How many steps are there in CG special effects synthesis?

Special effects movies have won the audience's enthusiasm with their unique visual arts and exquisite creative skills. Special effects editing is an indispensable element in the plots of many sci-fi masterpieces, which can sublimate the appreciation level of the whole movie. But most of us only see the display of special effects, but we don't understand the details and hardships of the production process. The following is a brief picture of the production process of special effects movies.

R&D plays a major role in the production of large-scale special effects movies. The R&D department is mainly composed of scientists, programmers and mathematicians, and provides plug-ins for existing special effects production software, such as Maya and Nuke, or some independent software for continuous technical research and development. For example, Avatar uses a large number of virtual character animations captured by performance, and the newly developed virtual camera can watch the preliminary synthesis effect at the capture site.

1. Model making: Models are divided into solid models and digital models. The three-dimensional model is made into a digital version by a three-dimensional producer. Digital models will produce different levels of faces. Fine high face model is used for final rendering, middle model is used for animation, and low face model is used for visual preview. Now large-scale special effects movies need low-profile models to make visual preview stories. The drawings made after conceptual design or the objects provided by the art department are the standards of 3D production, and sometimes solid models can be directly made to prompt or scan the production of virtual characters and important props.

2. Motion capture: an important part of special effects movies. Before Avatar, motion capture was mainly based on body motion capture, which was usually used for animation of medium and long-term virtual role actors or monsters such as dinosaurs and gorillas. Later, "Rejuvenation" tried to make highly simulated facial expressions of characters, so the technology of motion capture began to move towards "performance capture".

3. Motion tracking and matching: lens tracking starts immediately after film scanning, and the first thing to do is to reverse the lens trajectory. Using 3D tracking software such as BOUJOU and PFTRACK, it is very useful to shoot the lens parameters recorded on the spot at this time. However, whether the default tracking function of the software can cope with the difficult shooting action requires the tracker to manually adjust the parameters or use the newly developed project-specific tracking software to operate. After the accurate lens tracking is completed, the camera trajectory will be sent to 3D software or 2D synthesis software. In addition to the reverse of the lens trajectory, it is also necessary to track the motion trajectory of objects (characters, props, etc.). ).