Processing
Noise reduction refers to any process or technology designed to minimise unwanted background noise in an audio signal. In analogue recording, it typically involves dynamic pre-emphasis and de-emphasis systems such as Dolby or dbx, while in digital audio it can also include algorithmic or spectral noise suppression.
Noise reduction systems were developed to counteract the inherent tape hiss present in analogue recording formats. Hiss results from the random motion of magnetic particles on the tape surface, which produces a constant broadband noise. To reduce this without affecting the desired audio, noise reduction circuits manipulate the signal’s dynamics and frequency response during recording and playback.
The two most prominent analogue systems are Dolby and dbx. Dolby systems, such as Dolby A, B, and C, use multiband companding, applying gentle compression to low-level signals on record and complementary expansion on playback. This reduces hiss while maintaining natural dynamics. dbx systems use a broadband 2:1 compression on record and 1:2 expansion on playback, achieving greater noise reduction at the cost of possible artefacts if playback calibration is imperfect.
In digital audio, noise reduction can refer to signal-processing techniques that identify and remove unwanted noise after recording. These methods include spectral subtraction, adaptive filtering, and machine-learning-based restoration tools that can isolate and attenuate specific noise components such as hum, hiss, or background ambience.
Proper alignment, calibration, and consistent use of the same system for recording and playback are essential for analogue noise reduction to function correctly. Whether analogue or digital, the goal remains the same: to preserve the integrity of the original signal while minimising noise introduced by the recording medium or environment.
