Why Your Choir Sounds Like a Synth Pad

Everyone reaches the same point with sampled choirs. You load the patch, you play a chord, you hold it, and what comes back is beautiful and completely inert.

The library is almost never the problem. What you played is a pad, because you gave it none of the things that make a choir a choir.

A Choir Is Movement, Not Tone

Take a real choir sustaining one chord and look at what is actually happening.

The vowel is drifting slightly as the singers adjust. Individual people are breathing at different moments, so the ensemble thins and refills continuously. Consonants arrive and cut the sound. Intonation moves as the harmony moves under it.

None of that is in a held Ah on one MIDI note. What you have instead is a smooth, static, filtered texture, which is a fair technical description of a synth pad.

So the fix is not a better library or more reverb. It is putting the movement back, and the four sections below are the four kinds of movement that matter.

Change the Vowel During the Note

This is the biggest single improvement and most libraries support it with a controller or a keyswitch.

Instead of holding Ah for four bars, move it. Ah into Oo across a two bar swell reads as a choir shaping a phrase. Oo into Ah on a crescendo reads as the sound opening up, which is what actually happens when singers push.

Where the library puts vowels on keyswitches rather than a continuous control, the change is a hard switch and you need to place it on a note boundary to hide the seam. Where it offers a morph, ride it slowly, over a bar or more.

The annoyance is that vowel controls are wildly inconsistent between libraries. Some use CC1, some a mod matrix, some a separate patch per vowel, and the ones with separate patches force you into the two track approach below whether you wanted it or not.

Playing connects you to YouTube, which may set cookies.

Put Consonants on the Accents

A choir entering on a pure vowel fades in. A choir entering on a consonant lands.

That difference is most of what makes a choir sound rhythmic rather than atmospheric, and it costs one keyswitch. Where the library has consonant samples, put one on every accented entry and on the front of any note that needs to articulate a rhythm.

Where it does not, you can fake a surprising amount with a very short percussive layer under the entry, or by shaping the attack of the vowel with a fast volume ramp rather than letting the sample's own slow attack run.

Do not put a consonant on everything. A line where every note articulates sounds like a chant, which is a real effect and rarely the one you wanted.

Write the Breaths In

Singers stop. This is not optional for them and it should not be optional in your writing.

Put a gap every four to eight bars, sized to the tempo, and place it where a phrase ends rather than mechanically. Shape the note before it down rather than cutting it, because a singer running out of breath tapers rather than stopping dead.

The gap does not have to be silent across the whole choir. Staggered breathing, where the section splits so that somebody is always singing, is real technique and it is how sustained choral writing works. In samples you reproduce it by putting the choir on two tracks and breathing them at different bars.

That two track split solves several problems at once, which is why it is worth the RAM.

Split the Choir, Do Not Stack It

The default instinct is to load one patch and play a four note chord. That gives you four copies of the same recording of the same people singing the same vowel, which is dense and lifeless.

Split it instead. Sopranos and altos on one track with one vowel, tenors and basses on another with a slightly different one, breathing at different points, at slightly different levels.

Two vowels moving against each other produce beating and interference that a single patch physically cannot, and that interference is a large part of what the ear reads as many people rather than one sound.

The friction is CPU, and it is real: choir libraries are heavy, and running two instances where you had one on a large template is a noticeable cost. Do it on the cues where the choir carries something, and use the single patch for background texture.

And Use Less Reverb Than You Want To

Last, because it undoes everything above if you get it wrong.

Choir libraries are almost always recorded in a large space already. Adding a long hall on top produces the wash that people associate with epic choir, and that wash smears exactly the consonants and vowel changes you just spent an evening putting in.

Use a shorter send, check with the consonants soloed, and trust that the recorded space is doing more work than it appears to.

If the cue still wants that vast sound, get it from the arrangement and the low end rather than from a four second tail on the voices.