Noise Reduction | SpectraLayers Elements Tutorials
Reduces noise by creating a noise profile. See also Voice DeNoise if working with speech content specifically.
The first step is to select an area with noise, without any of the signal to keep. This area should cover as much noise frequencies as possible.
Then click to register the noise profile.
To perform the noise reduction, then click .
The graph at the top shows frequencies on the horizontal axis, and power on the vertical axis. The white curve corresponds to the noise profile you registered earlier, and the red curve corresponds to the threshold (based on that noise print) that you defined with the parameters below. If your signal goes below that red curve, it’s considered noise and gets cut from your recording.
- Tolerance: Threshold under which frequencies should be attenuated. The threshold is relative to the registered noise print.
- Tolerance: Tolerance amplitude from the defined threshold, to fully eliminated signal.
- Horizontal Smoothing: Increase for smoother denoising, at the risk of softening attack sounds too much.
- Vertical Smoothing: Increase for smoother denoising, at the risk of losing frequency precision.
- Reduction Ratio: Ratio of attenuation, in percents.
- : Choose to reduce the noise or the signal.
