Recently, ChatGPT's image generation feature has become a global sensation, and the enthusiasm is no less than when GPT-4 or Sora were released, with a variety of ways to play with it.
Taking advantage of this boom, OpenAI's CEO, Ottman, announced that the image generator is now available even for free users (twice a day).
This certainly opens up new creative horizons for more users
If you combine GPT4o with DeepSeek text function and AI video, the gameplay will be more colorful!
Using GPT-4o's multi-round image generation function, we can generate a series of consecutive images, and then use Conch AI to turn them into videos to make a cozy parent-child cartoon!
The process of making it is so simple that AI geeks can easily get started 😜
I'll start with the basics and then share the full production process
I. Basic operation
1、Real life photo to cartoon
Tool GPT-4o
Due to AI security restrictions, we are required to animate a real person's photo before we can proceed.
Upload a photo and specify the style using the prompt word, "Convert to Pixie Style" to get an anime version of the family photo.
2、GPT4o Continuous out of the map
Once you get a picture, you can continue to enter requests under the same dialog to keep GPT4o producing pictures
For example, type in "family eating, 16:9" and the AI immediately generates a happy dinner scene.
The newly generated image character image is exactly the same as before!
If there is a style inconsistency, you can add the keyword "
Pixar 3D Style"
For those who can't use gpt4o, you can use beanbags instead, but the failure rate is higher
3. Conch to video
We'll make the picture move now.
OpenConch AI (https://www.1ai.net/21428.html), select the "Video Generation" function.
Click Generate and wait a few minutes...
WATCH! A life-like and vivid family dinner is freshly baked! Doesn't it feel like you can't hide the traffic code in your hand?
II. Practical cases
Core Processes:
Real photo/story idea (DeepSeek) → AI generates comic image (GPT4o) → AI converts image toShort VideoClip (Conch AI) → AI Generate Voiceover (Conch AI) → Edit and synthesize final video
Here's a step-by-step process for you:
1. DeepSeek Write Scripts
Enter DeepSeek (https://www.1ai.net/24339.html）
Enter the prompt: I want to draw a parent-child cartoon. The theme is Family Adventures in the Amazon Jungle, the characters include dad, mom, grandson, and grandma, write 5 drawing prompts, the style should be funny, and no text is needed in the picture. Form Output.
Get the script
Note that you can manually adjust the drawing cue words based on the AI output, trying not to use adverbs
2、GPT4o Make a map
As described earlier, go to GPT4o and enter the drawing cue words generated in the first step to get different images
3、Do the video
Make videos with pictures separately, operate as before, cue words can be left blank
Remember! Be sure to pick the model 12v-01-live, great for picture to video!
4、Dubbing
Conch also has a super powerful free text-to-speech feature
We started by having DeepSeek write the voiceover verbatim:
"Write a 20-second spoken word verbatim script to be used as an off-screen voiceover for a short video, including the family driving to the forest, the grandson's cannibal flower bites his foot the family saves him, the family meets monkeys happily posing for a photo, taking out bananas to try to eat them results in a group of gorillas dominating the scene, and the family flies home in a big bird with the help of the big bird"
DeepSeek. Right away.
Go to Conch Dubbing Channel
Select the model, enter the verbatim script, select the timbre (we'll go with Bean), and set the parameters for the emotion (automatic is fine) and the speed of speech.
5 seconds to get a munchkin voiceover
5. Editing and synthesizing
Finally, use editing software to edit the video and voiceover together, a short video full of love for parents and children is completed!
How is it? Doesn't it feel that with the support of Conch AI, the GPT-4o's graphic generating function instantly becomes more powerful? We can use them to create more and more eye-catching pop-up videos!
The variety of playstyles is simply overwhelming!For example, funny cartoons, compulsion celebrities, classic IP second creation, digital people with goods, etc., are very suitable for everyone to send to the major self media platforms to attract traffic to cash!
Now the domestic video AI has begun to crush foreign similar software, it is also very convenient to use, the practicality is really very high!
Conch AI + deepSeek + GPT-4o is simply a traffic godsend for self-publishers to cash in!
If you don't feel comfortable using GPT-4o, you can also use beanbags to make diagrams.