I originally planned to do AR video effects, and let the virtual cat sit on the real computer. Although I made another video for various reasons, I still haven’t changed my mind. I still want to make an AR video. This needs to be done in unity, and it happens that my mentor Herman Ho knows how to do this and he helped me a lot in this regard. Even better, it didn’t take a lot of time since my animation was already done and I had good models.
But there are still some problems. The biggest problem is that the hair I made in Maya’s X Gen cannot be imported into unity. And the time I have now is completely incapable of reusing new software to make new hairs. So after I import my model and animation into unity. I’m going to directly modify the overall texture of the cat. I am very grateful to my tutor Herman who recommended the tutorial for the glowing crystal material. It saved me from having to make hair for the 3rd time. I made a cat with glowing crystal material according to an online tutorial. I love pretty and bright colors, and the previous video Cat on the Cloud used very bright colors as well. I’m very happy with the effect.
The next step is to use the AR technology in unity to make the virtual cat I made appear in the real world.
First of all, to make the virtual cat appear in the real world requires an anchor point, that is to say, a picture. After the computer camera recognizes this picture, the model will be generated on that picture. Of course, the picture needs to be converted a little bit, which is very simple, and the Vuforia Developer website can do it. As long as you find this website and register, upload the picture as the anchor point, and download it after the conversion is completed. I’m very happy with the effect.
The image I chose was the black and white still frame from the keyframe capture in After Effects. The color is not selected for the convenience of computer recognition. The whole picture should not be too complicated. If there are too many colors, it will be difficult to recognize. Especially because of my own preference, I rendered the whole photo colorful and too gorgeous. If the machine is used It will be very difficult to identify, and it is easy to make mistakes after the color changes due to different light effects from the outside world. It just happened that I grabbed a black and white still frame in After Effects for the convenience of class. So I used this image that was still framed as an anchor. It is also convenient to display after printing as a postcard.
The next step is to determine the location of the anchor point. Just place the image as the anchor point below the model in unity. Then the model will be generated at the position of the corresponding picture in the real world, and the animation will be played according to the camera lens. The lens of the camera can be moved, and the model will also display animations in various directions according to the positioned anchor points. It works really well, just a little tweaking of the size and position of the model relative to the model, and the cat is sitting on the computer, which is very interesting. I managed to make the virtual cat appear in the real world.
After the entire animation is rendered, it is time to start compositing. My topic is the cat on the cloud. Next, I need to add some clouds to the scene. According to the principle of lens language, I try to have different movement directions for short stories without moving. For example, when shooting a person walking, you can first shoot him from the left to the right of the camera, and then shoot obliquely in the next shot, so that he walks from the lower left corner to the lower right corner. You can change the direction again next time. In short, it is necessary to have different movement directions as much as possible, so that the whole short film can be more exciting. My lens is a fixed lens for convenience, so the whole animation needs to increase its interest when the lens remains unchanged, that is, its movement route. I plan to use the different moving directions of the cloud combined with the moving direction of the cat to create resolve this matter. Make my short films more exciting. Otherwise, if only the cat in the short film is moving, the whole short film will be too boring.
To make my animation more interesting, I don’t intend to make all the clouds move in one direction. First analyze the cat’s action path, from the left of the video to the right of the video, then turn to his left, and finally walk to the middle of the entire screen and sit down. And the cloud I’m going to add has three layers. The clouds in the foreground above the video and below the video are moving from left to right. The cloud layer in the background below the video spreads from the middle of the video to the bottom of the video. In this way, the movement path of the clouds below the video will form an L-shape from front to back, and then from left to right. This also happens to form a circle with the cat’s action path. Of course, I also tried to move the clouds with different action paths, but if the moving paths of the clouds at different levels in the video are completely opposite, it will appear very abrupt. But if all the clouds were moving in one direction, the whole video would look very boring. After many attempts, I settled on the current solution. Background clouds on the sky and foreground below, with layers moving from left to right. Clouds in the lower background spread downward from the center of the middle video.
