Suppose you’re shooting a movie. The actors do their work, and you also record their voices while they’re acting. Additional audio stuff like music and sound effects is added later.
But now suppose you want to dub the movie in another language. You can’t just slap music and stuff onto the project, but you (somehow) need to remove the voices of the original actors and then slap those of the new voice actors onto the film. Except if you cut that out, you’d also have to cut out all environmental noise, etc. And if you do that, you’d basically have to recreate every single sound required.
So how exactly does this work? Are movies shot with and without sound simultaneously? Or is there some technological means to separate the sound from the image?
In: Technology
I work in audio post production. In almost all movies/TV shows/Documentaries/etc, pretty much every sound you hear that isn’t dialogue has been added in post. All the atmosphere sounds, footsteps, any time someone puts a cup down or opens a door. Gunshots, explosions. Everything.
Sometimes even the dialogue is replaced for various reasons (audio quality/lines changed/etc), this is recorded in a studio and is called ADR (automated dialogue replacement).
Having all these things separate means that the person doing the final mix has much greater control over all the elements and can create a richer soundscape overall.
As well as a full mix, one of their deliverables is what’s called an M&E stem. This stands for music and effects, and contains a mix of everything except dialogue. This is what is sent to other studios worldwide to be dubbed into other languages.
There are sounds that were recorded in camera, sounds added in post like explosions, and sounds that are mimicking what was recorded on set.
There is a whole process where post production sound crews mimic all the sounds that were filmed on set to be used as a bed for the dubbed dialogue to play on top of.
When dubbing you use all the sounds added in post like music and effects, you add in your new dialogue, and you add in your mimicked sounds.
These mimicked sounds are commonly referred to as “MnE”
Every piece of sound is on a different audio track: Voices, background, sound effects. So each of those sounds can be turned on and off separately.
When mixing the sound for a movie, the dialogue, music, sound fx, and Foley are all on their own tracks. When a final mix is created, they also export what are known as “stems”.
These stems are the final mix of every individual sound, exported into the aforementioned layers (dia, sfx, Foley, music).
When the movies or shows go to another country, the dialogue track is removed and they have the sfx, music, and Foley. They can then add the voices in the languages of choice.
* I’m a sound utility in Atlanta, and I went to school for sound engineering for film at SCAD
Music, Dialog and Effects are three separate sets of audio reels prepared by sound effects editors. There is a re-recording mixer for each of those three classes of audio sources. A music editor, a dialog editor and an effects editor. Original dialog recorded on set by a mixer and a boom operator are part of the dialog tracks. Those tracks are enhanced by looping: rerecording the original dialog for better sound – frequently ambient noise removal. Airplanes, barking dogs, loud traffic, etc. Sound editors and dialog editors are different folks in different buildings. Changing the sound &/or the effects has no effect on the other. They are mixed together, later by the re-recording mixers. That makes the problems you imagine go away.
(Brand new here. I posted this to a commenter not to you. Noob error.)
As people have said, the recordings are multi-track so you can isolate ambient from music from effects from dialog.
If you don’t have multi-track, you can often isolate the vocal track out by being clever. There are tools that work via essentially using noise cancellation to remove that band. You see it used here and there. Kind of like SAP where you can still hear a bit of the original in the background.
Latest Answers