Color, contrast and lab!!

Our lives are full of tradeoffs and trade-offs. Trade-Off between form and nature, trade-off between price and performance, and trade-off between security and practicality. In print, especially in the reproduction and processing of images, the trade-off between color and contrast is the most delicate and peculiar-the two concepts are so tightly linked that it seems impossible to change one without changing the other.
Many artists have the suspicion that this trade-off is avoidable, and that a part of the solution can depend on the concept of Cielab, l*a*b*, or lab, known to connoisseurs.
This article is a discussion of actual production, not a pompous void, which gives the reader a series of interests. In all the material that has emerged in this area for several years, it is a brochure that can be convincingly liked: "5 minutes Photo CD appreciation" The secret of its success is that it is based on lab methods to quickly adjust the image to improve the quality of all photo CD images. Such a question is not mentioned anywhere else. It points to two features that lab does well: sharpening and setting the entire color range, but this is just the beginning of lab stories.
In fact, each person chooses in RGB or CMYK to make the revision and color correction, the choice of lab is respectful. Choose RGB or select the Cmvk method as a way to deal with a number of very loyal users, unless you are engaged in high-end color correction, otherwise both methods are feasible. In fact, RGB and CMYK are very similar, for people familiar with a color space, it is also very easy to operate in another color space.
But lab is not the same, it does well in some ways, but on the other hand it's bad. We are now going to take advantage of the good aspects of it (the good things will be more than the article says). Here we will use lab directly to modify the image and co-ordination the gradient range. Next time, we'll explore the correction of the curve under lab color space.
Color and contrast
We can use lab precisely because it is a color separation from the details of the image, which solves many problems that are difficult to solve in RGB and CMYK. If you are accustomed to operating in RGB color space, you are free to switch between lab and RGB, without loss of quality in the conversion. If you have already operated in CMYK color space, there is only a small loss of quality to lab, but the loss is offset by the benefits of this conversion.
Lab has three channels, and in this page of images you can see that A and B channels are almost indistinguishable, they just define the color portion of the image without defining the shape details of the image, which is different from RGB and CMYK.
Lab gamut space is much wider than any color space we are familiar with. As a result, only half of the tones are used in both channels, no white or black, and some are just grey.
Understanding the phenomena of these two passages is not laborious, as one scientist calls them: the opposing channels of interaction. In popular terms, it is in the A channel, the bright part of a region red than green, the dark part of a region of green more than red. The lighter (or darker) indicates the greater the color saturation. A pure ash, of course, represents a gray color. B Channel works the same, light means that a region of yellow than blue, black means that a region more blue than yellow. This can be explained in our example image, why a channel and B channel would be like that, in the r-g opposite color of a channel, emphasizing the contrast of red highlighting and trees, while the tree emphasizes green, in y-b opposite color B channel, y-b This contrast is weaker, because the tree and the yellow in the font is slightly more than blue.
The L channel is simpler and can be seen as a black-and-white version of a crisp image.
Without lab, whether you use a filter, a curve, or any other tool that fails to represent the thinnest color in the image, lab is a common way to make the body color more colorful. It's often the last thing you'll use. Of course, if you are in the no color of the L channel for revision or color correction, it will not have any effect.
The virtual mask applied to the L channel is more efficient than any composite channel applied to the RGB or CMYK color space. At the beginning of this article, you will understand why, the virtual mask means that there are some dynamic-like boundaries around the body touching each other, exaggerating the transition area to make the object appear clearer, the darker object getting a darker boundary, and the lighter object getting a brighter boundary, The result: Instead of the past from white to black, we have a compelling way of getting new boundaries from white to white to black to darker boundaries.
If the virtual mask USM is applied to the entire image, it is equivalent to applying the USM to each channel respectively. In this example, I sharpen the image in RGB, the result is better than the effect of CMYK.
Sharpen in RGB, will make yellow wine trademark on the word deepen, but it will also produce an offensive white glow around the logo, creating an exaggerated transition area in the blue channel, which is also replicated in the yellow version, and another catastrophe will occur in the second image, A combination of two contrasting colors at the edge of a red uniform. In RGB files, there is no such exaggerated transition area in the blue channel, and there is such a transition in the red and green channels, but they go from bright to dark in the opposite direction.
There is no doubt that vivid color is very beautiful, but it is not what we want through the USM desired effect. You will notice that the woman's neck has become green in contact with the red uniform, the edges of the uniform have been blurred by the change to a more gorgeous red, and the yellow star on the uniform is like the red pattern on the wine label, and the color becomes even more yellow.
If we use sharpening l channels instead, these color problems will be solved. On the wine label, the yellow color of the word will become lighter, but it is still very clear yellow, does not produce nasty white, in the white uniform and background contact Place, will produce a bright halo around, but this halo is more than the result of the dazzling green halo produced by the effect of better.
The problem of fuzziness
Many of the Divergent (Defocus) filters in Photoshop include: Despeckle, Median, Dust &scratches, and all kinds of fuzzy virtual tools are processed using digital equations, which you use with me, we sometimes have to use, These tools are not very satisfying.
