As church sound engineers, we often get wrapped up with mixing music and los e sight of the most important part of a worship service; speech. Speech intelligibility is of the utmost importance, and we have to do everything in our power to boost intelligibility for every member of the congregation.
The Room
Good intelligibility starts with the room and actually has little to do with the sound system itself. If you are in a highly reverberant space like a gymnasium, you are already fighting an uphill battle toward good intelligibility. I’ve done worship services in everything from a school cafeteria to a gymnasium to a highly controlled space with a sub-1 second decay time, so I have lived these challenges for many years. I’ve recently been involved with a live event that takes place in a soundstage used for shooting films. The room is truly a black box with almost no acoustic treatment, an absolute nightmare for amplified sound of any type, especially speech. For this event, we had a scenic company come in and wrap the entire room in velour drape. This drape is relatively inexpensive but it makes a world of difference for the room’s acoustics, and therefore the room’s intelligibility. Everyone would love to hire an acoustician to recommend acoustic treatments for their worship space, but the reality is that simple fixes like drape are often as acoustically effective as they are cost effective.
The System
This is a topic that is near and dear to my heart, and I’ve talked about it in many articles. One of the most important aspects of good intelligibility is choosing the right speaker system for a given room. I feel that this is criminally overlooked, and people often choose a system because it is popular or they heard it work well elsewhere. This is related to what I call “line array syndrome,” which is thinking that line arrays are the solution to all audio problems. Poor acoustics? Line array. Poor intelligibility? You probably need a line array. Cure for world hunger? Line array!
In fact, not all speakers are suitable for any given room, and line arrays aren’t suitable for a large percentage of churches. The main goal of any sound system is to put direct sound onto the listeners and keep sound off of reflective surfaces such as walls and ceilings. Certainly this goes hand in hand with acoustic treatment, but simply choosing speakers with an appropriate coverage pattern also makes a world of difference. Reflections from hard surfaces make their way back into your listening area, causing comb filtering and reverberation that will color the sound and hurt intelligibility. These sorts of reflections cause variance from seat to seat, making the response of the system unpredictable and inconsistent for the audience.
The Mix
I’ve had some conversations recently about how sensitive I am to speech mixed through a P.A. Mixing speech is an art form, and after so many years of focusing on mixing spoken word, I am very attuned to how different people handle it. The topic came up after I heard someone mixing spoken word entirely too loud. It wasn’t an intelligibility issue per se, but it sure made spoken word difficult to listen to for a long period of time. Anyone who does house of worship or corporate gigs can sympathize with me. If speech is too loud, it’s just painful and harsh to listen to. If speech is mixed too quietly, it can be unintelligible and listeners have to focus way too much brainpower on deciphering what is being said. From a mix perspective, there are several things I focus on when mixing spoken word for best intelligibility…
Microphone Selection: I always favor omnidirectional headset mics for speech. They are more forgiving in terms of their placement, so if the capsule moves for any reason, it’s still a workable sound. However, if you mix in a highly reverberant space, the omni can actually be your worst enemy for speech intelligibility. In this case, try moving to a cardioid headset, but make sure it’s positioned properly every service, or you may get poor results.
Microphone Placement: I’ve seen this time and again. It’s not always something we can control, but as much as possible, try to place and check your pastor’s headset or lav before each service. Once a pastor has a mic on, they will do things like put on a robe, hug someone, put on a jacket, or do something to otherwise move their mic placement. We’ve all experienced the headset miking the ear, where the capsule slides back on the side of the face, hurting intelligibility.
Level: As mentioned above, simply getting the level of spoken word right goes a long way toward good intelligibility.
Equalization: The way that spoken word or vocal performance is equalized has a huge impact on intelligibility. First, start with the all-mighty high-pass filter to clean up that low-end. Don’t get embroiled in the numbers — set it based on what sounds right. I typically start with a high pass at 125 Hz on any vocal, but I’ve taken them as high as 200 Hz in certain situations where a vocal was really boomy. Next, clean up any low-mid muddiness with a parametric filter. I almost always scoop out a little low midrange between 200 and 300 Hz. Finally, if a vocal is harsh sounding, you can address that with a parametric EQ centered around 2k Hz or 3k Hz.
Dynamics: For spoken word, dynamics are really challenging. If left unchecked, vocal dynamics for speech or singing can be very problematic. A regular broadband compressor goes a long way toward managing this. However, if speech or vocal performance is over-compressed, you lose all of the excitement. When dealing with speech, I spend a lot of time tweaking the compression to get it just right. Many of the previous topics such as mic placement, level and EQ will have an impact on your dynamics processing.
Finally, always make sure to get out and walk the room. Don’t be tethered to your console. This will give you a better idea of what the congregation is hearing, and it will help you make level and EQ choices that best suit the majority of your listeners.
Photo Caption: Placed in close, a headworn mic — such as this Countryman E6 — offers significantly improved intelligibility over lavalier models.
