Skip to main content
Audio

Noise Reduction

Noise reduction is the process of removing or minimizing unwanted background sounds — such as hiss, hum, air conditioning, or ambient noise — from audio recordings using specialized software algorithms.

Noise reduction is an essential audio restoration technique that improves the clarity and quality of recordings that contain unwanted background noise. Modern noise reduction tools work by analyzing a sample of the unwanted noise (a "noise print"), then using algorithms to identify and suppress that specific noise pattern throughout the entire recording while preserving the desired audio such as speech or music.

Advanced noise reduction tools offer multiple approaches: spectral noise reduction targets specific frequency ranges where noise is concentrated, adaptive noise reduction continuously adjusts to changing noise levels, and AI-powered noise reduction uses machine learning models trained on thousands of audio examples to separate speech from noise with remarkable accuracy. Tools like iZotope RX, Adobe Podcast Enhance, and DaVinci Resolve's Fairlight suite provide these capabilities.

While modern noise reduction is powerful, it has limits. Aggressive noise reduction can introduce artifacts — a hollow, underwater quality, or musical tones known as "chirping" — that sound worse than the original noise. The best approach is to capture clean audio during production (proper microphone selection, placement, and environment control) and use noise reduction as a safety net rather than a primary strategy. When reviewing production audio, a small amount of consistent room tone is preferable to heavily processed, artifact-laden audio.

Termes associés

Retour au glossaire
Noise Reduction — Glossaire | O'Yelen Studio