Image Processing – Noise Reduction - AWS Elemental Live

This is version 2.18 of the AWS Elemental Live documentation. This is the latest version. For prior versions, see the Previous Versions section of AWS Elemental Live and AWS Elemental Statmux Documentation.

Image Processing – Noise Reduction


In some applications, you can apply the Noise Reducer Preprocessor to improve output quality. It works by reducing spatial noise, which makes images compress better. But it changes the visual quality of the video.

  • The Filter field has four options:

    • Mean / Gaussian / Lanczos: All of these algorithms allow for varying blur strengths. Mean is the strongest filter (it operates on a smaller group of pixels) while Lanczos is the mildest (it operates on a larger group of pixels).

    • Sharpen: This algorithm sharpens the edges instead of softening them.

    • Conserve: This algorithm limits the pixel values to within the minimum and maximum values of the neighboring pixel values. It is designed to reduce speckle noise. This filter does not seem to be very valuable with video.

    • Bilateral: This algorithm tends to preserve strong edges and is the best compromise between noise reduction and visual quality.

  • Strength: The strength field of the filtering has its greatest effect at 3.


In most cases, enabling the Noise Reducer is not required, but can help output quality if the content will be compressed heavily. When enabled, the recommended algorithm is Bilateral, but the other algorithms may produce better results in certain use cases. Testing the different algorithms using the expected source content is the recommended method to determine the best option.

Location of Fields

Location of Field on Web Interface Location of Tag in XML
Streams > Advanced >Preprocessors > Noise Reducer > Filter



Streams > Advanced >Preprocessors > Noise Reducer > Strength