ADR — Automated Dialogue Replacement — is the process of re-recording dialogue after principal photography to replace unusable production sound. When done correctly, it's completely undetectable. When done poorly, it's one of the most jarring elements in a finished film. The difference between the two comes down to the precision of the recording setup, the performance, and the depth of the editorial work.

Why ADR Fails and How to Prevent It

ADR fails for three primary reasons: the recording room sounds nothing like the original location, the performance doesn't match the on-camera timing and energy, or the spectral character of the voice is different because of microphone choice, room acoustics, or processing. Each of these is a solvable problem if addressed at the right stage.

The most common mistake is recording ADR in a dead, over-treated acoustic space that sounds nothing like the original production location. A line recorded in a noisy, reflective restaurant should not be replaced with ADR recorded in a silent floated-floor studio without significant post-processing to restore the acoustic context.

The Professional ADR Recording Setup

A professional ADR stage uses variable acoustic treatment — moveable panels that allow the engineer to tune the room's reverberance to approximate different location characters. The microphone should match or closely approximate the production microphone — for exterior scenes, a hypercardioid (like the Schoeps MK41 or Sennheiser MKH 50); for interiors, a cardioid or supercardioid. Matching the microphone character is more important than matching the exact model.

The actor watches a looped section of picture (the ADR loop) and speaks in sync with their on-screen performance. Most ADR engineers use a three-beep countdown system followed by the loop. Recording multiple takes — performance first, sync refinement second — gives the editor options for both acting quality and timing precision.

ADR Editing: Timing, Phase and Spectral Matching

Raw ADR recordings rarely sync perfectly to picture on the first take. The editor's job is to use elastic audio, clip gain automation, and fine-cuts to lock the replacement to picture within one frame (40ms at 25fps). Most experienced editors work at the sub-frame level for close-ups, where even 20ms of timing error becomes visible.

Spectral matching is the most demanding part of ADR editorial. The replacement voice needs to have the same tonal character as the production dialogue around it. Tools like iZotope RX's Voice De-noise, EQ matching, and Dialogue Contour help shape the ADR to fit its acoustic environment. Adding a small amount of the location room tone under the ADR line — even just 2–4 dB below the voice — dramatically improves the sense of placement.

Integration in the Final Mix

In the final mix, ADR is processed through the same dialogue chain as production sound — the same EQ curve, the same subtle room treatment, the same dynamic range. The goal is for the mix to breathe consistently across production sound and ADR without the listener ever noticing the transition. Automation between the two must be handled with care — level riding on an ADR edit point is one of the most audible signs of a difficult session.