Image Sharpening with DxO
| | |

Ultimate Guide to Image Sharpening with DxO

Image Sharpening is a topic about which I’ve seen more disinformation than almost any other processing topic on the internet.

Image Sharpening with DxO software is fairly straightforward but there are a lot of choices in the tooling which leads to a certain amount of confusion in users new to the software. In this post I’m detailing the various options and hopefully indicating some good practices along the way.

Traditional thinking around sharpening images produced from RAW files is that there are two stages. Input or Capture sharpening which compensates for the softness introduced by the RAW file and Output Sharpening which compensates for the softness introduced by print media and /or screens as well as allowing for some creative uses to emphasise depth for example.

It is useful to hold this thought in mind, we no longer manually apply irreversible Capture sharpening to our RAW files, instead, a level of corrective sharpening is applied automatically in most processing software, but the principle of a corrective stage and an output stage is a good one which should keep us out of trouble!

In camera, the jpeg that you see in the display screen when you check the picture is sharpened in camera. The RAW file is not, because to do so would introduce permanent change to that file.

What is a RAW file really?

We need to dive in a little deeper briefly to look at the nature of RAW files and sensor technology.

Essentially a sensor consists of a layer of silicon and a layer of photosites which are light sensitive. These photosites simply measure the amount of light falling on them. The luminosity in other words. This is why Capture or Input sharpening is based on luminosity.

Above the photosites in most sensors, is a Bayer filter. This is an array of green red and blue filters, each corresponding to one photosite.

Thus, each photosite is sensitive to only one color, and in the majority of camera systems, each photosite corresponds to one pixel in the final image.

Each green photosite is surrounded by two red and two blue photosites in alternating red-green and green-blue rows. The sensor is divided into 50% green, 25% red, and 25%  blue photosites. The colours we see in our images are computed by comparing the RGB values of neighbouring pixels.

We are used to seeing sensors measured in megapixels. But this is a little misleading because a megapixel is an abstract concept. Despite the labelling on the image above, pixels do not exist on a sensor. Pixels exist on an image and are mapped to the photosites erroneously labelled here as Sensors.

Pixel values are computed from photosite information. 

Let’s take a look, using RAW Therapee with demosaicing turned off …..

RAW Therapy UI with Demosaicing turned off

What is happening here? And why is the picture green?

Let’s take a closer look at the catamaran in the image.

Close up of RAW file before demosaicing with visible red green and blue pixels
RAW file before Demosaicing

To obscure the picture more, many camera have a low pass filter or anti aliasing filter  which prevents moire (abstract patterns most often seen in textiles including small checks). 

When we save a RAW file, there is no processing done – unlike the in-camera jpeg, the task of processing – converting those red green and blue filtered photosites into a color range of thousands of colors, is done by the software in a process called demosaicing. 

Demosaicing describes the process of assembling real world color information from the information contained in individual photosites. If you think of the way a printer mixes a small number of inks to produce thousands of colors, then this is broadly analogous, the processing software is assembling an area of colour using evaluations of each photo site and its neighbouring photosites. 

Each pixel is computed by combining the information from at least four adjacent photo sites, a red one two green ones and a blue one.  So a 1920 x 1080 RAW file contains 6.2 million pixels which are computed from 6.2 million physical photosites + the ones used around the edges to make the computation of these edge values as accurate as those in the body of the image. 

The combination of manufacturer specific sensor and vendor specific raw processing is the reason that Adobe files for example look different to DxO files and Canon different to Fuji.

Some sharpening is applied during the demosaicing process refining the luminance data to correct the softness inherent in RAW files and in some cases softness inherent in the lens. This is analogous to input sharpening, but is applied to a copy of the original RAW file in memory and saved during the output stage.

Capture Sharpening applies to the RAW file – and overdoing it risks permanent damage so most image processing software carries out capture sharpening according to fixed or preset parameters which are carried through to the output file in the absence of any other sharpening.

A lot of software, for example DxO PhotoLab has default settings determining the amount of sharpening that is applied during the conversion process. Which means that many photographers are able to sail blissfully through life entirely ignorant of the amount of sharpening applied to every RAW file they process.

We’re going to take a look at all the methods used in the DxO catalogue to deal with sharpening.


Sharpening in PureRAW

PureRAW carries out lens sharpening within fixed parameters, outputting a DNG, tiff or jpeg.

