Immediately to March 3, there are people do not know Guangxi March 3 will be put 4 days long vacation? Overseas people expressed "envy and jealousy", but Guangxi people do not say, just a preparation for the "March 3". It can be imagined that only this holiday has all the important.
So can we utilize AI for this holiday season? Take a hands-on learning experience today.
There are AI tools that need to be used on this:
1. Raw map tool: i.e. Dream
2. Video tools: i.e. Dream/Keling/
3. Editing tools: Cutting and mapping
ralph lauren polo ralph lauren cue words out of the chart
Want the character to sing, first we need the subject character, here you need to generate through the cue word, here the cue word has been prepared for you.
HD photography, background Guilin landscape Li River, 20 year old high color beauty, wearing Zhuang costume, headband with unique Zhuang headdress, short Zhuang skirt, standing on bamboo raft singing mountain songs, bamboo raft moving along the water, full body portrait of Zhuang beauty, 8K resolution, sunny, vibrant bamboo forests on both sides of the Li River. Natural light and shadow, high definition details
HD photography, background Guilin landscape Li River, 20 year old high color handsome guy, wearing Zhuang costume, headband with unique Zhuang headdress, Zhuang pants, standing on the bank of Li River singing mountain songs, full body photo, UHD, 8K resolution, sunny, vibrant bamboo forest on both sides of Li River.
HD photography, background Guilin mountains and Li River, 20 year old high color beauty/handsome guy, wearing Zhuang costume, headband with unique headdress of Zhuang, short skirt of Zhuang, sitting on a bamboo raft and singing a mountain song, the bamboo row is moving along the water, close up, highlighting the people's percentage, 8K resolution, sunny and bright, the two sides of the Li River are full of life in the bamboo forests. Natural light and shadow, high definition details
Modify it yourself according to the men and women to use it.
ralph lauren polo ralph lauren text born picture
There are many AI drawing tools we just want to choose the right one for us.
That is, the dream URL: I found a treasure AI products, recommended to you: multi-style painting artifacts, limited time free experience, quick poke https://jimeng.jianying.com/
Open directly to dream
Enter prompt word, best 2.0PRO after model testing
After selecting a good input prompt word directly out of the map, the effect is as follows
After drawing an image you are happy with, click on Canvas Edit - Ultra HD - 2x Download to save it locally.
Here is a selection of 5 images to test.
ralph lauren polo ralph lauren polo ralph lauren pas cher video
Here is an example of a Kerin 1.6 model: open the Kerin-AI video
Upload the pictures in order to generate, Mr. into a separate dynamic video, for example: Zhuang A-Go A-Mei is enthusiastically singing mountain songs, the requirements of the singing action, the action is natural and smooth, no deformation, no defects, the bamboo row along the water!
I won't show the effect one by one, but save it after generating all of them.
ralph lauren video lip sync
Here, we use Dream, Spirit, and DreamFace. Before lip-syncing, we need to prepare the audio of the character singing, which can be exported via cutscene.
Here's a comparison for you:
i.e. dream-digital person-counterpoint - just upload a picture directly here!
Kerin-AI Video - Counterpoint
Effect Comparison
Instant Dreams (simple and effective, with good lip-synching and movement, you can just generate the whole song if you have the points)
Ke Ling (the effect is also good, first of all, when generating the video can not be overturned, the back of the lip-synching according to the length of the video to match, 5 seconds of video with 5 seconds of audio, if the generation of a section needs to be split into a small section of the generation of points to consume less)
But these 2 have a big common problem, that is, if the dream is to produce a finished product is too consuming points, but the spirit of the video lip-synching and the length of time can not be.
I ended up using it.DreamFace
If you don't know how to use it, you can read the previous section.
Quickly create your own unique digital person, let the pictures speak and sing in minutes!
Select our character - upload audio (can be 30 seconds) - final synthesis
The overall effect is all very good, just not very sharp. Also because the character's full-body photo occupies a small percentage of the screen, resulting in smaller features, so it does not look very vivid. Try our generation time to keep the characters close up to highlight the main image of the characters.
ralph laurenVideo Editing with Cutscene
1, open the cut screen PC Professional Edition, enter the main interface, click [+ start creating].
2, add the saved video clips, add background music, you can directly search in the cut screen
3, the appropriate location to add the right sound effects, you can double the effect of the video.
Learning point: text, the addition of a background can make the picture feel better.
Cautions and Operational Recommendations
1. Maintain originality: Do not directly copy the content of other people's cue words, by modifying the creative body of the cue words or adjusting the details of the screen video, to ensure that the work has a certain degree of originality.
2. Optimize the rhythm of the video: too long animation content may lead to viewers skipping, it is recommended to control the video time to ensure a compact rhythm.
3. Accurately locate the audience: choose a platform and time period where the target audience is clear to publish the video, for example, uploading it in the evening for better results.
4. Interaction to enhance fan stickiness: actively reply to fan comments, increase the interaction rate, attract more users to pay attention to and forward.