three 12 2-way boxes. Without proper design, two major flaws show up over and over in this type of arrays: 10 Silly Things Churches Do With Loudspeakers Page 45

46 1. Many, if not most, such arrays have excess overlap in the mids and the highs. Obviously three 90 degree horns that are only splayed 45 degrees are going to overlap by 45 degrees at each seam. 2. Putting three 12 or three 15 surfaceloaded drivers in a horizontal array forms what we call an accidental line array, the pattern of which is oriented in exactly the wrong direction. We will first examine the major issue of excess overlap. If we look at the EASE data for the particular cabinet used in this array, we can see the actual coverage balloon that could be expected from this HF horn at 2000 Hz (shown to the right here). Since the coverage from this single horn is often too narrow in the horizontal direction, many churches and some professionals who don t know better will combine three cabinets, thinking that each horn will simply add more coverage in the horizontal direction. We can model this imagined (but not real) coverage by turning off the part of EASE that calculates the interaction between multiple drivers. This produces an imaginary attenuation balloon of what the array builder was hoping to achieve Silly Things Churches Do With Loudspeakers Page 46

47 However, reality comes calling. If we ask EASE to show us what really happens because of the overlap between the three HF devices (horns), the resulting attenuation balloon is much more revealing. This lobed monster is in fact comb filtering in three dimensions. Of course, because these lobing patterns are caused by the variability of multiple path lengths in various directions, the shape and direction of the fingers of the lobes will be different at every frequency. To the right is the attenuation balloon for the 4000 Hz range instead of 2000 Hz. It is vital to understand what we have just discovered, so let s recap it. The array builder thinks he or she can add additional smooth, even coverage by bolting together three trapezoid-shaped cabinets that seem to fit together well enough. But what was hoped for is not what actually happens: 10 Silly Things Churches Do With Loudspeakers Page 47

48 It is important to understand why this happens. We need to visualize how the HF sections (the horns) of these speakers are interacting. This simple diagram shows a tight packed array in which the nominal coverage pattern of the horns is overlapping far too much. In this case, the person sitting in position A is actually within the coverage pattern of all three devices. Since he or she is also a different distance from all three devices, this listener will receive three HF arrivals that are offset in time from each other, and the resulting comb filtering will distort the response for that position. If the listener moves over a few feet, the comb filters move in frequency, but the response is still distorted. Whereas the listener may have been sitting in a lobe at 2000 Hz and a null at 4000 Hz, by moving over a few feet he or she may now be sitting in a position with a null at 2000 Hz and a lobe at 4000 Hz (pulling numbers out of the air here). A listener seated at position C is basically out of the coverage pattern of the leftmost horn, but he or she will still be receiving multiple arrivals with multiple path lengths from the horns in the other two loudspeakers. The same concept would be true for a person at position B or any other position. The fact of the matter is that the side panels of most, but not all, trapezoidal cabinets are constructed at such an angle that a tight pack (where the side panels are tight to each other) will result in excess overlap. This is because the angle of the side panels does not correctly compliment the coverage angle of the horn in the speaker Silly Things Churches Do With Loudspeakers Page 48

49 There are, however, some better loudspeakers available on the market that are designed to address this issue, and these do a much better job of arraying as a tight pack (in the HF section, at least). The HF coverage angles when tight packed looks more like this: In this case, we can see that listeners seating in positions A, B, and C are all located within the nominal coverage pattern of only a single horn from a single cabinet in the array. We might therefore say that these cabinets array well. Listeners seated in the much smaller, narrow overlap zones will still have some comb filtering, but the effect is minimized as much as possible. So far, we have only been considering the coverage from the horns of these cabinets in a tight packed array. However, we must also consider how the 12 or 15 drivers typically seen in these sorts of arrays combine. Unlike horns which can be designed to more or less achieve some level of constant directivity, the surface-loaded drivers inevitably lose directivity as the frequency decreases. We will explore the ramifications of this fact at length, but first we will examine how what we have learned about the dangers of tight packed arrays with excessive HF overlap plays out in the real world. Let s see how this worked in a real project with real devices Silly Things Churches Do With Loudspeakers Page 49

