Fine image Technology of Photoshop

Pull the picture is the image works of design and production of very important technology, the following two cover pictures are in Photoshop (hereinafter referred to as PS) first use the channel to make accurate selection, the characters and background separation, and then do the effect. But this is also inseparable from the previous shooting, early shooting time to be clear how the latter is to be produced, there is a general direction, so that when shooting as far as possible to choose a more pure background and the subject of the film color contrast relatively large to shoot, so in the later production time will be more effective.

We take a closer look at these two cover works, you can find many dozen moving details, a picture to have vivid details is to give people impact, of course, no high-precision pictures are difficult to have details. Many designers use ultra-high-precision images, and even tiny details are pulled out using the pen tool. Many people say that the most accurate in PS is the channel, I personally think that the most accurate in PS is the pen tool, channel is the most loss of detail of a method, of course, it is best to combine: Use a pen to pick up the more complete part of the diagram, the channel is responsible for more fragmented parts




Let's give a practical example to illustrate. Below this picture background and the character main body difference is quite big, is suitable to pick out as the material, here is not for the performance technique, but serves for our need.



First we need to use the pen tool to pull out the larger pieces of the picture in the complete area.

This is my path map, use the pen to use the fewest nodes to draw a picture. And a little more important. When we use the pen tool, it is best to set the display of the picture preview box to 25%, 50%, 100% or even higher, I generally enlarge the picture to 300-400%, why? The main purpose is to avoid the visual error when drawing.




I set the display mode to 66.67%, the circular edge of the image is very bad, in fact, the sawtooth does not exist.


We first enlarge the image to show, such as 300-400%, so easier to pick. Select the Pen tool and find a place in the image to set a location first. And then find a second point, if it is an arc do not release the mouse drag can be pulled out of two adjustment rods, drag out the two adjustment lever release the mouse, press CTRL, you can adjust the bar adjustment, The average skilled person in this step can make the required radian; press ALT key to adjust the individual adjusting lever without affecting the other adjustment lever, as shown below.


Careful use of pen painted outline, this time there may be not easy to observe the error, but do not worry, we can be in the end after the observation has no excess edge, and then adjust. This step may be some people know how to use the channel to pull the map, to the channel to select a main and background contrast a large channel to adjust, so you can make a good choice, but to adjust repeatedly. In fact, there is a better way, that is image/calculations (image/operation), the following figure.



Image/calculations (image/operation) is a good way for us to make most of our work in the channel, and for us to create a new alpha channel automatically, we can save a lot of work time, and more than directly using the duplicate color channel to adjust better! Click the operation, PS will pop up a large dialog box, let us set the operation. Important to choose the reverse image, so that more easily produce a strong contrast, for this picture I set, each picture settings will not be the same, which requires your own slowly debugging to their satisfaction, as shown.



Here is the resulting alpha channel:


The left image below is the channel panel diagram, and you can see that PS automatically creates an alpha channel for us in the channel panel, so that we can copy the channel and then adjust it, as in the middle diagram.

Because we're going to have to adjust the channel later, so we're going to copy the same file now so we can compare the channels and choose image/duplicate (Image/copy), like right.






So PS will pop up a dialog box, is to let you start the file name, because we copy the file is entirely to make the channel selection, so any name can be, and then PS will pop up a copy of the file, the following figure:


Adjust the position of the two pictures, so that we can copy the file side of the channel adjustment, where the real where the virtual will be at a glance.

Now go back to the original image of the channel (ALPHA1 COPY), from the figure can be seen on the body of the tape and background are bright colors, almost no difference. The path we just made comes in handy and converts the path to a selection in the path panel because the background is white and we want to separate the characters and backgrounds as much as possible, so we fill the selection with black. The following figure:


At this stage, there may be a lot of beginners are still confused, why do this? Why completely fill black? Why make it a black-and-white color? Why...... In fact, if there are any of these problems or have not yet understood the channel, it means that you are still addicted to the layer to your convenience, we do this in the channel only one--to get a precise constituency.

When you open a picture if your picture is in RGB mode, in the channel panel you can see four layers of similar layers, and here we call the channels, and you can see from the preview that the first one is colored, and the name is RGB, which means that this is a three-color channel; The following three are all gray, The name is R, G, B, these three channels are stored in the picture of the red information, green information and blue information, they are presented to you in grayscale, as in the first chapter of the school color, you casually change the value of a channel, that is, change the channel's tone, the channel information will change, So the color of the entire picture will change, this is the color function of the channel!

What can the channel do besides changing the color of the picture for us? In fact, the most powerful function to be a constituency, the precise selection for us to establish a gray gradient selection and fragmented complex selection, and in the channel to make a constituency, unlike in the layer of the simple use of magic wand, lasso tools and pens and so on ... It's not just the channels that make the selections, but the most exciting thing is, the white in the channel, that is, the selection, gray is translucent, and black is not the part, which greatly expands our previous selection in the layer of the simple, means more rich, you can use the pen in the desired selection of white spray, Or to spray black in the place where it is not, you can also use the color bands, the curves and other commands turn the gray part black or white, so you can use the color bands or curves to make the selections we want, and even you can use burn tools and Dodge brightening tools to make selections, and use them with pens, rigging, magic sticks, etc. Of course, you can also use some of the filters in the channel to make some strange selections, which expands the function of PS.

