Wednesday 14 May 2014

SPLIT FREQUENCY HEALING FOR THE BEST SKIN LOOK IN PORTRAITS


Split Frequency Healing is a technique used in Photoshop to retouch photos without affecting the photo's respective tone or texture.
Healing and cloning both have their advantages and disadvantages. The Healing tool will copy texture from one area and then merges the texture into the color and
luminosity of the destination area. 
The advantage is that you can use it on an empty layer and get natural looking results. 
The disadvantage is that you can’t control the amount of texture or the amount of
merging. 
The Clone tool doesn’t make this distinction and just pastes texture and color into the destination area. 
The advantage is that it can be used on an empty layer blended “darken” or ”lighten”
and limit the effect, getting good looking results. 
The disadvantage is that you can’t “clone” only texture.
Split Frequency splits an image into its visual frequencies using two layers, one to capture texture (High Frequency) and the other for toning and color (Low Frequency). In this way texture can be manipulated without affecting tone and contour and tones and luminosity can be corrected without altering texture.
To split an RGB image into it's low and high frequencies create two copies of the image (Cmd/Ctrl J twice) or (Cmd/Ctrl+Option/Alt +Shift+E to merge layers and then Cmd/Ctrl J twice). Name the top one High and the lower one Low.
Working on the Low layer apply a Gaussian Blur filter to blur to suit but around a setting of 15 is a starter. The RIGHT radius is the one that smooths sufficiently for the look required.
On the High layer go to Image-Apply Image and apply the following settings:-

Settings for 16 bits:
Settings for 8 bits:


Make sure in the box marked Layer it is the layer below named Low. 
Click OK.
The HIGH layer will now look like a High Pass. Set the Blend Mode of the layer to Linear Light. Opacity and Fill should remain at 100%.

Taking the layers together it will look exactly as your
original image. 
The only difference is that your image data is now separated into two frequencies: 
HIGH containing detail and LOW containing shape. 
The layers may be grouped to keep them together.
We can now use the Healing and Clone tools but make sure the Clone Sample mode is set to Current Layer.
What can be done on the HIGH layer? 
Use the Healing brush (hard edge to avoid blurred texture). Always use a bigger brush than the texture to be replaced. Blend texture with texture. Use the Clone brush (usually with soft edges) in normal and both darken and lighten mode, but now, 
only for texture. 
Use the Patch Tool to replace texture. 
Clipping different curves for different effects. 

What can be done on the LOW layer? 
Using the Healing and Cloning tools in the usual way to replace, smooth or contour color and shading. Create a new layer over it and paint in normal mode, on low opacity to even
out tones. 
Create a duplicate of the LOW layer. 
Blur it further to eliminate middle frequencies. 
Mask and paint with white on the mask to reveal the effect where needed. 
This is called “Band Pass” and it has a lot of variations.
The most important part of this technique is that texture is maintained or created whilst toning is created without the two affecting each other. Some experimentation will be necessary to become proficient.



MATCHING SCREEN IMAGE TO PRINT


There are many tutorials and sources of advice on tackling the perennial problem of getting printed images to match the image on the computer screen. In my photography I have probably tried all of the methods, mostly using a combination of Hue and Saturation, Color Balance and Curves layers in the Soft Proof viewing (Duplicate image and then viewing side by side go to View-Proof Setup-Custom and inserting the printer/paper profile to see how the print will look).

Now, in Adobe Photoshop CS5 and CS6 we can take the image and convert the screen profile to the printer/paper profile and I find this works every time without all of the tweaking and tuning that I previously had to do.

How do we do this.  First we should have the basic requirements in place; that is the monitor properly calibrated (I use the Spyder2Express) and have an ICC color profile for the type of printer and paper (I have an Epson R2880 printer and the ICC profiles come with the software for Epson papers or other paper suppliers will provide ICC profiles free of charge, as with Permajet, the other papers I use)

Now take this image.



I want to print to the Epson Premium Glossy Photo paper. If I Soft Proof the image as in these two captions the appearance dulls down. In other words the printer reads the image in accordance with the color profile it uses, which may have different color readings to the workspace used in Photoshop.

You can see that the workspace used is shown in the bottom left corner and is Adobe RGB (1998). We want to convert it to the RGB readings of the printer profile. To do this go to the Eyedropper tool from the menu on the left set it to the Color Sampler tool and a Sample Size of 5x5 in order to pick up a good sample. Now pick up four points by holding down the Shift Key and clicking around the image; pick points that give a spread of contrast and color in the image. We should see the RGB readings for each of these points in the Info box to the right of the screen.


We convert the profile by going to Edit–Convert to Profile and in the dialog box, which comes up, insert the printer/paper profile where it says Device to Simulate.  Other settings are as shown.  Click OK. Note that in the two pictures below the RGB readings have changed, but the image has not, when the Convert to Profile dialog box comes up and is ready for conversion. 


Click OK in the Convert to Profile dialog box and then Print the image in the usual way as shown.


So far this is the best method I have found for matching screen colors and contrast to the print.