Re-recording Mixer

What they do

Work starts at picture lock, on a dub stage calibrated to a known reference level with the speaker layout the film will be shown in. First they audition the delivered tracks on their own, listening for distortion, unintelligible lines, mismatched noise floors and edits that will not survive being loud. Then premixes. The dialogue pass is the hardest: location tracks, ADR and room tone made to sound like one continuous performance in one place, using EQ, de-noising, de-essing and matched reverb. Effects and backgrounds follow, then music, often with the music editor in the room. Then the final pass, reel by reel, with the director and producer present, deciding what an audience should be able to hear at every moment: which line survives the explosion, whether a cue enters under the last word or after it, and where music and effects are competing for the same job. Panning and space are part of it, whether the format is stereo, 5.1, 7.1 or an object-based immersive mix. The output is not one file: dialogue, music and effects stems, an M&E for foreign versions, near-field and consumer trims, and a master that meets the delivery spec's loudness target in LKFS and its true-peak ceiling. Then the fixes, as the cut keeps moving.

What they decide

The calls that belong to this role and no other.

• The relative level of everything, moment by moment, proposed to the director and producer
• How dialogue is treated so it sits consistently: EQ, noise reduction, reverb matching between location and ADR
• Whether a line is intelligible enough, and what gets moved out of its way to make it so
• Panning and spatial placement, including object placement in an immersive mix
• The stem structure, and therefore what the deliverables and trims can be
• Whether a delivered track goes back to editorial rather than being repaired on the stage

Principles of the craft

• Dialogue intelligibility is the floor, not the goal: you can always win a line by turning everything else down, at the cost of the scene's scale. The craft is carving space with narrow EQ moves, ducking one element instead of a whole bus, and moving music out of the voice's range.
• Mix the intent of the scene, not the meters: metering keeps the master legal, but the reel decides whether a battle builds and whether the quiet scene after it actually feels quiet.
• Perspective is level plus room plus filtering, not just a fader: a voice heard across a hall is distance, reflection and a loss of top end. Matching ADR to a location track is the same problem solved backwards.
• Stems are a deliverable with consequences: how you split the mix decides what the foreign version, the trailer, the airline cut and the accessibility track can be. Convenient bussing today is somebody's rebuild later.
• Mix for where it will be heard: a cinema with an immersive layout and a phone speaker do not forgive the same things, so a near-field check is part of the day rather than an afterthought.
• Loudness compliance is arithmetic, dynamics are the art: hitting a platform's LKFS target is a measurement anyone can make. Choosing what to spend the available dynamic range on is the mix.

Common mistakes

• Mixing loud all day, losing perspective, and finding the quiet scenes have no dynamic room left
• Solving intelligibility with a broad dip across the whole music and effects bus
• Processing ADR until it is clean but no longer in the room the scene takes place in
• Judging the mix only on the stage monitors, with no small-speaker or headphone check
• Heroically repairing a badly track-laid session on the stage, at stage rates, instead of sending it back to editorial
• The competent mixer's mistake: a mix the room and the director love that misses the platform's loudness or true-peak spec, or stems that sum correctly but cannot produce a usable M&E

How to learn it

Exercises and routes in, not a reading list.

• Mix one dialogue scene where a music cue and a heavy effects bed both want the same seconds. Do it three times: ducking music only, ducking effects only, and EQ carving only. Have someone listen without picture and write down the dialogue.
• Take a location line and an ADR line of the same words and make them match. Then match them in the other direction. Note which is harder and what that tells you about noise floors.
• Mix a scene, print stems, then rebuild the mix from your own stems and try to make a change the director might ask for. Whatever you cannot do is a flaw in your stem structure.
• Mix to a loudness target with metering, then listen on a laptop, a phone, headphones and in a car. Write down which decisions did not travel and why.
• Take a scene with one big impact and mix the ninety seconds before it quieter. Measure nothing, just note how much bigger the impact became for free.