activate the channel panel
2. Gaussian blur for red channel copy, as shown in figure
Gaussian blur for red channel copy
3. Use the color scale command to adjust the copy of the red channel. The parameters are shown in the figure below:
Use the color scale command to adjust the copy of the red channel
4. Set the foreground color to black, and choose the appropriate brush tool to erase the parts that don't need sharpening, such as the foreground.
set the foreground color to black
5. The selection has been made, then load the red copy channel as the selection, return to the layer panel, copy the background one layer, and don't cancel the selection. Perform filter-artistic effect-painting daub, and the parameters are shown as follows:
Selection has been made
6. Then copy the background copy into the background copy 1, move the background copy 1 under the background copy, change the transparency of the background copy 1 to 3%, and change the background copy into color filter, in order to brighten the picture slightly.
Then copy the background copy into the background copy
The repaired picture
This will make the original blurred picture much clearer. You might as well try this method, which is very useful!
Note: Some friends say that it is not effective to make it. Please read the following text carefully to understand the principle of making it:
Find the most suitable channel, and the principle of finding it: the white area (the area that can be selected) in the channel contains the parts that need to be processed. For example, if there is a picture of a portrait, the clothes and face of the middle figure need to be clearer, then we need to choose a channel with the widest white coverage and relatively brightest among the three channels, G and B, which is equivalent to obtaining the necessary selection for pixel processing
Copy the channel, then process the Alpha channel (the channel obtained by newly building or copying the primary color channel is the Alpha channel), and then use the filter-style. The width of the edge depends on the size of the area you need to deal with, and the value is generally large; Smoothness is suggested to be set to a value within 3, because a higher selection will produce transparency, so we can first get a large selection and then correct it, just like carving eyes smaller and nose larger to have more processing space). The purpose of this operation is to remove unnecessary parts (it is always impossible to copy all the selected areas in the channel that you need to deal with), and at the same time, it can improve the brightness of the needed parts to a certain extent (the brightness improved in the channel is to strengthen the opacity of the white part of the selected area or the edge)
. Smart people may ask, just to improve the brightness of the operation, this Gaussian is not all in vain? Here I appreciate and affirm your question.
actually, this idea is completely wrong. If you carefully observe the channel, you will find that the white area obtained after the operation of illuminating the edge has a sense of "sharpening" visually. In fact, it only improves the brightness of white, and at the same time, it also improves the brightness of many details by the specified value accurately and equally. We should carefully deal with the redundant area when matting, and it is the same here. Gaussian blur can help us achieve this goal more quickly and comprehensively.
Maybe someone will point to it just now. Yes, at this time, we can also use the classic command of image-adjustment-color scale to adjust the white field and black field in the histogram of this channel (the white and black triangle arrows under the histogram are not the two below the bottom black and white gradient! ), drag the black field to the right, drag the white field to the left, and keep the gray field in the middle unchanged, so as to reduce the contrast in the whole channel and make the white reach the range we need; Remember, once enough is enough, don't go too far, drawing is rigorous. How's it going? What else is there to say?
in the end, it is natural to make a copy, and then correct the unnecessary white areas, and use lasso tools or brushes with hardness of 1% to correct them. Be careless and careful, and quickly cover the unnecessary parts with R G B= black
Hold down Ctrl and click the channel that has just been processed, and return to the layer panel. Select the background layer-ctrl+j
as the picture in the example. The author uses filter-artistic effect-painting daub to give an appropriate value, then sets an appropriate transparency and adjusts the layer blending mode. The actual effect is not ideal anyway. The method I use here is to call "overlay" mode+5% opacity
for the layer I just copied. The inspiration of this example is that no matter what style of photos are processed, the sharpening effect can be achieved by reducing the contrast of some edge parts after obtaining the appropriate selection, and it is more natural. Of course, if you are willing to call sharpening directly, remember to do it on the new layer, and at the same time adjust the appropriate layer blending mode and add some opacity to achieve the same or even more. Finally, I would like to send you a sentence. The method is used to inspire the brain, not to imitate it blindly.