Which kind of desk do you mix on?
Some steps in this lesson are different on an analogue desk. Choose yours, and the lesson shows the steps for it.
The pastor reaches the big point of the sermon and suddenly shouts. A compressor is an automatic hand on the fader: when a sound gets louder than a level you choose, it turns it down. It catches the loud moments, so you can set speech where you want it and trust it. It can't lift a whisper: that's still your fader.
Swipe the drawing sideways to see all of it.
Above the threshold line, the compressor turns the sound down; below it, nothing changes.
The controls
- Threshold: the level where the compressor starts working. Lower means more of the sound is compressed.
- Ratio: how firmly it turns down above the threshold. At 4:1, a word that goes 8 dB past the threshold comes out only 2 dB past it.
- Attack and release: how quickly it reacts, and how quickly it lets go.
- Make-up gain: turns the whole signal back up after compression.
- Gain reduction meter: how much it's turning down right now. Watch it as closely as you listen.
Swipe the drawing sideways to see all of it.
Follow a level going in along the bottom, up to the line and across to what comes out: below the threshold the two match, and above it, the flatter the line, the firmer the compression.
How much is enough?
Speech starting point from the compression cheat sheet: ratio 3:1, attack 10 to 20 ms, release 100 to 200 ms. Lower the threshold until the loudest words show about 3 to 6 dB of gain reduction, and check the meter rests between phrases.
The threshold number depends on your gain; set it by the meter. If the meter never rests, you've gone from control to squashing, and the voice starts to sound flat and tiring.
On an analogue desk
Many analogue desks have a one-knob compressor on some channels. Turn it up while the pastor speaks at full level, until its light (if it has one) flickers on the loudest words, not on every word. The knob adds level too, so check the mic isn't closer to feedback. If only some channels have one, plug the pastor into one of those. No compressor at all? Ride the pastor's fader (lesson 4), or ask your leader about a compressor plugged into the channel's insert.
Try it: tame a sermon
In the sermon below, the computer sets each phrase's level: normal speech, a stronger point, a much louder big moment (the shout), then a much quieter one (the whisper), over and over.
How to use this
Practise: setting gentle compression on a sermon by ear and by the gain reduction meter.
All six steps, and what to listen for
- Press Play (headphones, low volume) with the compressor out. Which moment jumps out, and which one gets lost?
- Press "Cheat sheet settings" (3:1, attack 15 ms, release 150 ms). Lower the threshold until the shout shows 3 to 6 dB on the gain reduction meter and the meter rests at 0 between phrases.
- Press "Squash it" (10:1, threshold low). Listen to the voice go flat, and watch the meter never rest. Notice the whisper came up too, with the breaths and the room: the make-up gain turned the whole squashed voice back up. Then put your setting back.
- Try a very short release, then a very long one, and read what each does on a desk. You'll hear little difference here, but with the long one the meter lets go slowly after the shout. Put it back to 150 ms.
- On your gentle setting, normal speech is left alone, so it needs no make-up gain. Turn the make-up gain up to +3 dB: every part comes up 3 dB, the whisper too, and the gaps between them stay the same. Put it back to 0. Then switch the compressor in and out. It's level-matched, so judge the sound, not the loudness.
- Last round, the whisper, three ways. With your gentle setting, watch it go through untouched. Press "Squash it" while the whisper plays and listen: it comes up, but so do the breaths and the room, and the voice goes flat. Then press "Back to my setting" and ride the fader up for the whisper, and back down before the next words.
Listen for: the shout coming down a few dB while normal speech sounds untouched. It still stands out, which is why your hand stays near the fader.
Done when: the widget says "Gentle and steady" for your setting, you've heard "Squashed", read "Pumping" and "Too slow", heard make-up gain turn every part up alike, compared it in and out, heard Squash it bring the whisper up with the breaths and the room, and ridden the whisper.
Sunday: the compressor catches the shouts; your hand on the fader still looks after the whispers.
Note: This compressor smooths its own release, so a very short or a very long release shows less here than on a desk. The settings are the cheat sheet's starting point.
This interactive needs JavaScript.
Without JavaScript: An example from this practice tool. The shout peaks at –12 dBFS, normal speech at –24 and the whisper at –38; to the ear, the whisper sits about 15 dB under normal speech. At 3:1, attack 15 ms and release 150 ms, a threshold of –20 dBFS shows about 4 dB of gain reduction on the shout and none on normal speech, and the meter rests between phrases: gentle and steady. The whisper goes through untouched, and +3 dB of make-up gain turns every part up 3 dB, so it's still about 15 dB under normal speech. At 10:1 with the threshold at –40, normal speech's loudest words are turned down about 10 dB and the meter never rests: squashed. Its +12 dB of make-up brings the whisper up about 12 dB, but normal speech as a whole comes up about 6 dB, and the breaths and the room between phrases about 9 to 10 dB, so the whisper is still about 9 to 10 dB under normal speech. Only your fader brings the whisper up on its own.
Check yourself 1 of 3
During the sermon, the gain reduction meter on the pastor's mic sits at 8 to 10 dB and never drops back, even between sentences. What do you do?