50 This is the sanctuary we showed at the outset with the 3 cabinet tight packed array flown in the center. We first constructed an EASE model of the space, which allowed us to evaluate the coverage from the existing tight pack, as well as to develop a better solution for this congregation. The sanctuary seats about 1500 people, and it included a balcony and under-balcony seating areas. For the sake of this illustration, we will only be looking at the coverage in the 2000 Hz octave band. In a full design, the coverage and array interaction in all of the frequency bands must be carefully considered. If we model only the center speaker in the array, the coverage over the center section is not too bad, although the hole seen in the middle seems anomalous Silly Things Churches Do With Loudspeakers Page 50

51 If we again set EASE to ignore how the array elements interact, we can roughly map the expected coverage from the tight packed array onto the seating area. We can see that the coverage does indeed widen out and again doesn t look too bad aside from the holes in the center of the patterns (which can also be seen in the balloons illustrated on page 46). When we set EASE to calculate and show what actually happens with the interaction of the 3 arrayed horns with excess overlap, the reality is uneven, lobing coverage. The astute reader will have already realized that this is what happens when the mangled attenuation balloon shown on the top of page 47 is mapped onto a real seating area. It is also an important part of the design process to consider where else the lobes from the array might be going. In this 3D view from inside the space, we can see that the lobes go up too, and they are shown here hitting the reflective ceiling, from which they will be reflected back down into the seating areas Silly Things Churches Do With Loudspeakers Page 51

52 It is vital to understand, therefore, that bolting three horns together doesn t automatically produce the expected results. This is an important enough concept that we will present the very same EASE mapping side-by-side so that the differences can be easily observed. A Better Way It is often better to avoid the many issues of tight packed arrays completely. In this case, that is exactly what we did. Instead of a center cluster, we employed an exploded array. The exploded array consisted of 5 loudspeakers flown in an arc across the front of the seating area (the cabinets seen flanking the center cabinet shown here are flown subwoofers) Silly Things Churches Do With Loudspeakers Page 52

53 We designed this solution using high quality Renkus-Heinz loudspeakers which employed a large, well-controlled mid and high horn, along with vertically oriented pairs of LF drivers located above and below the horn for improved LF vertical pattern control. Here is the coverage provided by this superior design that replaced the existing tight packed array: What a difference! Now all of the 5 main seating areas on the main floor are covered very well. We are not spraying a lot of sound all over the side walls (notice how the coverage is falling off nicely right at the outside edges of the seating areas). Even more important, we have dramatically reduced the amount of sound spilling onto the platform, which will give the congregation a huge improvement in gain before feedback. We have carefully placed the seams of the exploded array in the aisles where people are not sitting. In the seating areas, each listener will be hearing predominantly only a single arrival from a single cabinet, thereby minimizing comb filtering. By applying the correct amount of delay to each of the four non-center cabinets in the exploded array, we keep 10 Silly Things Churches Do With Loudspeakers Page 53

54 the perceived image shifted to the center and we virtually eliminate any audible comb filtering in the seating areas. This delay is a vital part of the design. Finally, we used a zoned delayed ring of compatible loudspeakers to nicely cover the balcony seating areas, and that is not reflected in the coverage map above. To recap, in this real world project, the existing tight pack array, with all of its associated problems, was eliminated and replaced with a well-designed system that greatly improved the sound for the congregation by fixing the coverage, lobing, comb filtering, and fidelity issues. This was accomplished by putting the right loudspeakers in the right places, increasing the quality of the loudspeakers, and carefully selecting and placing loudspeakers such that sound was directed where it was needed and kept off of where it was undesirable. We also carefully tuned the system, measuring and setting delays precisely, shading low frequencies from the flanking array elements as needed, and applying EQ to make the whole system perform together as a high-performance unit. We even received the obligatory comment from the ushers that the under-balcony fills were not on. We took this as a compliment meaning that the delays, EQ, and levels were such that the operation of these speakers was completely transparent to the listener Silly Things Churches Do With Loudspeakers Page 54

