band PA: how to get good intelligibility when a singer talks while music is playing?

Question

I'm running the PA system in a church. During worship, the lead singer will sometimes say a few words while the band is already playing a song intro. These words are invariably difficult to understand. When this happens, I try to improve intelligibility:

• increase the volume of the lead singer as much as necessary
• decrease the volume of the band by 10 dB (more than that is useless because the band volume via the PA will drop below the band volume via their monitors, and I can't quickly change monitor level)
• switch off the reverb effect on the lead singer's microphone

This helps a bit, but not quite enough to my taste. I'm considering routing the lead singer's microphone through 2 channels on the mixer, one set up for singing and the other for speech, so I can use different EQ settings. Are there other things I can do to improve intelligibility?

The system is capable of delivering good speech intelligibility (STI=0.65). It's just speech over music I'm having trouble with.

Answer 1

Knock a 3-6dB bandpass low-Q filter in the master bus for the band around the speech frequencies then use a seperate bus for the vocals that pushes a simlar bandpass filter with a little bit of gain. Then balance to taste.

Your idea of splitting the vocal channel between singing and and speech is a good approach. This will allow seperate control of reverb and eq.

Answer 2

The very first thing I would do in your case, is to put a -18dB/oct highpass filter, tuned at 200Hz or a little more, in the singer mic. it would depend on your mixer, some of them have a fixed 80Hz or 100Hz HPF (which I consider too low frequency for the HPF), other allow you to vary the cutoff frequency. if your mixer has a fixed cutoff frequency, I would search for a external filter, like a parametric equalizer. bass response in vocals mics, could affect the intelligibility terribly. when I'm mixing, I cut off almost EVERYTHING at 200Hz, except obviously the low frequency instruments like kicks or lines like keyboards and electric bass.

Obviously all singers have more acoustic output when they are singing, than when they are talking, for compensating that, you can do one of the two following, or even both: 1) request this singer to put his or her mouth closer to the microphone when he or she is talking, but beware of the proximity effect that ALL directional mics have, the bass response is incremented much when the mouth gets closer to the mic. Again watch out the HPF, very important 2) compensate adding a little gain in this talking passages (if you know previously when they came in the songs), if this case, to have two preconfigured channels is a good idea. but beware of feedbacks.

I would not recommend to use a compressor to equalize the dynamic range, because it could fire up the microphone feedbacks, and they could add much stage residual noises when the singer is not singing or talking. moreover I would recommend to use a noise gate, but the threshold should be very low because it could filter out the talking passages. every channel (vocals and instruments) should have its noise gate with their threshold carefully configured.

as for the reverb, you could use different kinds of reverb, in the singing passages use a long reverb, and in the talking passages, use a short a subtle reverb.

Generally churches have very bad acoustic response, so I would try to directionate the PA speakers directly to the audience as much as possible.

And of course, be sure your PA have good acoustic response, in terms of frequency flatness and in terms of distortion. keep as flat and natural as you can, for both mikes and PA. don't abuse of coloring and/or graphic eq's. correct only a few needed bands, and nothing more.

Consider trying different mics, some has better mid-presence than others. I like very much the german BeyerDynamic mics.