Before the existence of amplification technology, we relied on the capacity of a speaker or singer to project their voice by simply increasing its level through brute force — just straight up making it louder by forcing more air out of their lungs over their vocal cords. We also had building architecture intended to direct and maximize the increased level produced by that person on the stage, but in those days, there was no mechanism for capturing the voice with a transducer, amplifying it, then presenting it through yet another transducer with the intention of making it louder. Public address systems were indeed developed primarily to increase the level — the quantity — of the sound. The means through which the quantity was increased also presented an opportunity to increase the quality of the sound. In fact, in some cases, the increase in quantity creates a situation that requires an increase in quality, since the amplification mechanism can have a negative effect on the quality of the sound. Whether it’s a necessity or just a serendipitous boon, there are ways to increase the intelligibility and clarity of the voice through a number of means, and they are worth pursuing.
Mic Selection
There are several ways to go about increasing intelligibility, and we will list them here in decreasing order of ease-of-implementation, from the most painless, easy solutions to the more complex, challenging ones. Topping the list is microphone choice. One of the important components of intelligibility is consistency. It simply won’t do to have level increasing and decreasing wildly as the result of a handheld microphone moving closer to and further away from the mouth of the speaker or singer. A lav mic is preferable over a handheld by way of minimizing these fluctuations, but a headset mic is superior, since it maintains a more consistent mouth-to-microphone distance (which can fluctuate with a lav mic due to the occasional turning of the head).
Headset mics come in two varieties — cardioid and omnidirectional — and comparative inconsistency can result from the cardioid variety. Omni headset mics are much more forgiving and will present a more consistent signal in terms of level than their directional counterparts. That’s generally true, whether the mic is headset, handheld, or podium-mounted (although out of the entire bunch, handheld omni mics are probably going to be the most susceptible to feedback). Headset omni mics are almost certainly going to be the best bet to accomplish consistent level.
Mic Technique
The proper use of the mic also figures in by way of determining consistency. Of course, in the case of handheld mics, we beg pastors and singers alike to hold that mic close to their mouths, a bit off-axis, and please try to avoid pointing it toward speakers of any kind. It seems like this kind of pleading falls on deaf ears, but if we put ourselves in their shoes for a moment, it’s not hard to sort out that in the throes of passionate performance, when the Spirit is moving, the last thing on their mind is going to be how the sound guy said to hold the mic. But we still have to try. So we do. And we try to gently guide our spoken word pastors toward headset mics. It’s important for even headsets to be properly deployed. A mic flopping around 6-10 inches away from the pastor’s mouth is almost certainly going to be even worse than an inconsistently positioned handheld or lav. So take care to ensure that the headset mic element is firmly located at the corner of the pastor’s mouth, and placed in a way that will drastically reduce the likelihood it will move from that location.
Signal Processing
Even if we get the microphone part of the equation right, there are other things we can do to maximize intelligibility. Signal processing can be very helpful. It’s important to also note that signal processing can be harmful to intelligibility, so let’s sort out what we should or should not do. Let’s talk first about the things we should avoid. One very fundamental factor determining clarity is the ratio between direct signal and reflections (or reverberation). We want the direct signal high and the reflections low. Adding reverb to the pastor’s voice might create an illusion of a bigger space, and this can be a temptation for live streaming, particularly if there’s no video or if the video doesn’t reveal the true size of the sanctuary — “wow — listen to that — this pastor must be preaching to a huge congregation!” Not only is this a false representation in an attempt to increase the credibility of the pastor and church, it decreases intelligibility. Save the artificial reverb for the singers, and even in that case, it’s a good idea to reduce the reverb returns whenever a singer prays or reads scripture or otherwise speaks (as distinct from singing). Reverbs, delays, and various forms of modulation all reduce clarity, so play it straight and avoid this type of thing.
So what kind of signal processing will increase intelligibility? Filtering and equalization are almost certainly the best place to start. It’s a good idea to limit the signal to only that which is important, hence focusing the listener on the specific thing we want them to hear. The male human voice can sometimes contain energy below 80 Hz, for example, but there’s nothing down in that low end that truly contributes to intelligibility. There are, however, distractions down in that frequency range — rumble, thumps, bumps, and other stuff that’s not helpful. I strongly recommend filtering it all out with a high-pass filter, and if you have access to digital (read: non-resonant) filters, don’t be bashful with a steep slope either. Shaving a little off the top won’t hurt much either (and may help tamp down feedback too). Most adults aren’t hearing anywhere near 20k Hz anyway, so don’t fear a little LPF.
Invariably, we’re equalizing to avoid feedback to begin with, so do what you need to do there, but if there’s any way you can avoid digging in too deeply in the 1k Hz – 4k Hz domain, that would be good. That range contains information that is important to intelligibility. Of course, there can be some harsh nastiness in that area that needs to be controlled, but generally speaking, the sibilants and transient components of fricatives that help us to understand speech live in that area, and we’re best to try not to cut it much if we don’t have to. And indeed, we’re almost always better off cutting than boosting, although a small boost here and there is fine.
Another signal process that can be helpful to us is compression. As we said in reference to mic choice and deployment, consistency is important, and reduction of dynamic range can help to bolster it. When we effectively make the quiet syllables louder and the loud syllables quieter, it smooths out the overall delivery, and that consistency definitely contributes to intelligibility.
The Room
One last item: room acoustics — indeed, the most challenging of potential solutions we list here. We always want to keep reflections under some modicum of control in our space, and this will always help increase intelligibility by way of increasing the direct signal to reflections ratio. By selecting and using the right mic the right way, processing the signal appropriately, and keeping a handle on the room acoustics (or at least looking at how your speakers are aimed), we can achieve clean, intelligible speech in the sanctuary. It can be done!
John McJunkin is the chief engineer and staff producer in the studio at Grand Canyon University.