Channel can be said to be the core of PS, almost all complex things in the production can not be separated from the channel, although in the new version of the PS layer function enhanced a lot, but the charm of the channel is irresistible. If a friend still does not understand the words, you can simply remember this sentence: In the channel to do a constituency, White is a constituency, black is not, gray is translucent.

Go on with our tutorials, look at the figure, now we want the figure, but now the background is white, and the characters are black, according to the above, black is not, white is a constituency, then we just do the reverse, this time we just use Image/adjustments/invert ( Image/Adjustment/anti-imaging) ctrl+i. So the black part becomes the white part, and the white is black, and this is the constituency we want.





But now take a closer look, the background is not pure black, but dark gray, if you now extract the selection, but also retain the background is only translucent, and the hair is also gray, so we need further adjustment, using ctrl+l or curve adjustment can be.
I'm using a color-level adjustment, carefully dragging three sliders to make the background black, and the main body into white, this time we have just copied the document to adjust, you can find that the part of the hair is very virtual, there are many people use the color level on the adjustment, and finally the virtual place to adjust the very real, This makes the constituency will be very blunt, especially when the hair loss of movement, we can use ctrl+l to adjust it, we can also use the pen to make the background into pure black.




Now the background and the main difference is very big, the background is already pure black, but in a look at some parts of the hair or not, this time you can not use the color level in the continued adjustment, because if you adjust, the hair tip will become very solid, and then do the selection to pull out will be very false, this time we can use spray pen , highlighting and other tools for deep processing, for large areas can use spray pen for spraying, but it is best not to use the brush with feather edge.

For the more virtual and fragmented place using the highlight tool to smear, if the smear too much we can use the deepening tool to darken, especially in the hair section more attention to this if it is too bright, we will pull out the edge of the picture with a background color, Therefore, it is best to use the tool of deepening when adjusting the hair tip slightly spit a little bit rather than let it have a background color, this is a compromise approach. As shown in figure:


The final figure is as follows:


I used the highlight tool, burn tool, pen tool repeated adjustments, to a certain extent to the layer to pull out to see what deficiencies, so repeatedly compared.

Press CTRL to click on the channel and you'll find that you have a selection on the channel map. Then go back to the layer and click on the picture layer, and then use CTRL+J (copy the current selection to create another layer in the original location) so you pull out the main body, but don't worry, this time to pick out the possibility of a flaw, We're going to have to look at it to finalize it, and we're going to create a new layer underneath the layer that we just pulled out, to fill in the white, because the original background is dark so we have to fill the bright color to see if the edges are residual? Look at the layers panel.

There is also mentioned in the use of the pen when there is no very accurate to pull out, this use of the edge of the body if found to have remnants of the edge we can go to the path panel to select the path and then continue to adjust the path, think that the appropriate time to observe the character's hair is very stiff or If there is, then on behalf of you in the channel adjust the color level of the past, the virtual place to be solid.

Then we went to the channel to remove the newly created channel, then copy an operation of the channel, in the just adjust the path to the selection, fill the color, start over again, and then in the layer to pull out, remember to click on the picture layer, rather than you pull out of the layer and white fill layer, This way you will make a layer on top of the white filler layer.




Then we drag the layer down to the white padding layer and hold the mouse up. Then hide the last time you did the selection, and then see if you pulled the map better than last time, if there are remnants or to the channel to remove the channel from scratch until you think it is perfect. We can also use mask for unwanted places to delete, or with the background of the place can also use mask to virtual, and finally the following figure.




Hint: Some friends may adjust the channel when the character's hair tip part of the adjustment is too real, will appear in the following picture:



You can see the edge of the hair has a very clear background in the inside, this requires you in the channel in contrast to copy the image to carry out a number of details on the trade-offs, discard some details to achieve the overall effect, the hair tip adjustment of the virtual point can be. This is a need for attention.
This time we can easily add background and special effects to the picture. A brief introduction to the beginning of the cover of the production process, first in the pick of the character layer above the text, arrange the hair in place, and then delete the font layer, and then press CTRL click on the pick of the layer, so that you made a just selection, and then use the Eraser tool to erase the place, you can safely erase, Because this time eraser only works in the character's constituency, outside the constituency is not to be erased, as shown:



Through the above practice I believe that we should have a little understanding of the channel it? When you understand that White is a constituency in the channel, I think you can use the mask more easily.

Learn to use PS Channel to pull the map I think there is a great advantage, for the plane, and even film and television is also very useful, in doing film synthesis when you can even put a film out of the TGA format and then in PS frame a frame of the object to pull out to achieve realistic picture effect, although this is a huge labor, very time-consuming , but the things will be very good, after all, is a part of the work of the fruits.