Non-Stationary Noise
Definition
Interference whose spectral content changes over time: a vehicle passing, a crowd noise that rises and falls, or a door slamming. Adaptive filtering is required because a fixed noise estimate based on a single silent interval will not model the interference throughout the recording.
- Defining trait
- Spectral content changes over time
- Typical examples
- Passing vehicle, rising and falling crowd noise, a door slam
- Mitigation approach
- Adaptive filtering
- Contrast
- Stationary noise, e.g. a constant hum, keeps a fixed spectral profile
Common questions
Why can't a single silent segment be used to model this kind of noise?+
A noise profile taken from one quiet moment only captures the interference present at that instant, so applying it across the whole recording either under-suppresses noise that changes character later or removes speech energy where the noise profile no longer matches.
What techniques handle non-stationary noise in forensic audio enhancement?+
Adaptive and spectral-subtraction filters that continuously re-estimate the noise floor, along with Wiener filtering and, increasingly, neural denoising models trained to track changing background interference, are used to keep pace with the shifting noise.
Related terms
- PESQ / STOI
- Perceptual Evaluation of Speech Quality and Short-Time Objective Intelligibility: instrumental measures that predict listener intelligibility without needing human subjects. STOI scores between...
- Spectral Subtraction
- An enhancement method that estimates the noise spectrum during a speech-free interval and subtracts it from every subsequent frame. Simple and effective...
- Stationary Noise
- Noise whose spectral characteristics remain approximately constant over the duration of a recording. Tape hiss, HVAC fan noise, and mains hum are...
- SWGDE
- Scientific Working Group for Digital Evidence. A US multi-agency body that publishes consensus best-practice documents for digital forensic disciplines, including image authentication,...
- Wiener Filter
- A minimum mean-squared-error filter that computes a frequency-dependent gain function from the signal-to-noise ratio in each frequency band. Less aggressive than spectral...