Do we all have to be online at the same time?

No. Each participant records their part whenever they want. That is what makes group dubbing realistic between people who are neither in the same place nor free at the same hours.

What happens if someone drops out?

Their role stays open and can be taken by someone else, or by the host. Recordings already made are not lost.

How many people can take part?

As many as the scene has speaking characters. A two character scene is dubbed by two people, a four character scene by four.

Together

Dub a scene with friends, without ever meeting up

Q: How do you avoid two voices overlapping?

Each participant only records their own character's lines, at the moments already identified in the transcription. The split comes from the scene itself, so there is nothing to coordinate.

A three character scene dubbed by one person always sounds like one person. With several, everyone brings their own voice, timbre and acting, and the result changes nature. All without anyone having to be free at the same moment.

The real obstacle, and how it falls

Dubbing together used to mean getting everyone in the same room, or at least in the same call. That is where it fails: finding a slot that suits three people is harder than dubbing the scene itself.

The answer is to split the work. The scene is analysed once, the lines are divided between characters, and each participant receives only their own. They record whenever they want, at home, hearing the other characters in their original version to play off. When everyone is done, the voices are assembled.

Nobody waits for anybody. That is the difference between a project that ships and one that dies in a group chat.

How to do it

Pick a scene that suits it

Confrontations and fast exchanges work best: two or three characters answering each other. Avoid scenes where one character monologues while the others listen, the point of being several disappears.

Declare the exact number of characters

The step everyone rushes, and the one that ruins most sessions. Automatic detection regularly merges two similar voices, and you end up with three roles instead of four.

Name the roles before sharing

"Speaker A" means nothing to someone who has not seen the scene yet. Use the real character names.

Send everyone their own link

Each role has its own link. The participant opens it, sees the scene and a rythmo band with their lines highlighted, and records.

Assemble

Once every role is recorded, the voices are mixed together over the music and effects track.

Count the people who SPEAK, not the ones on screen. And give the exact number: overstating is as harmful as understating. Ask a model for five voices in a scene that has three and it will obey, by cutting a real character in two.

The traps

What goes wrong

Volume gaps: three people, three microphones, three rooms
An echoing room next to a furnished one, audible instantly
Someone who disappears halfway through
Roles named "Speaker A" that nobody can match to a character

What to ask everyone

Headphones, always, or the original voice leaks into the take
A constant distance from the mic, around twenty centimetres
A furnished room, never a bathroom
Record the whole role in one sitting, the voice stays consistent

If someone drops out, their role simply stays open. What the others recorded is kept, and anyone can pick it up.

Open a group session

Frequently asked

Do we all have to be online at once?

No, and that is the whole point. Everyone records whenever they want.

What if someone drops out?

Their role becomes available again. Existing recordings are kept.

How do you avoid two voices overlapping?

The split comes from the scene itself: each person records only their character's lines, at the moments they are spoken.

How many people maximum?

As many as the scene has speaking characters.

Can one person dub a group scene alone?

Yes, and it is a common use. See the dedicated guide.

Sources

On the practice: fandub.

Written by the Dubba team · Published 5 August 2026