55 Accidental Line Arrays The final thing we must consider with tight packed arrays especially the homemade ones is how the low frequency (LF) drivers combine, couple, and interact. In the lower frequency bands (60 Hz, 125 Hz, 250 Hz, and 500 Hz), the wavelengths are much longer than the 2000 Hz band we have been examining up to this point. For example, a 2000 Hz sine wave has a wavelength of about 7 ½ inches, while a 250 Hz sine wave has a wavelength of about 4 ½ feet. Because the wavelengths change but the path length difference between the arrayed cabinets does not (the physical dimensions are the same), this means that as a practical matter we are dealing with a different kind of interaction between devices in the LF range (although the physics is actually the same). In the HF range of the horn, we were primarily concerned with excess overlap. In the lower ranges of the LF driver, the driver does not have near the directivity of the HF horn. This is one reason that calling a 2-way cabinet with a 12 surface-loaded driver a 60x40 box or any other such nomenclature is really disingenuous. The 12 driver is anything but 60x40 at 250 Hz. As such, it is safe to say that for any box with surface-loaded drivers, we are always dealing with we would probably call excess overlap if it were occurring in the HF sections. The attenuation balloon of the same 12 2-way loudspeaker that we been evaluating as a tight packed array is presented below. The balloon on the left is a single loudspeaker s balloon for the 500 Hz range. If we array three of these cabinets in a tight pack, we intuitively expect to get a wider, bigger pattern, such as the one shown on the right that is the EASE simulation of what would happen if there were no interaction between drivers Silly Things Churches Do With Loudspeakers Page 55

56 Armed with this concept, a church expert decides that it would be wise to tight pack three MI (music store) boxes, most often 12 2-way or 15 2-way models. Cabinets are bolted together and flown above the pulpit. The church is expecting to get this: 250 Hz 500 Hz But, when the system is turned on, the church wonders why there is such a feedback problem at the pulpit and near the center of the platform. They also can t figure out why the sound is muddy in some seats and thin in others. The answer is that the homemade tight pack cluster is not doing what they thought it would do. If we set EASE to include the interaction between array elements, we now can see that the church is in reality actually dealing with this: 250 Hz 500 Hz When we see these patterns, and we realize that the pulpit is right underneath this thing, we can understand that it is no surprise that the system is feeding back at 500 Hz. This pattern might actually be useful if it were rotated 90 degrees and turned on its side! Because of this, we often refer to these homemade disasters as Accidental Line Arrays Silly Things Churches Do With Loudspeakers Page 56

57 For the average homemade tight pack array, the problems are usually worst in the 250 Hz and 500 Hz octave bands. In the 125 Hz octave, the wavelengths are now around 10 feet long for a source that is only 3 or 4 feet across. As such, the signals from all three drivers arrive to most listeners within a quarter-wavelength of each other or less, and therefore the result is constructive summation instead of destructive cancellation. Moreover, in the 125 Hz range and below, we usually need the acoustic output of all three drivers combining coherently. At the upper end of the LF driver s range, in the 1000 Hz octave band that is usually right below the crossover frequency, we see that the wavelengths are getting short relative to the size of the driver (usually 12 or 15 ). As a result, each driver begins to exhibit some significant directivity, and this pattern control tends to limit the amount of destructive interference produced from the driver(s) on the opposite side of the array. We are sometimes asked to explain why this accidental line array effect happens. This is essential knowledge for church sound technicians, especially because line arrays are currently such a popular option for churches. We should first consider that in the operating range of about 500 Hz and below of a 3 box tight pack array with surfaceloaded drivers, the vast majority of the people in the congregation are seated in a position where they will be receiving a substantial arrival from each of the three drivers Silly Things Churches Do With Loudspeakers Page 57