Added another animation after the clouds were all done. Finally, after the cat sat down, a group of birds flew from the lower right corner of the video to the upper left corner of the video. This is a slanting upward movement trajectory that is different from the previous trajectory. It also makes the boring time after the cat sits more interesting.
Next was to color the whole scene, I couldn’t make my work black and white, that would be too bad. After I arranged all the cloud layers, I grabbed the key frame of the last frame directly in After Effects, and chose to export to photoshop format when exporting, so that the work I exported will be divided according to the layering of the previously synthesized video. Many layers. This also makes it easier for me to add color when coloring the video at different levels. Of course, in the final export, not only can the work in photoshop format be directly uploaded to After Effects, but also it can save time for subsequent synthesis and modification and make it clearer.
In order to know what color is more suitable, I found some references on the Internet. It was a sky reference. Excessive, the sky is white clouds, the ground is a city, very beautiful pictures, and the colors are also very beautiful. Although the overall color looks very complicated, in fact, the entire color change is changed according to the color temperature. Put the photo and the color temperature chart together and you will see that the color shift is exactly the same. The left part slowly changes from the brightest solar white to yellow, orange and red. The part on the right is slowly changing from cyan to blue in the sky, and finally a little purple.
Even better, due to lighting reasons, the left two special effect lights, the left one is red light, and the right one is blue light. And because the cat is moving, but the light is not moving, so the video starts when the cat is on the left side of the video as a whole is reddish, and when it is on the right side, the overall color is blue. Now the video background modification can solve this problem. . The left side is yellow light, because the dark side will change the hue, so the color will be reddish, the right side is blue light, the dark side will become bluer or even a bit purple due to the hue change. These match exactly the lights I rendered on the cat earlier. It also perfectly solves the problem that when the video is pre-rendered, the color of the cat is reddish on the left side of the video, and the color is bluer on the right side of the video. However, I found that the color at the intersection of yellow and blue is very dirty, and the color transition between the top and bottom of the video is not perfect, so I plan to add other colors in the future. Make the whole work more colorful.
After adding the background, the next step was to color the dark sides of the clouds. Because the cloud layer is moving, and the part of the dark side is different in each frame, so I plan to add the color directly and use a soft light coating, so that the software can automatically calculate and add the color to the dark side. Does not change the white of bright clouds. And because only 1~2 layers are added, the final overall calculation and rendering will not be particularly troublesome. After all, I don’t have enough time.
When looking at the overall color effect, I found that the color at the junction of yellow and blue is very dirty, and the color transition between the top and bottom of the video is not perfect. The background of the whole video is only blue and yellow, which is too simple overall. So I plan to add additional colors in the future. Make the whole work more colorful. In the end I plan to continue adding colors based on changes in the color temperature table. The yellow part changes from orange to red according to the color temperature, and the blue part changes according to the color temperature. Finally, a perfect transition can be formed between purple and red. I turned this layer into a soft light layer and it worked really well, but because the color saturation was too high, I lowered the saturation some more to make it more compatible with the tone of the entire video.
When I finally watched the overall effect of the video effect, I found that the two layers of clouds below the video were not too far apart, so I added a gradient layer of shadow between the two layers of clouds. Enhanced the effect of depth of field and front-back relationship between two layers of clouds.
Finally, I directly dragged the photoshop format file into After Effects, divided each layer, and adjusted the layer settings. After adjusting the color of the cat through the curve to make the color more obvious, my work will be rendered finished.
For lighting, I use 1 white main light source plus 2 colored special effect lights. I once learned how to light photography, and in this case, with one key light and two effect lights, I could do what I wanted. But photographic lighting belongs to real physical lighting in the real world. In Maya software, the software is used to simulate light. There is a certain gap in this aspect, but it is also very valuable for reference. And if the lighting can be adjusted during the lighting process, try to adjust it at one time, because the coloring and changes in the later stage often cannot replace the effect of the original lighting, especially the special effect light I plan to use. It is almost impossible to adjust to such an effect in the later stage.
