no way to compare when less than two revisions

Différences

Ci-dessous, les différences entre deux révisions de la page.

@@ Ligne 1: / Ligne 1: @@
+This series is aimed at learning to make something of the old photos in my possession, and others in the public domain due to their age. You, the reader, are welcome to tag along and I hope glean some small insight and perhaps an idea or two from time to time. No promises are made as to quality of the content, or potential errors and omissions. I am a computer scientist, not a true artist or a professional of image restoration. So please take all this as a best effort, but with no firm guarantees — much as is the case of most open-source software.
+In the previous part of this series, we focused on clearing up dark splotches and other defects on a photo that had been taken using 35mm photographic film, then scanned using a flatbed scanner and then converted to a positive. As announced at the end of Part 8, we will finish this series by taking a look at old photos taken using the digital cameras that became mainstream. Not all were created equal, and some of the earliest had sensors that lacked the quality we are used to seeing today in modern equipment. In four successive parts, we will tackle some specific defects of early light-sensitive chips: lack of sensibility, which is a drawback when shooting dark scenes; lack of contrast; lack of resolution; and finally lack of color dynamic range. Taken together, these weak points tend to give early digital photos a lackluster, washed-out and blurry feeling in contrast with those made using even rather common mobile phones today.
+To begin with, let us take a look at an interior photo I took in Trinity College, Dublin, in year 2006. It posed two typical challenges to previous generation digital chips: on the one hand, the lighter parts where direct sun comes in through the ceiling and washed out, to the point that some of the joints look thinner than they really are, and on the other the lower half of the image has registered very little light at all. The chip inside the camera simply could not cope with such a huge variety of lighting levels in a single scene.
+Our aim will be to maintain whatever quality is available in the upper half of the image, while enhancing the darker half to pick out as much detail as is hidden among the shadows. A simple strategy would be to use color curves as discussed in part 3 of this series, to correct this imbalance. So, we are off to menu option Filter, then Adjust and “Color Adjustment curves”. We raise the lower (dark) part of the curve, while maintaining the upper half on the diagonal line to retain the original response. With just three control points adjusting the lower half of the curve, the darker part of the image clears up nicely, and we can now see some action at the end of the corridor.
+More color is now to be seen. However, the lower half of the image is still not perfect, there are several darker shadows and the end result is quite dark and has less detail than one would like. Luminosity could, in theory, be increased yet more by giving the color adjustment curve an even steeper incline at the extreme left. This, however, can give weird effects to lower-to-mid tone colors, which tend to group together around a drab gray. This is especially visible in the woodworking around the windows in the first floor, which has a completely washed out look. Basically, this is the result of a section of the adjustment curve that has a very flat slope, which can be seen on the curve between the second and third adjustment points. A flat slope means that similar colors will be drawn together, and contrast is lost.
+So, applying a color adjustment curve directly does not seem to be a good solution. A more advanced technique consists of two stages. In the first, we shall decompose the image into three separate channels. The first, known as Hue (H), gives us the actual color of each pixel, as defined along the chromatic circle. We can think of hue as defining whether we are speaking of a red, or are more close to a green, or perhaps our pixel has a tinge of yellow. The second channel, Saturation (S) represents the force of the color. A gray is a color with low or extremely low saturation; as saturation increases, colors will have more character and seem vibrant. finally, Value (V) tells us whether the pixel is a dark, or a light color. Contrary to other image manipulation programs such as GIMP, in Krita one does not easily perform HSV channel separation decomposing a layer into its different components in separate images. But the color adjustment curve can be used to tweak a single channel. Select “Color adjustment curves” once more, but this time choose channel “Lightness” (this being the Krita term for Value) instead of RGBA (see next page, bottom left). We can now push up the lightness of the lower portion of the image, without altering color balance.
+The first floor woodwork now looks like natural wood, not gray paint. The dark red at the end of the lower hall section is now a more natural color than in our previous try.
+We can go a bit further with this tool. Suppose we now wish to increase saturation, raising color intensity slightly in the lower half of our image. We could go, once more, into “Color adjustment curves” and select channel “Saturation” using the appropriate curve. But doing this would increase saturation all over the board, for all parts of our image. What we actually need to do is increase saturation, but only for the darker colors. For this reason, the tool we will use is in menu option Filter, then Adjust and option “Cross-channel adjustment curves”. In this, we can specify we wish to adjust Saturation, but making our adjustment dependent on another channel. This is called the “driver channel”, which in this case is “Lightness”. So, choose this curve and increase saturation for lower values of lightness – this is the left half of the curve.
+The lower part of the corridor now has slightly more vibrant colors, though remaining quite realistic. The woodwork on the first floor also has a tad more character than in either of the previous images.
+These tools are quite advanced, and showcase some of Krita’s excellent range of options to rework our original photos that are slightly off. With some practice, we can often go quite a bit further in restoring old and dark photos, beyond increasing the whole image’s luminosity. However, in other cases old digital photos can have quite the opposite defect, coming out too clear and with colors washed out by too much light in the original scene for the camera’s chip. We will work on this scenario in the next part of our series. Until then, take care!