High school began to learn painting, and the modeling teacher always emphasized composition, proportion and perspective. The perspective mentioned by the teacher is divided into one point (parallel), two points (angled), three points (overlooking or looking up), scattered points and so on, which was difficult to accept at that time. But I can't find a more effective way to solve the two-dimensional virtual three-dimensional space. Every time I finish painting, I can't help but feel that the picture is dull and I can't fully express my feelings for the object.
When I was in college, I read a lot of biological materials in the library and made an in-depth study on the principle of human eye imaging. I took off my nearsighted glasses, naked eyes, picked up my binoculars, stood on the tallest building in the school, made numerous experiments and observations, and finally got some experience in cambered perspective. Using this method to draw scenery, still life and scenery has achieved good results. Although it is still not immersive, the real touch has made some progress.
After work, I have the honor to learn ink and wash landscapes with my teacher. When I was a child, the muzzle, trenches and hills appeared in front of me again. The majestic Taihang let me go out early and return late again and again, just to touch the pen with a pen? Look up at the sky and look down at the wall? Magnificent, listen to that with your heart? Thousands of miles of monuments are constantly looking forward to it, and the wind and rain are blowing relentlessly? Generous, but no matter what artistic conception, it is always difficult to show Taihang. The gorge turns 19, Qifeng confronts? Real feelings.
After watching Avatar, I deeply studied the time difference shooting principle of simulating eyes with two lenses, the imaging principle of polarized vertical film played by double-sided projector, and the working principle of some software and chip-controlled electric glasses with the function of synthesizing 3D images, from which I gained enlightenment. The simple understanding is to record the images seen by both eyes respectively, and then transmit the images seen by the left eye to the left eye only, and the images seen by the right eye only to the right eye, so as to realize the 3D stereoscopic synthetic image function of the human brain. In theory, this should be consistent with the images obtained by people's fixed-point observation.
The difference is that movies can take advantage of the time difference of 1/24 seconds left by human vision, and the left and right projectors work alternately, so that the images people see are still concentrated on the same screen, while static painting can't take advantage of this time difference. So at the beginning of the experiment, I can only take photos of the same group of objects in the left and right eyes with the help of a single-lens camera (in order to make the experimental effect more obvious, I chose an angle slightly to the left and right than the left eye). In the process of observation, an opaque partition is always placed on the middle vertical line of the eye connection line to separate the field of vision of the eye. It's a bit like a blind spot in a horse's eye, but it doesn't affect the experimental results. The results show that the stereoscopic effect of brain synthetic images even exceeds the normal observation results of people. I think it is probably caused by the wandering of two eyeballs when human eyes observe objects.
With such experimental results, I further affirmed the feasibility of binocular perspective theory research. The next question is how to combine the imaging of the left and right viewpoints. After simple computer superposition processing, there will be ghosts as vague as taking off special glasses and watching a curtain 3D movie directly. Of course, what we want is not the kind of graphic art that we can appreciate only with special glasses, so I have done a lot of experiments with the confusion of how to synthesize the left and right viewpoint pictures.
A series of experiments ended without results, and finally had to return to the starting point of plane perspective. When I was a beginner in perspective, because I couldn't find the focus, the teacher often urged me to close one eye and measure the direction of the still life structure line with a pencil, every time I closed my eyes. This is actually a single-focus perspective, just like the lens of a single-lens camera. What I originally wanted was realism, which has nothing to do with where the focus is and where the visual center is. This also reminds me that the visual center (focus) marked in the schematic diagram of cube Parallel perspective is mostly outside the still life. The definition of visual center point is: the vertical center from the viewpoint to the apparent horizon, that is, the vertical foot from the viewpoint (human eye) to the apparent horizon. The line of sight is defined as an imaginary line between the viewpoint of light with the human eye as the vertex and any part of the object, which extends without being blocked. The apparent horizon is defined as a horizontal line with the same height as the human eye. We might as well sit down and demonstrate: if the viewer is still, the point of view remains the same; Three objects are placed in front of the viewer's left, his front and his front. When observing three objects, there will be three directions of sight and apparent horizon. In linear perspective theory, the compass direction of eye level is not strictly defined, only the height and level are clearly defined. That is to say, the horizontal line in one direction will always be perpendicular to the line of sight, so the visual center point will certainly appear between the extension lines of the tangents of the left and right silhouette of the object, which is equivalent to looking directly at three objects at the same viewpoint and getting three independent perspective images. According to this result, I tried to use this perspective method of moving focus from the same viewpoint in Sketch of Taihang Mountain, which broke the original composition method of grouping single viewpoints and achieved a more realistic effect of two-dimensional assumption of three-dimensional space. In the process of surveying and mapping focusing, I don't keep one eye closed all the time, but take the method of alternating left and right eyes and finally taking the midpoint. This method can prolong the sight distance, just like artillery measuring sight distance with their thumbs, and get more accurate distance.
Of course, this method of measuring the sight distance alternately with both eyes, moving the visual center point (focus reference) of the same viewpoint within the overlapping sight distance of both eyes, and taking each object or each distance as an independent reproduction object for independent line perspective has not yet reached the final stage. I will continue to use experiments to demonstrate the feasibility of this method. Let's call it the perspective theory research of simulating the principle of 3D imaging, and we are willing to discuss it with friends who are also pursuing the three-dimensional authenticity of plane reproduction.