Because my cat is white as a whole, but only white color will make the whole animation look very monotonous, especially after I will add white clouds, so I have to find a way to make the overall color more diverse, Otherwise, the wife is monotonous. So I plan to challenge one of the difficult problems in the design world, the colorful white.
Contrary to the colours used in painting, white light is a collection of all colours of light and it is impossible to add other colours to white light. When painting, mixing all the different colours of paint together, black will appear. And the situation with light is the exact opposite of what happens when painting. He discussed prism dispersion in great detail in his book Opticks (1704). He shone a beam of white light into a triple prism, which resulted in the refraction of different colours of light. This means that the white light contains all the other colours of light. Conversely, by superimposing all the different colours of light together, white light is formed. That is why it is almost impossible to add colour by light on a white object. Because white light already contains all the colours, no matter how much more light of other colours is added the result is white. And in many games the overall lighting is very dark. Not only for the sake of creating a scary atmosphere, but also because darker scenes can reduce more detail. And often times to reduce memory, and to speed up the computer. So ways are found to make the whole thing darker so that lower resolution mapping can be used, rougher modelling and the player can’t notice the details of the scene. And because my work is very white and very bright, more care must be taken with the details, otherwise any flaws will be magnified.
So it’s very difficult. The cat I made has white fur, and the scene I want to add later, such as the white cloud, is also white. So if I don’t want to increase the color of the whole work, the whole work will be black and white, which is very boring. I hope my work can have more colors and be more beautiful. But if you want to reflect white, you must use white light for irradiation. If the white cat is illuminated with light of other colors, then the cat will become another color, but if you want to reflect the whiteness of the cat, you can only use white light for illumination, then this work has There will be very few colors, only black, white and gray. So it is very difficult to add color to make the white cat colorful without changing its own color. What’s more difficult is that what I want to do is that the video is dynamic, so I can’t make all the colors fixed by painting, it will look very fake. This is the reason why brilliant white has become one of the conundrums in the design world. And I wanted to challenge that by making a colorful white cat.
The solution I chose was to use special effects lights to do it all, which is an advanced trick with lighting. First use a white key light to render the whole cat white, and then use two other special effect lights of different colors to color the shadow part of the whole cat. Since there is no way to change the color of the key light, which is the natural color of the cat, the best case scenario is to color the shadows to make it more colorful. In this way, more colors can be added to the whole scene without changing its inherent color, which is white, and because it is a shadow part, it will not appear too overwhelming. Whether it is special effect light or hidden color, it is a very difficult advanced skill.
The situation where only two special effect lights are turned on.
Although the overall theoretical requirements are very complicated, in fact, only three lights are needed to complete it. First, the main light of No. 1 plus the special effect lights of No. 2 and No. 3. The more light, the better. When adding a lot of light sources in a mess, it is often unclear what effect you want, or how to increase the light to achieve the effect you want. Therefore, adding lights indiscriminately has no benefit other than increasing rendering time.
The white of the key light is very easy to adjust the value, as long as it is bright enough to show the natural white color of the cat, but also has enough shadows that I can increase the color of the shadows with the effect of light. The difficulty lies in the two special effects lights. I need to use special effect light to make the whole scene add more color to the dark side of the cat, but the light cannot exceed the key light, and the light produced by it cannot change the inherent color of the cat’s white color, which is very difficult. And what kind of lighting can be used to add more colors to show better effects, which also requires constant testing and consideration.
