The last few days have been spent figuring out how to makeDigital Humanfrom, for example, using a third party such as a Zen mirror, i.e., a dreamAI Digital HumanThe first thing I did was to find out that a lot of the problems were either costly or required a lot of points, etc. I couldn't afford to use it at all, and if I needed to orate or sing for a few minutes for a long time it was really very troublesome, and the result was not the most desirable state. Here again to ponder for a long time, found a good use of AI software: DreamFace.
Dream FaceThis is an app that uses AI to create oral videos:https://www.1ai.net/30947.html, a solution for creators to provide a full process of creation. Let photos speak and sing, easily, online has appeared to break the permanent version (Android version), unlock all the privileged features, unlimited use.
App directly download APK to install on your phone.
Once installed, it will start with a delayed screen and no associated features. Wait a little while will or back off and refresh a few times.
After showing success.
Click here for our video of the Digital Man.
Before uploading a video digital person, we need to generate a digital person using an AI tool, here we will use the previous guitar girl as a case study.
Step 1: Make a digital person
Here's the i.e. dream out chart used:https://jimeng.jianying.com/
Cue word: An 18-25 year old Asian girl standing on stage, holding a guitar, standing in front of a microphone, wearing a floral patterned dress with a clean background. Ultra HD, 8K resolution to ensure detail. The style is more realistic, but with a touch of dreaminess
Here you can upload your own photos to refer to the raw image
Choose the character you are satisfied with for HD HD, canvas exported without watermark image. (I won't show the process here)
After getting the image, for the MV effect, we need to get the character to move first so we get the hand motion that pops him up.
Here's the Koring out video used: https://klingai.kuaishou.com/
Enter cue word: beautiful woman playing guitar slowly with both hands. Fixed shot.
This way we get the 5S video of the girl guitar alone
Step 2: Create a video digital person
Then we uploaded the video to DreamFace - Video Digital People
Select our role - upload audio (here to test 30 seconds, beyond the total prompt failure, the system shows that the longest 180 seconds) - the final synthesis
Next add our audio, following the step-by-step illustration below.
Just need to wait a little while will be a finished singing digital man will appear, we just need to split the song audio into 30 seconds a piece, a song time of 2 minutes or so, or 4-5 audio segments, in accordance with the previous way to replace the audio can be.
This way we get the whole music (the same logic applies to the other mouthpieces)
Step 3: Compose the clip
Finally, we'll edit the video with subtitles and narration and we're done!
Of course, if you also want to generate an avatar digital persona to use as a narration or image, then we can just guitar girl's figure to generate one using the instant dream vivid mode, the
Trimming using Paint-in-Picture + Mask function
In fact, there are a lot of useful functions inside the app, picture characters talking and singing, animals talking and singing waiting for you to discover.
It still looks good, doesn't it?Everyone can get hands-on, easy to learn, hurry up and try it~!
