A few days ago, my best friend came up and said he had a friend to play withDance VideoIt's now more than 300,000 years ago
But I can't dance. Ask me how I can make my own dance video
It's easy to look for videos like this, so you can't just go up to millions
Well, it's easy for me to see
JUST RECENTLY, A SUPER-UTILISTIC AI TOOL, WHICH ALLOWS FOR EASY ROLE REPLACEMENT, HAS BEEN DUG UP. USE THE GENERAL, 30 SECONDS EASY。
I'VE BEEN WORKING ON IT, AND I'VE GOT A LOT OF A.I. SOFTWARE TO DO THAT。
Take a look at that video with me today
Catalogue of articles:
- Tool Analysis
- Operation Presentation
1. Tool analysis
GENERAL MANSION IS A CORE MEMBER OF THE ALARION MASTER MODEL FAMILY AND A FULL-FORM AI GENERATION TOOL FOR FOCUSED IMAGE AND VIDEO CREATION。
IT EMERGES IN THE AREA OF AI IMAGE AND VIDEO GENERATION, WITH CORE STRENGTHS FOCUSED ON FOUR DIMENSIONS OF TECHNOLOGICAL POWER, OPEN-SOURCE ECOLOGY, LANDSCAPE ADAPTATION AND CREATIVE EXPERIENCE。
What we're using today is its digital human function, the Wan2.2-Animate model, which is a more effective action generation model that supports both action mimics and role-playing, moving the action of a video role to a static picture role, replacing the role of a post-visual character, and perfecting the new role into the original video environment through photo-reforming。
In addition, the Wan2.2-S2V model produces a minute-scale film-based digital person video with a static map and an audio。
Compared to the imitation function of a dream, the most fragrance is that it's completely free of charge, regardless of points! It's extremely friendly for the little white boy who just entered
Let's see how it works
2. Operational presentation
Step 1: Find reference video
First of all, we're going to find a short-flowing video on a short video platform, which can be dance, oral, clip, video, etc。
Step 2: Prepare to replace the picture
When preparing to replace a picture, care needs to be taken that the ratio of the replacement to the original video is as consistent as possible and has a clear face, which increases the probability of successful replacement。
Special attention is paid to the fact that if the image is multiple and the person ' s face is not clear, it is easy to replace it。
Step three: Replace with a common one
And we can see here three things -- mouth-to-mouth, role-replacement, action imitation。
• Spectrum: According to the upload, the character's mouth is matched to the case。
— Role replacement: Replace the person in the video with a new role while maintaining the original video background。
- Action imitation: while keeping the new graphic elements unchanged, the characters in the picture are imitated by reference video action。
If it's a dance video, we can use a role replacement or a motion imitation, and you can just choose according to your needs。
But if it's a video clip, we're going to use the role replacement, so here I'm going to use the role replacement。
Once the target video (reference video) and role image (replaced character) is uploaded, it can be generated
That's it. The rest is waiting。
Because the tool is free, it is created in a long queue. I'm running a video in about 7-10 minutes。
If you want to be quick, you can fill the membership. But I think it's more fragrance
Just a moment, and we'll have a video
It should be noted, however, that only 2-15 seconds of video is currently supported。
If the reference video is longer, it can be repeated in a clip of 15 seconds before it can be edited and then remixed with a clip
Summarize:
It's super simple. Step: To find a video-preparating role picture that wants to be rewritten, and to recaptulate it with a common meaning, the whole recapturing action, not counting the time it's waiting to be created, and not for 10 seconds, it's extremely friendly for the rookie
But there is still something to be noticed:
- To the extent possible, replacement of role pictures is consistent with the original video and has a clear face, which increases the probability of success
- Only 2-15 seconds of reference video is supported for upload, and if the video is too long, it can be produced in a clip and then collide
- Free versions wait longer, but I think it's okay for a little while for a fee
ALL RIGHT, BOYS, LET'S TRY IT