The last two effects lights I settled on were red and blue. These two colors are complementary and will be very noticeable together, and will have a very nice purple transition where they meet, so that there are at least three different colors. And after I adjust the brightness, it will not seize the natural white color of the subject. I have used many different colors, among which the color of the yellow light is very inconspicuous and cannot show its color, and because it is dark sleep, the light effect appears more gray than yellow. And many relative colors, such as red and green together or in the middle transition will appear very gray, which is not good-looking at all. So after I tried many times I ended up using red and blue as complementary colors for the two effect lights.
Finally, I rendered the whole animation and found something very interesting when looking at the effect. After I put the picture into photoshop or After Effects, it turned out that the color rendered by the light on the cat was not obvious. The cat is almost all white. But after I put the rendered picture into Davinci, I found that the color is very bright, which is very interesting, but it also increases the difficulty for me and the subsequent Compositing. I need to add color to the cat and cloud in After Effects Then put it into Davinci for final Compositing and export. This also meant that I was going to use the compositing software After Effects for grading, and then use the grading software Davinci for compositing.
Fram 700
There are still some small problems in the final rendered video, because I refer to the real-world photography lighting, but my work is an animation, the cat in the video is moving, and because it is the light calculated by the software . So after the final rendering, there are still two small problems in details, but the impact is not big.
First of all, when the cat turned around in the middle, because the left side of my two special effect lights was red light, and the right side was blue light, so the cat was on the right side at that time, so there was too much blue light. For this problem, unless the light moves with the cat, or the blue part is slightly reduced in post-grading, it can have a very good effect.
And the second small problem is that after the cat finally sat down, the light spot in the right eye was too bright. And this problem, so can not be changed. Because the reflected light spots in the eyes are the reflected main light, if the size or intensity of the main light is changed, it will have an impact on the overall brightness. If it is a real world light, then the overall brightness can be lowered. Change the main light source to ambient soft light and then add special effect light. Because in real-world photography, the brightness is not fixed and the absolute requirement is the light ratio. For example, the light ratio between my main light and my two special effect lights is about 10:2~10:1. So as long as the light ratio is kept constant and the overall brightness is reduced, the same effect can be produced. But because it is in the software, the intensity cannot be reduced, and because it is the main light, if the main light is not enough, the whole will be very dark, and the special effect light will overwhelm the guest and change the overall color, which will be even worse. This is also the difference between software lighting and real lighting.
Unfortunately, at the beginning of September, my hair reached the neck and the most complicated head was done, because the Ragdoll is a longhair cat, and its cheeks will be longer and the inside of the ears will also be It has long hair, so it needs to be made in 3 to 4 times. That is to say, it takes 3~4 strings of nodes to complete the hair production of the head. The back part of the body and tail is very simple. I had expected to finish making the hair by the end of September. But the school’s computers were overhauled in September and were not open for use until the end of September. And by the time I started using the school computer at the end of September, the Houdini software was not found in the library of the school computer. Although Houdini can be seen in Curry later. However, other students who use Houdini said that the school’s Houdini is not stable enough. Some computers can be successfully installed, some computers cannot be installed, and some libraries are not even available. So I had to change the software that made the fur and recreate the cat’s fur. After considering different software, I ended up going to use Maya’s X Gen for the hair. Since my animations were made in Maya, it would be easier to continue making the fur in Maya without having to worry about any compatibility issues. And luckily the hair making in Maya is much simpler than Houdini and can be done on time.
Due to time constraints, I didn’t draw the textures on the cat completely according to my cat Moqi. I directly used the whole cat as a white cat to complete my work. However, I modified the color of the cat’s eyes, nose， mouth and paws. I made the cat’s eyes blue because the eyes of the puppet cat are blue, and the bluer the cat, the better the appearance, and then Moqi’s eye sockets and a circle around the mouth, and 4 paws changed the paw pads to pink. Because I had some problems with Houdini, the software I used to make modules, I temporarily switched to Maya’s X Gen for hair in less than two months. This didn’t take me long enough to finish the texture, so I simply changed it to a white cat instead of continuing to texture my cat Moqi.
