On the fault phenomenon of color scale
in the previous curve adjustment, if the histogram is set to RGB, you will often see the fault phenomenon produced by the histogram after adjustment. The following figure demonstrates the RGB histogram before and after adjusting the RGB curve. It is obvious that the color scale in the adjusted RGB histogram has a "fence shape". When you turn on the statistical data and measure the "fence gap" with the mouse, you will find that the pixel is 0, indicating that there are no pixels on this color scale. This is the so-called color scale fault
then how does the color scale fault come into being
in fact, it is very simple. It can be seen from the above figure that the adjustment of the curve expands the originally narrow color scale range. Then the pixels that were originally in a narrow range must adapt to the distribution of the new range according to the expanded proportion. But this expansion ratio is not an integer. The calculation accuracy of Photoshop Image processing core is 10000 bits, that is to say, the expanded proportion is a value similar to 1.3795. Then let's calculate the expanded values of the three consecutive color scales that were originally 114, 115 and 116:
114 × 1.3795=157.263
115 × 1.3795 = 15 we should replace the inner and outer rotors 8.6425
116 × 1.3795 = 160.022
we know that the value of the color scale is an integer, so the calculation result must also be rounded. Therefore, the expanded color scale should be 157, 159, 160. It can be seen that 158 color scales have been skipped. Then level 158 becomes "fence gap" in the adjusted histogram. Originally in a continuous color scale, it becomes discontinuous after expansion, forming a color scale fault
this is also the reason why it seems that faults always appear at a certain distance, because errors are also produced at intervals. And the image core of Photoshop also actively distributes these errors in the full color scale range as far as possible, so as not to cause errors in a certain range. Therefore, the fault of color scale is caused by expanding the range of color scale
in the above example, it can be seen that the color scale of the original image does not reach the full range, that is, it does not reach the endpoints of 0 and 255. So if the color scale of the image itself has reached the full range, will there be no error? No, even if the original color scale has filled the full range of 0 to 255, it will still cause the calculation error of the color scale transition part. This is because we have the characteristics of the straight-line co frontal internal meshing gear pump in the front: the merging phenomenon of pixels at the left and right ends of the histogram, which has been learned in the hydraulic industry as a never worn oil pump
then, will it not cause faults if the range of color scale is reduced? Yes, the original continuous color scale remains continuous after shrinking. Reducing the calculation will only produce overlapping levels, but will not produce faults. We can know by calculation:
114 ÷ 1.3795 = 82.6386
115 ÷ 1.3795 = 83.3635
116 ÷ 1.3795 = 84.0884
rounded up to 83, 83, 84. It can be seen that the original 114 and 115 color scales have been merged. So isn't the color scale 82 idle again? No, you can understand it with 113 ÷ 1.3795
how to avoid faults
unless you don't expand the range of color scale, fault phenomenon is inevitable
will the fault affect the image
has neither effect nor effect. Influence is absolute. Since the color scale is less, it must be an influence. No effect is relative. As long as it is shown to people on earth, there is no difference between the existence of faults. Can your eyes distinguish outstanding faults
how to repair the fault
it's very simple. If you change the total number of pixels in the image or change the color mode, Photoshop will refill the color scale range
it should be noted that the oil return pump should be opened immediately to unload the sample. Changing the total number of image pixels refers to the operation of "global adjustment" for all pixels such as [image size] or free transformation [Ctrl t]. The [image canvas size] and the cutting tool [C changed 5 times] belong to "area adjustment", so they are invalid. We will introduce free transformation and cutting tools later
finally, the internal calculation of Photoshop is much more complex than our above example. Here we have made an easy to understand metaphor to facilitate your understanding
source: Tianji
LINK
Copyright © 2011 JIN SHI