Have you ever seen videos where a woman is made of noodles, fire, water, and flowers, dancing in a magical way? These kinds of transformations are taking over Instagram and TikTok right now, and they're absolutely mesmerizing. Today, I'm going to show you exactly how to make your own stunning videos like this.
I'll guide you through the entire process, from crafting the perfect prompts, generating images with AI, and turning them into share-worthy videos that will grab everyone's attention. For those of you new here, welcome. I'm Gary, a content creator and expert in AI video making.
I can help you create viral videos like this on YouTube, Instagram, and TikTok. If you have a video idea you'd like me to reveal, just leave a comment and I'll explain how to make it happen. Now, let's dive into today's topic. Just like any AI video, the first and most important thing is that you need to know the right prompts to create eye-catching images like this. I, like you, don't know exactly what commands the video creator used, but that's not a problem.
First, you need to pick an image that you like from the video. After that, I use ChatGPT to create the prompt based on that image. Specifically, I type the following into ChatGPT. Provide a detailed description of this image, capturing as many elements as possible. Include specifics about the colors, textures, any people or objects present, and the setting.
Describe the atmosphere, any notable features or interactions, and the overall mood of the scene. ChatGPT will give you a detailed prompt, but it might not be 100% accurate just yet. Next, I need to test creating the image with the prompt provided by ChatGPT. To generate the image, I use Leonardo AI, which I personally believe is the best text-to-image tool available right now.
If you're not familiar with it, feel free to try it out yourself. In Leonardo AI, copy the prompt from ChatGPT and paste it into the prompt box on the image creation screen. Next, select the model Phoenix 1.0, which is the latest model. Keep the prompt enhance, style, and contrast settings as the default. For generation mode, Choose Ultra.
Set the image dimensions to 9 by 16 and select a number of images as 4. After that, click Generate. Once the image is generated, you'll see that it's already quite close to the original image. If you want it to match even more closely, do the following.
Click on the image icon, choose Image to Image, and upload the original image. Then, click Generate again. You can adjust the strength setting. In this example, I chose 0.3. The higher the strength, the more the generated image will resemble the original.
You can generate several times to get the best possible image if you want. Repeat these steps for any other images, and you'll have the result you're looking for. Next, I'll use these images to create a stunning video.
To do that, I use Kling AI. It's an amazing tool with very high quality right now. In Kling AI, select AI Videos and choose Kling AI version 1.6.
Next. upload your image. For the prompt box, this is optional and you can leave it blank.
However, for better results, I recommend using this prompt. Animate this noodle character gracefully dancing in a bowl of noodles. Focus on flowing, elegant movements that make the noodle strands sway and ripple.
Include natural flowing effects where the noodles appear to blend seamlessly, giving a sense of constant motion. For creativity, the higher this value, the more the output will follow the prompt. In this example, I set it to 0.5 as the default.
Select Professional mode and set the duration to 5 seconds for this example. As for motion brush and camera movement, these options aren't available in version 1.6, so you can skip them for now. Once everything is set, click Generate and review the result. You should see a beautifully animated video dancing gracefully with smooth, flowing movements and stunning visual effects.
But now, how can you come up with even more creative scenes beyond the images you find on social media? Here's the answer. I continue to use ChatGPT to generate additional prompts based on the original ones. In ChatGPT, I type, replace noodle with ice, stars, smoke, sand, glass, rose, flowers, and your noodle prompt.
ChatGPT will create various different commands based on these elements. Then, you repeat the process from creating images in Leonardo AI to generating the video using Kling AI, just like we did earlier. Here's the result of the videos I created while making this tutorial. If you enjoyed this tutorial, be sure to hit that subscribe button.
Give this video a thumbs up and don't forget to ring the bell for notifications on future tutorials. Let's create more viral content together. See you in the next video.