A meter that never rests means you've gone from control to squashing, and the voice sounds flat and tiring. Raising the threshold lets normal speech through untouched. Lowering it squashes even more, and a compressor never brings quiet words up: that's your fader. At 3:1, a word 9 dB over the threshold comes out 3 dB over, a 6 dB turn-down: about as hard as speech should be pushed. Aim for 3 to 6 dB on the loudest words only.
Gates on speech: gently, or not at all
A gate turns a channel down when the sound drops below its threshold. On drums it's common. On a speech mic it's a choice, not a rule: it can help a lectern mic that hears the room between sentences, or several open speech mics. Set badly, it chops the starts and ends of words, and the room pumps in and out with every phrase.
If you use one on speech, set it gently: a range (how far it turns down) of about 10 dB rather than full off, the threshold below the quietest word, a hold (how long it stays open after a word) of a few hundred milliseconds, and a slow release (how gently it closes after a word ends). To find the threshold, start with the range at 10 dB. With the speaker silent, raise the threshold until the room no longer opens the gate; then ask them to read softly, and lower it until every quiet word opens it. If your desk has an expander mode, use it: it turns the sound down gradually instead of shutting it. For four or more speech mics, use the desk's automixer if it has one: it turns up whichever mic is being spoken into and turns the others down. One close headset in a quiet room usually needs none. Often simpler: keep mics that aren't in use muted until they're needed.
On an analogue desk
Analogue desks have no gates. Keep the lectern mic muted while nobody is at the lectern, and get it close to the reader.
How to use this
Practise: setting a gate on a lectern mic gently, between the room and the quietest word.
All three steps, and what to listen for
- Press Play, then "Speech starting point" (range 10 dB, attack 2 ms, hold 300 ms, release 500 ms).
- Find the threshold between its two limits: raise it until the room alone no longer opens the gate, then check every quiet word, the whisper too, still opens it.
- Press "Drum settings" and listen to the words chop and the room pump. Then press "Back to my setting".
Listen for: the room coming down a little between sentences, while every word comes through whole.
Done when: the widget says "Gentle", and you've heard "Words chopped".
Sunday: one close headset in a quiet room needs no gate; on a lectern or a panel, set it gently or not at all.
Note: The reader is the same voice as in the compressor above, with longer pauses between sentences.
This interactive needs JavaScript.
Without JavaScript: An example from this practice tool. Between sentences the room peaks at about –54 dBFS, and the whisper at –38. With the range at 10 dB, attack 2 ms, hold 300 ms and release 500 ms, a threshold from just above the room, about –53, to –42 dBFS opens on every word and turns the room down about 10 dB between sentences: gentle. Any lower, and the room's own hiss holds the gate open. Above –42, the whisper's first sound is cut, and from –38 the whisper is lost. With a deep range, no hold and a fast release, the words are chopped and the room pumps in and out.
Check yourself 2 of 3
A gate on the lectern mic keeps the room down between sentences, but now the ends of the pastor's quieter words get cut off. What's the best fix?
A gate on speech only works when it's gentle: it must open for the quietest word and close slowly, so endings fade instead of being cut. On a gate, a higher threshold means more words fall below it and get cut: the opposite of a compressor, where raising it does less. A faster release doesn't make it reopen sooner; it closes sooner after each word, so endings vanish. Lower the threshold, slow the release and keep the range at about 10 dB, so a word it misses is quieter, not gone. Still odd? Switch it off and keep unused mics muted instead.
Keep speech dry
Reverb and delay make singing sound full and musical, but they blur words, and in church the words are the point. Speech gets no reverb and no delay, ever. Speech mics (for the pastor, the lectern, readers and presenters) send nothing to the effects. A worship leader's or singer's mic keeps its reverb for the songs, so that reverb has to come off whenever they speak: the welcome, a prayer, a reading, talking between songs.
Swipe the drawing sideways to see all of it.
Each singer's mic sends a little to one shared reverb, and the reverb comes back on its own return, so the move is on the return, not on every channel.
A DCA is one fader that turns a set of channels up or down together; the effects DCA moves all the effect returns at once (Level 2 sets it up). Bring the effects DCA down (on a desk without DCAs, the effect return faders) between songs, during prayers and whenever someone speaks, so spoken words are dry and clear. Ease it down over a second, so the last chord's echo fades rather than stops. Push it back up on the count-in of the next song. Use the DCA, not the mute: a mute can't be eased, and on most desks it also takes the reverb out of the singers' in-ears.
On an analogue desk
On an analogue desk the effects come back on one fader or knob, often labelled FX RTN or FX: wherever this lesson says the effects DCA, use that. Ease it down whenever someone speaks, and up on the next song. Only a knob for it? Turn it down for speech, or use the desk's effects mute or footswitch if it has one. Keep the FX knob on every speech mic at zero.
Check yourself 3 of 3
Between songs, the worship leader talks to the congregation on the vocal mic that has reverb on it for the songs. What do you do?
Speech gets no reverb and no delay, ever, and the leader's mic still has the songs' reverb on it. One DCA move takes every effect down, eased so the last chord's echo fades, and brings it back for the next song. A shorter reverb still puts reverb on the words. A mute can't be eased, so the last chord's echo stops dead, and on most desks it also cuts the reverb in the singers' in-ears.