I recently saw a lot of people recommending GPT Image 2 stories。
A picture of everything can be done, and all video messages, such as character setting, scene design, spectrometry, mirrors, lights, colour boards, etc., are directly marked。
Match Seedance 2.0, you'll be able to make a direct and efficient film, free to produce
But does this combination work? How's it going
Today, I'll give you a full picture
This time I'll useLiblibthe LibTVIt's got Image 2 and Seedance 2.0 in the same stream, not to be used to cut back to the platform, with a variety of practical tools, and it's full of efficiency
Official website:https://www.liblib.tv
It's been a long-standing national unicorn on the AI photo track, and this year it's going into the video track to launch LibTV. The short play companies are now using it to pay millions of dollars a month
IT'S NOT AN EXAGGERATION TO SAY IT'S THE BEST IA VIDEO PRODUCTION TOOL IN THE COUNTRY。
1 Storyboard in place
Before that Banana 2 could do a nine-gauge spectroscope, but it had only images and no lens information。
Image 2 is a step-by-step spectroscope design and generation that can directly produce a professional storyboard。
There are various Image 2 storytips on the Internet, but they're all the same. Once I know the structure, I've just sorted out a template for the hint:
All you have to do is simply provide your own creative ideas, for example, and I want to produce a video about blonds dreaming。
As can be seen, the storyboards generated are complete, the spectroscopy is logical, and the roles are well aligned。
AND HERE'S THE POINT, THE RESOLUTION OF THE STORY PAGES MUST NOT BE TOO LOW, AT LEAST 2K TO ENSURE THE QUALITY OF THE VIDEO IN THE BACK, OTHERWISE THE LITTLE WORDS ON IT CAN EASILY BE CONFUSED。
If it's not clear enough, LibTV takes a high-intensity magnification, and it's used。
2-Video One Straight Out
Then we upload the storyboard to Seedance 2.0
Because all the footage is clear in the storybook, the video message here is short:
💡
Generate video by the content of the story
It's really good, but there's a lot of problems with the process。
Three, the problem
First, because the storyboard contains too much information, is prone to error and, if it does, it is difficult to modify it precisely。
When it's not careful, the video that comes out of the back goes down。
For example, the spectroscopy of the pen sharpener, the product is clearly deformed, so it can only be scrapped and redrawn。
But the regenerated storyboard, which contains video messages, changes, and it's difficult to control it precisely。
AND IT'S AN ALL-EMBRACING STORYBOOK THAT, IN TURN, LIMITS THE USE OF AI。
SO LET'S CHANGE OUR MINDS AND LET'S NOT TAKE AI TOO SERIOUSLY。
FOUR, UNTIE AI
We can pre-set the basics of roles, scenes, styles, and the storyboards only keep the central spectroscopy and lens description。
Would it have worked better
The same thing with dogs for daydreams:
This time I'm just giving Image 2 a picture of the dog and the direction of the story, not with that complex frame word on it。
Is it clear that this video is better than the first edition
It's a better picture and a more comfortable narrative。
And here's one thing, sometimes the video that we create, like this dog video, the middle and the last picture, there's some unwanted text messages。
In this case, you don't have to change your software to cover the screen, and you can use LibTV's smarts for subtitles, and you can erase the house, which is very practical。
THE MORE FREEDOM WE LEAVE TO AI, THE MORE SURPRISE IT WILL SURPRISE US。
This time, I took a keyboard picture of the desktop, and Image 2 helped me to think and create a spectroscopy of the ads。
And then I sent the storyboard to Seedance 2.0, once generated。
A rather nice keyboard product short, and it came out。
Conclusions
Overall, the conclusions are clear。
Image 2 story editions, with Seedance 2.0, do make a good film。
But don't fantasize that by a storyboard alone, you're gonna get high quality AI video generationit's all bugs and cards。
In particular, the video of the story is complex
After identifying the content of the role, scene and style, the Image 2 output version ensures consistency and consistency。
Actually, if you want to learn an AI video, or you want to find inspiration, just brush LibTV。
Now, the AI video content is raging, and when I brush my voice, it's almost every third IA, and I find that a lot of it is made of LibTV。
LibTV has accumulated a great deal of high-quality work, and it's played a million times in tremors。
Even better, you can see people's complete creations directly here, and it's easier to follow。
And LibTV did a very new Creator+Agent double-entry design in the industry。
IN SHORT, YOU CAN DO IT ON YOUR OWN, AND YOU CAN DO IT IN A FRAME, AND YOU CAN DO IT IN HIGH QUALITY, AND YOU CAN GET AI'S ASSISTANT TO HELP YOU DO THE WHOLE PROCESS FROM SCRIPT TO FILM。
THIS PATTERN OF PEOPLE WORKING WITH AI FEELS LIKE THE REAL DIRECTION OF FUTURE CREATION。
website: https://www.liblib.tv/
glithub: https://github.com/libtv-labs/libtv-kills