58 In other words, people at positions A, B, and C are easily within the pattern of all three surface-loaded drivers. However, the person at position A is nearly equidistant from each driver, so the sound from each driver arrives at A within the quarterwavelength window we mentioned earlier. The person sitting at position C, however, receives the sound from the most distant driver (the one pointed at position B ) later than the sound from the closest driver (the one pointed at their position). As the difference in time between these two arrivals approaches a ½ wavelength for specific frequencies (usually in those problematic 250 Hz and 500 Hz octave bands), those specific frequencies will be cancelled to one degree or another at that position. Once we realize that we have summation from all three boxes at position A and cancellation from various boxes at position C, it is quite easy to understand how the doughnut-shaped patterns we examined earlier are produced. Both arrivals in phase 90 out of phase 180 out of phase 180 out of phase, but 2 nd is lower in level Sound from Driver 1 Sound from Driver 2 Resulting summation or cancellation We can demonstrate the summation and cancellation that makes line arrays work by adding two signals together in a DAW. Above, we can see that two signals arriving in phase produce a +6dB summation. If one signal arrives late (90 out of phase), the signals still produce a larger summation. However, if one signal is a bit later (180 out of phase), the signals cancel each other out completely. If one of the 180 signals is reduced in level, as would usually be the case, then the cancellation is less than complete Silly Things Churches Do With Loudspeakers Page 58

59 The trick is to make this effect work for us instead of against us. This is why line arrays composed of vertical lines of drivers are usually a good thing (but not without their own set of problems!), and line arrays composed of horizontal lines of drivers are usually a bad thing (such as the tight packs we have been considering here). If we rotate the 500 Hz doughnut that we examined earlier by placing the drivers above and below each other, the pattern is now more often useful than problematic. The diagram below illustrates the basic geometry that produces the pattern to the right. How a mini line array (a bipole) works This is the underlying principle by which all line arrays achieve their directivity. The length of the array, the number of elements, and the processing applied to each driver determines the frequency range over which a more sophisticated array may be effective in achieving pattern control. However, the concept is vital and quite useful even for point source designs, where we should be looking for every bit of low frequency pattern control we can get Silly Things Churches Do With Loudspeakers Page 59

60 It is important to understand that the extent to which any trapezoidal cabinet is able to control its coverage pattern in the lower octaves determines its suitability to use in a tight packed horizontal array that avoids or reduces the now-dreaded accidental line array effect. Surface loaded drivers are essentially all the same regardless of manufacturer when it comes to pattern control a 12 cone radiates like a 12 cone. On the other hand, some clever co-entrant designs with Danley SH69 relatively large mid to low frequency horns can extend pattern control several octaves lower than surface loaded drivers, and these boxes are the design of choice for tight packed horizontal arrays. Danley Sound Labs is one manufacturer that has done a great job with designs meeting this requirement; Frazier is another. The other option for designers is to put the accidental line array effect to good use by using loudspeakers and array designs that use the effect to great benefit. For example, Renkus-Heinz was an early producer of a type of array that combined HF elements that did not suffer from excess overlap with vertically aligned LF transducers, such as the early (now discontinued) TRAP array (below left). Later, cabinets such as the EAW MQ series used vertically oriented LF drivers, and the Craig Jansen designed the TD415 with EAW which used a pair of bipoles to great effect. Renkus-Heinz s STX series also uses bipole LF drivers Silly Things Churches Do With Loudspeakers Page 60

61 Finally, we would like to show how carefully avoiding building an accidental line array from tight packed trapezoidal cabinets can make a huge difference in the coverage and GBF (gain before feedback) in a building where a high performance system was desired by the church. The left mapping above is a tight packed array of 3 typical MI trapezoidal cabinets in the critical 250Hz octave band (middle C is 262 Hz). The coverage is too narrow and has a hot spot in the front rows. Worse, the so-called array is putting the 250 Hz octave all over the platform (where the open microphones will be). On the other hand, the coverage modeled for single high-output cabinet with 3 vertically oriented 12 drivers covers the seating area like a glove (right). Even better, the mapping reveals a high amount of pattern control, no hot spot, and a huge reduction in the amount of sound being spilled onto the platform from the main speaker. Just say no to homemade tight pack arrays! As we have stressed throughout this entire article, better design always produces better results. By avoiding common pitfalls, a church will save money in the long run and enjoy superior and safe results every week, making it more effective in its mission. That, after all, is what it is all about Silly Things Churches Do With Loudspeakers Page 61