An Introduction to Photo Editing with PhotoLine: Raw Conversion
If your camera is able to save raw images, you may be able to obtain images of higher quality than you would otherwise; but it takes some doing as raw conversion is a complicated process. I personally recommend RawTherapee, a free and open-source application, for raw conversion, partly because it is one of the few that can use .dcp camera profiles. (This is another subject; you will need an X-Rite ColorChecker, as well as ColorChecker’s own calibration software and a .dng converter: Adobe - Windows | Adobe - Mac | digiKam, all free downloads. The process is actually fairly simple, and the results are amazing.) But PhotoLine is able to process raw images as well. Ideally, you should use an ICC profile specific to your camera, specified via Edit – Options – Color Management – Devices. You may be able to download one; or you can make one with a ColorChecker and Argyll CMS. (The process is more difficult than with .dcp profiles, and the results not as good.) But if you must, you can process your images without one.
Note that some of the adjustments described below may be applied to non-raw images as well.
Every camera uses a specific raw image format; for example, Nikon uses NEF and Canon uses CR2. If you are using the latest version of PhotoLine, it will probably recognize your camera’s raw format. When you open a raw image, two layers are automatically created in addition to the background image. Do not make any adjustments to the background image itself, or it will be converted from raw to a conventional RGB image. Save a master file as a PLD, and the final converted file as a JPEG or whatever.
Sensor array interpolation/demosaicing
First, note that the background’s Layer properties panel includes an entry called Pic. Data that specifies the sensor array interpolation method. Raw images are not like ordinary RGB images, and the pixel data must be put together in what may be a complex way. The default VNG Interpolation Cubic seems to work for this image; note how much more intense the chromatic aberration is using Cubic Interpolation. There is also a little bit more roughness to the fine lines with the latter.
Which brings us to our next subject. Chromatic aberration is not included in the default adjustment layers; create a new one via Layer – New Adjustment Layer – Chromatic Aberration. This image has red and cyan color fringes (look for them at sharp light-dark boundaries, especially near the edges of the image), indicating a need for Red Correction. Width Compensation is intended to compensate for aberration scale differences farther from the center of the image. The status of the checkbox is the reverse of the caption; this image seems to do a tiny bit better with it unchecked.
Raw white points
The vector layer is a child of the background layer; click the arrowhead beside the image thumbnail to see it. The relevant control is Correct Highlights. The RGB sensors of digital cameras sometimes have slightly different dynamic ranges, which can cause artifact colors, usually magenta, in the highlights; if this is the case, Correct Highlights enables you to clip the white points and eliminate false colors.
The Raw Adjustments layer is a rather busy layer; we are going to hit the highlights. Histogram Correction (sometimes called “Levels”) is not my preferred tool, because you can usually do the same thing with Curves. The first and third arrows under the histogram adjust the black and white points of the image scale; sliding the origin and endpoint of a curves adjustment does the same thing. And the middle arrow is comparable to the midpoint of a curve. The only time you actually need to use Histogram Correction is if the range needs to be widened. Use View - Mark Extreme Values to see clipped highlights and shadows, and move the white and black points to get rid of them. In the image below, the shadows are clipped and the black arrow needs to be moved down, if possible. Color casts in the extremes may be dealt with by adjusting the RGB channels separately; it is hoped that you will not need to do this.
Beneath Histogram Correction is Curves which we have already covered; adjust as needed. This can take the place of Color Correction and White Point as well; if there is a neutral gray in the image, click the gray eyedropper on the Curves panel, then click the gray area in the image. Next is Light/Shadow, which is a rather coarse tool, especially the Width sliders; curves will probably do a better job. For this image, increasing the Light gamma was a quick way to darken the highlights a little.
Color Temperature is a convenient way of warming or cooling an image. Correct Color Temperature is a logical choice here as it determines the source color temperature, which for midday sun is about 5500K. Set Color Temperature adjusts the image as if to display the indicated temperature. If you used the correct color temperature setting on the camera when you took the picture, you may not need to use this adjustment.
Noise reduction is a subject all its own. Noise tends to be worse in the shadows, and small sensors produce more noise than big ones. Cameras do a lot of image processing when saving JPEGs, including noise reduction. If you are processing your own raw images then you have to deal with the noise yourself. This image did not have a lot of noise; I enhanced a shadow area to illustrate how to deal with it. It is best to view the image at 100% for this.
Since noise tends to be worse in the shadows, and since noise reduction can obscure image details in well-exposed areas, you may want to limit noise reduction to the shadows. It will then need to be on its own layer (Layer – New Adjustment Layer – Denoise). You may also want to do it more than once, first for the shadows and then differently for a clear sky, for example.
Color noise is more objectionable to the eye than luminance noise (“Intensity Noise”); a little of the latter may actually contribute to the impression of sharpness. Noise reduction is a balance between getting rid of objectionable value differences (noise) and smoothing out actual image details. In the Denoise panel, Size refers to the size of the chunks that are removed as noise. Setting the maximum value for Color Noise reduced it from the first to the second image below.
More controls are provided for dealing with Intensity Noise. Threshold determines the value difference that is considered noise; 0% does not remove anything; 100% removes a lot, and can make the image look like plastic. Appropriate values are often in the single digits. Intensity just limits the overall effect. So what you need to do is to find the magic balance between Size and Threshold. The adjustment in the third image below may be a little too much, depending on how it looks after sharpening.
Now we are going to apply a Color Filter as we previously did with the saturation curves. This will limit the effect of the layer to the shadows. Right-click the denoise layer and select Color Filter. Move the Background upper arrow down until the noise starts to reappear in the shadows; move it up again to an appropriate level. Then raise the upper half of the split arrow. Look at a well-exposed middone with little or no noise, and move the upper half of the arrow to the point where Denoise has no effect. This will probably take some trial and error.
Plugins such as Neat Image are available that give more control over the noise reduction process.
Appropriate sharpening is an essential part of raw processing. Unsharp Masking is a sharpening method that we have already addressed. Deconvolution may be better for raw capture sharpening, and raw images need a greater amount than JPEGs. Again, cameras sharpen JPEGs; when processing raw images, you are on your own.
Two more adjustments, Lens Correction and Rectify, may come in handy, but unfortunately they are not adjustment layers and so they must be performed on a stamped copy like sharpening: Edit – Merged Copy, Edit – Paste as Layer. Showing the grid helps with both of these; View – Alignment Helpers – Show Grid. Adjust the grid appearance via Edit – Options – Display – Grid; adjust Grid Distance for the size of the image; maybe set Type of Grid to Line, and the color to gray. Lens distortion makes straight lines look bowed in or out, as if on a curved surface; use Filter – Digital Camera – Lens Correction to straighten them. And for perspective distortion, such as when looking up at a building, use Filter – Digital Camera – Rectify and place the four points so they define what should be horizontal and vertical lines.
Next: advanced selections and masks