

The image above shows zoomed-in examples of the different steps, and below is the full edited image. To bring back some of the detail, I increased Lens Sharpening Global to 1.8. A lower value would have been enough for most of the picture, but 90 gives a better transition between the hill and the sky. I decided to see what it can do to this picture.Įnabling DeepPRIME and bumping Luminance up to 90 eliminates the noise completely, but the image appears slightly blurred. The latest instalment of DxO PhotLab added an improved version of the PRIME denoising algorithm called DeepPRIME. It has a fair amount of noise and does not appear very sharp. It is a single frame taken without a tripod at ISO 5000. I have been reluctant to publish the image in my previous post, Heidelberg Castle at Night. Other edits are the same for all the tree images. EXIF info for the image used in this comparison is as follows: Canon EOS R5, ISO 2500, F/5.6, 1/250th 145mm. Finally, the right image adds Lens Sharpening (Global 1.8). DXO DeepPRIME requires the use of DXO PhotoLab as the raw processor. The middle image adds denoising (Luminance 90). The left image is without noise reduction or sharpening.

DxO PhotoLab 4 DeepPRIME noise reduction in action.
