Over the past year, our team wasAI VideoTHE FIELD SETTLES A WHOLE SET OF HANDS-ON PROCESSES. TODAY, AN AI VIDEO INDUSTRY EXCHANGE WAS ALSO HELD IN ORDER TO SHARE THE MOST UP-TO-DATE WORK。
TAKE YOU THROUGH THE WHOLE IA VIDEO, FROM CREATIVITY TO THE PRODUCTION CHAIN, AND TALK ABOUT THE USUAL TOOLS AND TIME/COST DISTRIBUTION。
WHAT CAN A.I.V. DO
COMPARED TO THE VIDEO NUMBER OR STATION B YOU USED TO USE, MANY OF THE BLASTS ARE ALREADY AI-GENERATED。
What we're going to do is to tear down the process that adults can use。
Overall process: text graph → audio → cut
We call it Five-step "Video-Screen" method.
Man (creative creativity):Artificial and then refined with the ChatGPT, Gemini, Claude, etc. Output of an implementable script text。
Script:It's a little bit of a scene: role, scene, transition, rhythm. It is managed in a single form to facilitate multi-person collaboration。
Style and character setting:The role-finals are generated using tools such as MidJourney or Dream. Ensures consistency of style and, if necessary, output of three views and reference maps. Now we can combine with Nano Banana
Figure:IMAGE MATERIAL BY LENS. THIS IS THE MOST TIME-CONSUMING SEGMENT (APPROXIMATELY 301 TP3T-601 TP3T TIME)。
Video Generation and Edit
Tools: Dreams, Vidu, VEO3 etc。
- Generate short clips (commonly 5 seconds), then select, edit, and harmonize the rhythm。
Join the sound, play, finish the final film。
Common tools and features
- It's a dream: To cover most day-to-day needs and to generate efficiency。
- whelks : Good at fighting special effects, action-type shots。
- Vidu: better suited to the binary style。
- MidJourney / Dream: character design and style。
- Multimodular Platforms (e.g. inspirational canvass): Future trends that link images, videos, audio。
Small trick:
- You can read more official documents and specify the elements of the hint。
- To be familiar with the characteristics of the different platforms, it is possible to achieve “what kind of lens is used for which tool”。
Time and cost distribution
The time distribution varies widely among projects with different subject matter. For example, a three-minute piece:
- Script writing:10%-20%
- Mirror Design:10%-20%
- Diagram:30%-60% (MOST TIME-CONSUMING)
- Video Generation:20%-30%
- Cuts and plays:20%-30%
Cost elements:
Most of the tools are subscription-based, with prices ranging from hundreds to thousands of yuan per month。
Video links often cost most of their cost。
What really determines the quality is the solidity of the front script + spectroscopy, not the last video model。
Our observations and our thinking
- Process standardization IT'S THE KEY TO THE SCALE OF AI VIDEO。
- Role/style Consistency Decide whether it's professional or not。
- Mirror = tone: Good work has been "stereotyped" before the video。
- The tool is being updated very quickly: The sequencing will change and more comprehensive integration platforms may emerge in the future。
THE AI VIDEO HAS MOVED FROM "SHOW TOYS" TO "DELIVERABLE PRODUCTIVITY"。