People reach for a voice changer for three different reasons, and the reason decides which one to use. Comedy wants an obvious effect. Anonymity wants an effect that survives scrutiny. Narration wants an effect nobody notices at all.

The four families

Sixteen effects sounds like a lot of choices. It is four decisions.

Pitch. Chipmunk, Deep, Helium, Monster, Ogre. These shift the voice up or down. They are the most immediately recognisable and the least convincing as disguise, because pitch shifting leaves the rhythm and the accent exactly where they were.

An audio waveform shown before and after silences are closed up, the take shorter.

Character. Robot, Alien, Cartoon, Duck, Ghost. These do something structural to the tone rather than simply moving it. They read as a performance, which is exactly what you want in comedy and exactly what you do not want in narration.

Space. Echo, Reverb. These place the voice somewhere: a hall, a corridor, a large empty room. Used lightly they are the only family that can improve a dry recording rather than transform it.

Lo-fi. Megaphone, Telephone, Underwater, 8-Bit. These simulate a channel the voice came through. Telephone in particular is the classic way to mark a voice as remote without a word of explanation.

For the underlying concepts, Apple's Final Cut Pro user guide is a clear primer on levels, headroom, and why a mix that peaks correctly can still sound wrong.

Applying one

  1. Open the clip in Zella and go to the Audio controls.
  2. Pick a voice changer.
  3. Play it back before you commit to it. Every one of these sounds different on a voice than it does in your head.

It applies to the clip's voice, so different clips can carry different treatments - which is how you do a two-character sketch with one person and one microphone.

Clean the voice first

The order that matters here is the same one that matters everywhere in audio. Remove the room noise before you apply a changer, not after.

A voice changer is a transformation of whatever signal it is given, and it transforms the hiss with equal enthusiasm. Reverb on a noisy recording gives you reverberant noise. Megaphone on a hissy take gives you a hissy megaphone. Getting the voice clean first makes every effect afterwards sound deliberate instead of cheap.

What survives a phone speaker

Most people will hear your video through a phone speaker with no low end and a great deal of compression, often in a noisy room. That changes which effects work.

Survives well: Telephone, Megaphone, Robot, Chipmunk. These live in the mid range, which is the only part a phone speaker reproduces properly.

Survives badly: Deep, Monster, Ogre. Everything that makes them distinctive is in the bass, and a phone speaker does not have any. What the viewer hears is a muddy voice rather than a deep one.

Depends entirely on restraint: Echo, Reverb. A little is atmosphere. A lot is unintelligible, and unintelligible is worse on a phone in a cafe than it is in your headphones.

If a video is going to a vertical feed, check it on an actual phone at actual volume before you publish. It is the single most useful five seconds in an audio workflow.

Voice changers are not anonymity

Worth saying plainly, because it comes up. A pitch shift does not make a voice unidentifiable. Cadence, accent, vocabulary, breathing and timing all survive it, and those carry more identity than pitch does. Somebody who knows you will recognise a Chipmunk-shifted version of you.

The effects that genuinely obscure a voice are the heavy ones - Robot, 8-Bit, Alien - and they obscure it by destroying most of the signal, which means they also destroy most of the intelligibility. That is a real trade, not a setting.

Where they actually earn their place

  • Telephone on a quoted line. Reading out a message or an email in a Telephone voice tells the viewer it is not you speaking, with no cutaway needed.
  • A second character in a sketch. One person, one mic, two clips, two treatments.
  • Reverb under a title card. A single word with a touch of space behind it reads as a chapter marker.
  • Megaphone on an announcement. A tiny bit, on three words, for emphasis.

Where they do not: across a whole talking-head video. Nobody watches four minutes of a processed voice, no matter how good the effect is.

Frequently asked questions

How many voice changers does Zella have? Sixteen, and the same sixteen on Mac and iPhone: Chipmunk, Deep, Robot, Echo, Reverb, Megaphone, Telephone, Underwater, Monster, Helium, Alien, Ghost, 8-Bit, Duck, Cartoon and Ogre.

Will a voice changer make me anonymous? Not reliably. Pitch shifting leaves cadence, accent and breathing intact, and those identify a speaker more than pitch does. Only the heavy effects obscure a voice, and they cost intelligibility to do it.

Should I clean the audio before or after applying one? Before. A voice changer transforms the noise along with the voice, so remove the room first and the effect will sound deliberate rather than cheap.

Which one works best on a phone speaker? Anything that lives in the mid range: Telephone, Megaphone, Robot, Chipmunk. Deep, Monster and Ogre depend on bass a phone speaker cannot produce.

Can different clips have different voices? Yes. The effect applies per clip, which is how one person and one microphone make a two-character scene.

Do these run in the cloud? No. The audio processing is on-device, on both apps.