Controls

Luminance/chrominance threshold sliders allow you to adjust the way how plugin detects image details and discriminates them for noise. Higher values make plugin suppress more noise at the expense of losing some image details. Lower values leave all image details at the expense of some more noise left.

Luminance/chrominance suppression sliders tell plugin to leave some small amount of noise in the image. Completely noise-free image may look unnatural because all real world images contain some small amount of noise. Specifying maximal suppression for luminance and chrominance noise (in decibels) lets you to leave some noise to make the image more lively and possible to retain some details that could be lost in the process of denoising.

Light/dark areas correction sliders enable to separately tune noise reduction in highlight and shadow areas. It is known that typical digital photos have more noise in shadow areas and less noise in highlight areas. By setting light areas correction slider to negative values you can lower the noise detection threshold in highlight areas and leave more details there. By setting dark areas correction slider to positive values you can suppress noise in shadow areas stronger. If you set these sliders to positive values and the noise reduction is still not enough – try increasing the luminance/chrominance suppression sliders – they constrain the maximal amount of noise suppression and you may want to increase it.

Developed in Moscow State University, Russia

 

MSU NOISE REDUCTION

A plugin for advanced noise reduction which includes an automatic noise level detection system that adapts to noise of different grain and chromaticity. Unlike conventional denoising algorithms that just blur the image in even areas, our approach based on wavelet decomposition allows to suppress noise "throughout the detail" without general blur of the details.
           

Original noisy image

Denoising using conventional denoiser

 

MSU Noise Reduction

 

For example if you increase the threshold and the suppression, you can get the above result with even less noise:

And if you are creative, you can improve this result even further by adding some saturation and applying a gentle unsharp mask: