I believe this is an easy-to-understand tutorial for big and little kids on the whole process of breaking down a video script idea to video landing. This post wants to try AI inskitaspects of the application, just to throw some light on it.
I. Elements of creation
ancient Chinese state near present day ChongqingskitThe creation of this work is similar to the previousAI VideoThe creative elements of the class are basically the same, but the content of the story and the language of the camera need to have higher requirements. If you want to make the AI short drama similar to the texture of the movie is more difficult. I hope to improve a little bit in each of my attempts and try to avoid the PPT feeling.
This attempt to create a micro-skit may not be the same as the popular sense of the content, leaving aside the sense of fun and funny. The theme is "survival", and the story is centered around three main characters: a deer, a wolf and a human. The difficulty in creating this story was maintaining consistency across multiple characters, which was easier with the animals I chose than with the characters. I had a rough idea for the story, and then optimized the script and subplot through a large language model.
In line with other AI video creation elements, the main ones include"Script Creation" "Visual Graphics" "Graphic Video" "Audio Creation" "Post Processing"and other aspects, which are now among the basic elements of AI video-type creation.
II. Planning and preparation
This creation is based on the theme of "Survival" and centers around three main characters: a deer, a wolf and a man. The general idea is that in the early morning of a winter's day, when it is snowing, a buck is struggling to find some food and is eating berries. Suddenly noticing a few wolves in the distance ready to attack, the buck runs quickly through the forest and escapes, with the wolves in hot pursuit. After a fierce battle, the buck finally manages to shake off the wolves, and just when the buck thinks he has escaped the wolves and can survive, a gunshot rings out. The last scene is a family sitting around happily eating dinner, the buck's head is hung as a decoration in the house, underneath is a photo of father and son holding the deer head in their hands, smiling happily.
With this simple idea in mind, give it directly to GPT or Beanbag to generate a creative script based on my content. Include subplots, etc., and revise and refine it according to your own ideas to finalize the entire story script.
Improve the split-screen form to identify screen descriptions, shots, dribbles, and sound effects to make it easier to clarify the idea of video creation.
III. Creative production
3.1 Visual component
Once the split screen script and split screen form had been created, it was time to start working on the screen. In the beginning stages there was no certainty as to what style would be used, so different style models were tried out, and ultimately the choice was made to use a real feeling model.
3.2 Background music and sound effects
Use Suno to generate the background sound. If you don't need lyrics, open "Instrumental" and enter the melody you want to generate. For the sound effects, you can find them on the audio website, or search for them in the final cut and add them.
This time, the music part of the attempt more times, but did not generate a piece of background music can be completely adapted to the video content, is generated after the music, through the post-production editing splicing of the background sound, the sound effects of the part of the direct use of the sound effects in the screening cuts of the material.
IV. Post-integration and optimization
Carry out the actual production according to the pre-determined sub-scene script form. Import the generated material and background sound into the video editing software, edit, splice, add special effects and other operations to make the video smooth and rhythmic.
Finally, the video cover was created, highlighting the theme of "survival" through the contrast of warm and cool colors and the expressions of animals and people.
After completing the creation of the entire video, you can use tools to enhance the resolution and frame rate of the original video in order to make the picture clearer. Using Wink, frame fill the video to enhance the smoothness of the entire video, and then use lossless zoom to enhance the resolution to 2k or 4k.
V. Conclusion
In this AI creation, more effort was put into the scripting part of the story, hoping to show the multi-dimensionality of survival through this story: bucks look for food in order to survive, wolves look for deer as food in order to survive, and their hunting behaviors are purely driven by the need to survive, with no additional greed or malice. Humans, on the other hand, hunt more for entertainment or decoration than for "survival". Later on, I felt that it would be monotonous to simply tell the story, so I chose to use the scene of a father telling a story to his daughter, and gave people food for thought through the little girl's final question.
Comparison of surprise is with the AI technology to generate such animal fighting scenes, the actual generation effect is much better than expected, just in the birth of the map will occasionally appear wrong, and sometimes there will be a wolf with antlers or a wolf face of the deer, you can try many times to draw a card, or manually synthesize the PS. Compared to the reality of shooting animal fighting scenes, need a professional team to follow the shooting, long time squatting animal habitat, subject to nature and animal habits and many other factors, time-consuming and labor-intensive and difficult high cost, AI directly generate pictures and video is much more convenient.