Articles - Singing together

home - Articles

recording studio

links

Singing together

A tutorial on faking multiple voices.

(works in progress - needs audio files and effects screenshots)

The voice is often the most important part in a song. It deserves the right amount of care for a nice sound.

But often a good take with a nice sound isn't enough. It lacks "dimension", "width", "presence". Or it needs that special "commercial" quality.

Another song may need just more voices singing. Doubling (like in many Pop songs), a choir, unison, harmony.

These are some common (or less-common) effects we can try when another singer (or another take) isn't available. From doubling to a whole stadium. Or when we are looking for special effects.

One voice

Voice is most often recorded through a single microphone. It is a mono signal.

When the song is mixed, the voice may sound too small or too focused at center. This is because a dry mono source hasn't any width in the stereo field and we perceive it as perfectly focused between the speakers.

The goal is to add some "movement" and "width" to the voice before applying reverbs and other effects.

Here are some tests with several kinds of effects.

Choir, Chorus

Audio file - two copies of the same track, slightly time-shifted

Audio file - two copies of the same track, slightly detuned

Audio file - two copies of the same track, slightly detuned (with mod)

Audio file - stereo chorus

Two slightly delayed copies

Stereo chorusing

Periodic time-shifting (gives periodic pitch-shifting)

Doubling, widening

Audio file - one voice from a pitch-shifting effect

Audio file - the same, random pitch

Audio file - the same, random pitch and time

Real doubling

Singing it again or

making another singer play the same part.

Faking the second voice requires:

random pitch variations plus random time variations.

The crowd

Sometimes some "special effects" are needed.

The "live event" sensation needs noises, reverb, echoes and the crowd. And, for a good impact, the crowd must sing with the singer, of course.

This is a trick to get a crowd singing together with the singer.

The "crowd" sound:

- has an overall timbre that is an averaged spectrum from several hundreds persons (averaged formants)

- cannot follow song tempo exactly

- cannot raise and fade out exactly and abruptly

- isn't crisp and clear like a single voice

- doesn't have clear, clean waveform

The concept of this trick is taking random waveform (hundreds of voices and phases) through an averaged spectrum (the "formants").

This is the final result:

Audio file - Singing stadium

(Black Roses - "The Viking")

Dynamics

Equalizers

Reverb

Mastering

Effects

Synths

Drums

Development

Tolcbocs - free!

Plate reverb

recording studio

collaboration

custom development

licensing

disclaimer