Spectral Editing
Definition
The visual and surgical editing of audio in the frequency domain, allowing specific frequencies at specific moments in time to be identified, isolated, and removed or modified. Used in noise reduction, sound design, and dialogue repair.
In Simple Terms
Editing audio by looking at a visual map of frequencies over time. You can literally see and surgically remove a dog bark, a phone ring, or a cough from a recording without affecting the speech around it.
In Practice
A spectral editor is used to remove a dog bark from a dialogue recording by visually identifying the bark's frequency and time position in the spectrogram, selecting it precisely, and attenuating it without affecting the surrounding speech.
Common Confusion
Spectral editing is surgical, not magical. Aggressive removal leaves audible artifacts — phasey "underwater" textures or smeared transients in the affected band. The cleaner the source recording, the better spectral editing performs. It is a repair tool, not a substitute for capturing audio properly on location.
Sources & Verification
- iZotope — RX Audio Editor user guide and whitepapersiZotope Inc.
- Wessler, T. & Avendano, C. — Audio Restoration: Spectral Repair TechniquesAES papers (various)
- Cinema Audio Society — Dialogue cleanup workflowshttps://cinemaaudiosociety.org
Last verified: 2026-05-05