In later versions of the PRIME algorithm eg. DeepPRIME and PRIME XD2 the user can choose between four levels of sharpening.

DxO PhotoLab

PhotoLab applies lens softness corrections using the lens profile if recognised. This can be turned off or on.

This sharpening can be added to using the Unsharp Mask. The Unsharp mask is a generic filter that applies contrast to edges in the image, effectively tricking the eye into perceiving them as sharper. In my opinion, the unsharp mask is too aggressive to be used globally across an entire image. It can introduce unwelcome artifacts and change color information through its use of contrast. I am much happier using the lens softness correction which applies to luminance only.

The unsharp mask can also be accessed in the context of local corrections such as Control Points under the Sharpening slider. This adds the possibility of using the unsharp mask for creative purposes.

Nik Sharpener

This is by far the most flexible sharpening available from DxO.

Nik Sharpener contains Nik Presharpener and Nik Sharpener Output. If you want to use Nik Presharpener and as we will see there are good reasons to do this, not least of them being control over the amount of sharpening in different areas of the image, you need to turn lens softness correction off in PhotoLab if thats where you are invoking Nik from because two layers of sharpening at this level runs a risk of introducing noise.

Let’s take a look at Nik7 PreSharpener in use –

There are three controls

Adaptive Sharpening

Balance Sharpening and

Image Quality – Normal or High ISO.

Then there is Selective Sharpening which can be applied to Control Point, Lines, Polygons or Luminosity Masks 


Color Ranges

Output Sharpening

Output Sharpening applies to user preference ie for creating perceived sharpness or sharpening for screen or print. This process should produce a tiff or jpeg and is not permanently applied to the raw file.

PureRAW applies noise reduction and lens softness correction to a converted RAW file – tiff jpg or dng.

There are four levels of sharpening, Soft, Standard, Strong and Hard.

In DxO PhotoLab you can apply sharpening via the lens softness correction Global, Details and Bokeh

and then via the Unsharp Mask

Intensity – entire image equally

Radius – thickness of the edges to be sharpened

Threshold – threshold of sharpening ie below threshold no sharpening (we don’t want to sharpen noise for example)

Edge Offset – Additional Sharpening to edge of the image

These are all about perceived sharpening on screen and like the ClearView slider should be used sparingly.


The boundaries between input sharpening and output sharpening have become blurred to the extent that many people don’t realise that input sharpening is happening as part of the demosaicing process.

All other sharpening, falls into the realm of creative sharpening but it is useful to keep in mind that sharpening is essentially a corrective process and should be used sparingly in the creative sense. That being said some sharpening oil the foreground is a good way of adding depth to a landscape image.

In terms of which product is best, it’s a matter of horses for courses. If you are using Lightroom your images will certainly benefit from using PureRAW as you will get all of the benefits of the DxO demosaicing process, leading to a better quality DNG file to work on in Lightroom. If you are using DxO PhotoLab, you don’t need PureRAW.

PhotoLab users will enjoy the additional Sharpening tools and the flexibility of using these tools in local adjustments. This opens up the options for creative sharpening considerably.

The real star of the show though is the Nik Collection tools, Nik 7 Presharpener and Nik 7 Sharpener Output. These are very powerful tools covering the full range of input and output sharpening. Users need to turn off sharpening in the programs they invoke Nik 7 Presharpener from because applying two layers of luminance based sharpening risks increasing noise.

Nik 7 Sharpener Output produces very realistic sharpening using a variety of controls. To my mind the images I’ve produced using these tools are slightly more realistic with better contrast and focus than those I’ve produced using PhotoLab alone.

Bear in mind that these are very small differences and opinions are subjective, others may find they prefer PhotoLab to ink Sharpener, it is certainly more straightforward.

Further Reading

.DxO PureRAW 4 Review

.Noise Reduction in DxO PhotoLab

Get the Software

Try out DxO PureRAW 4:
Buy DxO PureRAW 4:

Try out DxO PhotoLab 7:
Buy DxO PhotoLab 7:

Try the Nik Collection free for 30 days at:
Buy the Nik Collection at


I’ll keep you in the loop with regular monthly updates on Workshops, Courses, Guides & Reviews.

Sign up here and get special prices on all courses and photowalks in 2024

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.