Digital Seeing Hearing Understanding.

Go to the overview of the Blog's Topics.

Trichromatic Image Creation - Building The Trichromatic Image

by Uwe Heimburger

Tool: Affinity Photo (version 2)
Input: Three filtered DNG monochrome image files

Affinity Photo supports layers and using different Blend Modes similar to what Adobe Photoshop offers. The layers can be used to create a trichromatic image from three optically filtered monochrome captures we get from the K3iiiMono.

Layer Structure

If we already have created a template for the layer structure - portrait and/or landscape orientation - we open a new file using this template.

Otherwise we create a new Affinity Photo file of size 4128 x 6192 px and then generate a basic layer structure and name the different layers in a helpful way, details later. The image below shows how it should look like at the end of the process.

Correct order and grouping of the layers is critical. The Procedural Texture Layers - details follow below - need to be moved above their corresponding monochrome image layer.

For every group layer we have to set Blend Mode "Add", the other layers to "Normal".

As I experienced, all images - even if we shoot from a stable tripod - will be slightly misaligned. The reason seems to be the manual change of the three filters. It can't be done without impact on alignment.

When I started with my process experiments, I didn't align the images in post. I know now, the reason of color fringing I saw, was misalignment. A PentaxForums member wrote about the Alignment Feature of Affinity Photo.

So after creating the base grouping structure, we now open our Red, Green and Blue files using the "File > New Stack ..." dialog.

We see some options to import the images.

We first push the Add-Button. The file browser opens. We select our three images and press the Open Button. The images are shown in a list view (figure above). If not already selected we select "Automatically Align Images". As second option we choose option "Perspective" or "Scale, rotate or transport" in the dropdown menu below. If we like to see the alignment directly we also select "Live Alignment".

In our case we let the default as is, i.e. "Perspective". Then we press the OK-Button. After a few seconds a new Affinity Photo file is opened and a Layer Group of the aligned images is created. Each of Image Layer contains a Perspective Correction Layer and the dng files are developed (raw conversion).

We select all Image Layers and copy them into the Layers Tool where we already prepared our trichromatic structure. Then we move each Image Layer below the Procedural Texture Layer into its color group.

We can now close the Affinity Photo file where the images were aligned. No need to store it. We proceed working with our trichromatic structured file.

Red, Green and Blue Color Separation

Theoretical Part

Image processing apps usually work on RGB basis. So every layer will contain the three channels Red, Green and Blue. For grayscale values of RGB data the following general rule applies: R = G = B.

Our goal is to get RGB channels with the different data from the three monochrome files. So for each monochrome image, we need to preserve the gray values of the RGB channel representing the color of the color filter used. The gray values of the other two channels need to be set to zero.

Having done this, we can blend the different color layers using Blend Mode "Add". Result will be a proper color image.

All this may sound more complicated than it is. Sorry for that. The practical part below, hopefully will bring light into the dark.

Practical Part

In Affinity Photo we find different methods that we can use to create seperate color channels. I use Procedural Texture Layers (Menu "Layer > New Live Filter Layer > Colours > Procedural Texture ...").

The tool's panel looks like this ...

In this panel we have to create an equation for each color component, done simply by pressing the plus button. These equations we will apply to its corresponding "image layer" (the monochrome image data).

The example above shows the setting for the red layer. The only thing we have to do is to write "R" into the equation field for the red channel. If the Procedural Texture Layer is positioned correctly in the layer hierarchy, the channel group image representation will show the channel's red tint.

After that we copy this layer two times and change the copies for G and B.

At the end the Procedural Layers need to be positioned correctly in the layer's tree hierarchy and we should see the proper trichromatic color image in the image view area!

The image might look flat, and color might not really conform to the original color calibration card. The reason is, the colors are determined by the optical filters need to be balanced and corrected according to the color card reference.

For the next steps let's call the image our Trichromatic Color Calibration Image.


To generate an appropriate icc profile based on our Trichromatic Color Calibration Image we need to export the image using the following settings:

  • TIFF 16 Bit
  • no compression
  • no RGB profile

Go to: Calibration & Profiling

Go back