Unless, we need a fuzzy image of the case, when encountered a very rough image, in RGB or CMYK color space to apply these filters, will always damage the details of the image, and can only be adjusted at a lower level. lab--when the image is extremely blurry, unlike other color spaces-it has two channels that don't contain details and only color, is that a good thing? In a typical picture, we can almost use our ideas to blur A and B channels, more or less the same (reason), the L channel will be very sharp. If there are image details in a and B channels-we do not want to damage these details, the Gaussian blur ratio can be set to a maximum of 3.0, which usually destroys the image, and the set value in a and B channels is acceptable. The gray noise left in the L channel is much weaker than the noise retained by the color channel.
In addition to due to rapid photography and inappropriate filming conditions caused by noise and coarse particles to do, the scanning work of people know: RGB blue channel, will become CMYK in the yellow channel. It's usually more stable than we think. The human eye is not sensitive to the color of the yellow version, we usually use a Gaussian blur to get the smallest detail loss, without taking into account that we could have solved the problem by blurring the B channel without losing the image details.
Another aspect of using this technique is the processing of colored originals of the screen, such as a remake of a photo from a print. The best traditional approach is to scan with a very high resolution and slightly fine-tune the virtual focus, and then sample down. A better approach would be to blur A and B channels, because the dots in the color channels often produce moire in print, which is likely to appear in print. The reserved dot pattern will only be confined to monochrome, we can remove the dot pattern or all in the L channel.
Another very pleasing use of the Blur tool is the application in image collage. When we synthesize an object into another already existing image, we usually use some kind of anti-aliasing to avoid jagged edges of the two image contacts. But because of the precision of the graph, the reasons for the different backgrounds of the two images or other reasons may lead to some unsightly transition colors on the edges of the two image contacts. It seems difficult to decorate these incredible colours without any difficulty. However, if you go into lab color space and blur A and B channels, the problem can be solved.
The decoration and color range of the image
Take a look at the outdoor image on the second page of this article and make the following assumption: If someone is stupid, he put the red words together with the image below, without using the layer and has already saved the file, now we have to delete the text from the image below the restore font, and these stupid people, as you know, have no backup of the image files.
At this point, you may want to hear another stupid final solution, and he advises you to fix the image part of the damage by not only using the "clone" tool, but also using Photoshop's "smudge" tool. This sounds like a ridiculous thing to do unless you happen to be in the lab color space.
However, if you intend to make such a change, you must feel that it is much easier to recognize that the L channel is modified individually than to fix the entire color file. When you remove the fonts in the L channel, you hire a 4-year-old child to handle A and B channels. Once you explain to the child how the "smudge" tool is used, they can also do this kind of work. If you devote yourself to this work, the methods used are more complex and more effective, and there will be a wealth of color changes hidden in the L-channel where the original need to be repaired.
All in all, The L channel is ideal for image modification, and the L Channel sharpening tool works better than the virtual Mask tool filter, and the brightening and dimming tools are also effective in the L channel, because brightening and darkening the image does not produce a new basic color. If someone asks you, the best color space is CMYK or RGB for image modification, retouching and image compositing? You can say: Neither of the two is said.
Thinking about the color range
The most important color correction technique is to set the high light and dark tone, using no color information of the L channel is a very sensible way, it can avoid in the color correction process to produce some very jumping, do not want to see the color. The entire halftoning range is definitely a key factor in color reproduction, normal scanning color compared with the scan will certainly have some gaps, so this requires us to make corrections. In RGB or CMYK color space This can be done with curves or distributions, but to do a complete color correction, you have to separate each channel to handle. Unfortunately, Photoshop does this by using a main curve feature that seems to be so easily corrected that we are puzzled by the function and do not have to do more detail on color correction.
It is not appropriate to set a color range with a CMYK main curve, because if you work on a three color version (channel), it will destroy the black version. The RGB main curve changes do not have this shortcoming, but it also has the shortcoming may cause some small problems to become the big question. The RGB main curve will theoretically have the same effect on all three channels, but in fact it is only when the three-channel start point is the same. If we want to use the RGB curve to highlight the light, and not to think about the light of the starting point, as we have already seen, there is a clear and obvious main hue in the highlight, and we do not want it to appear.
If you decide to use only one curve instead of three, the solution is not the RGB main curve but the lab's L channel, and it can set the color range without changing the color that has been tuned, even for people who are unfamiliar with the curve. In this most convenient way, convert the file to Lab color space, set the brightest and darkest areas of the image, and set the information panel in Photoshop to read the lab color. The L channel is defined as 100 (white) to 0 (black), such a maximum range in the printing does not represent the details of the image, for the general print to say, the maximum and minimum value is about 95 and 10, the curve of two endpoints toward the center of the movement, until the known white spots and dark points reached the above value.
Sharpen the L channel, Blur A and B channels, use the L channel to create a color range, and the way lab handles images doesn't look scary, right. In order to make a more beautiful image, from the beginning of RGB to the final CMYK, if necessary, in the middle of a very short time in lab sharpening, correcting tonal range, when needed, can also be used to reduce noise processing.