A unique program has emerged in Russia — a digital operator and director 2 in 1. The software was created by domestic developers in the "Automation" department at the video technology workshop of the Higher School of Economics.
The system is a virtual operator and director, so the process is divided into two parts. The program operates autonomously and independently focuses on the figure of a person in the frame. Settings help to focus alternately on different characters in the story, as well as return the camera to the desired position if the person is "lost."
You can set parameters under which the AI will automatically observe the dynamics of movement and composition that are correct for video shooting.
You can also control the shooting process remotely. The digital operator with integrated AI understands voice commands. You can ask it to change the angle of the camera or shift the focus to another person.
All that is needed to work — a specialized video camera that supports remote control of the shooting process, and a powerful computer on which the "Automatic Director" will be installed.
Our software is written in Python and is already fully ready for use. In particular, several events were filmed on the university site, such as presentations and lectures.
Will AI replace directors and operators?
According to experts, this program will be successful among bloggers who work in the genre of informational interviews, but do not yet have the opportunity to hire a full staff for video production.
The informational topic is a key point here, as the AI will be focused on changing the camera based on the movement of the characters.
Also, the "Automatic Director" will easily leave operators and directors of sports broadcasts and events where multi-camera shooting is used without work, but is unlikely to replace a living person in the film industry at this stage.
The method is not suitable for people whose task is to create a multifaceted image of the hero, to show his psychology. After all, in this content, the viewer is interested in watching the reactions and emotions of the person in the frame — to questions, statements. And at such moments, the hero is often silent, his facial expressions speak for him, which is difficult to read even with the most sensitive motion sensor.
Earlier www1.ru reported that the Kandinsky neural network was taught to create videos based on text.
Read materials on the topic:
Russian online cinema KION has launched a neural network-based service for film recommendations
Russian engineers-programmers taught the robot arm to "think" AI