Logo
Junwoo
2023-09-07 20:27:27

Tutorial
Creating Characters with Midjourney and Editing Fairytale Videos with VideoStew

These days, we're seeing one or two examples of generative AI being well utilized.

We're going to explore how to use generative AI like Midjourney and VideoStew to create animated video fairytales.

First, because various solutions are used, I'll provide a general workflow.

  1. Create a fairytale script using chatGPT
  2. Generate characters and background images needed for the content via Midjourney
  3. Edit the video by dunking all the ingredients into VideoStew

While services that crop out background images and crop images web services appear along the way, they're not of major importance. So, just remember this three-step solution flow.

Now, let's get started.

Creating a Fairytale Script with ChatGPT

First, let's create a fairytale in chatGPT. This doesnโ€™t require any prompt know-how and is simple enough for you to freely test and generate fairytale scripts.

Use Prompt: Make a short fairytale that can give dreams and hopes to children, with a young boy as the protagonist.

We've made a simple fairytale script. The reason why "a young boy as the protagonist" is specifically mentioned in the prompt is because it's better for using generative AI like Midjourney when there is a definite character.

Creating Characters and Background Images with Midjourney

Now it's our turn to review the script and see what characters are needed and what background elements are needed.

Firstly, we need the protagonist "Minwoo" and his monkey friend "Ruby". As for the backdrop, it seems that we'll need a variety of places like "school classroom" or "zoo".

Shall we shoot one creation prompt at a time?

Creating Drafts for Main Characters in Midjourney

The first task is to generate various poses for the main character.

Basic Prompt: [Character Description], [Style], multiple expressions and poses, character sheet, [Outfit] --ar 16:9

Example of Use: Short boy, brown hair, Korean, black eyes, pixar styled, multiple expressions and poses, front view, side view, back view, character sheet, blue jean pants, pink t-shirt --ar 16:9

The important thing here is that we need to create a character sheet that includes front view, side view, and back view. Since the characters can't always face forward to blend in with the background.

Creating Emotion Variations for Main Character

If you want to create a character variation that has a crying facial expression, a smiling facial expression, or a celebratory pose, it's important to maintain the character's consistency.

You may think that you can just add expressions or poses to the same prompt, but this can lead to completely unexpected results. So remember the reference link if you don't want to suddenly change the protagonist in your fairytale.

First, you need to separate each of the 10 characters from the character sheet that we created earlier. You can use free online crop tools like iloveimg for this. Prepare 10 separate main character image files.

The prompt used to create character variations is as follows.

[Reference link], [Character Description], [Style], expressions and poses, character sheet, [Outfit] --ar 16:9

To get the reference link, you must upload all 10 images to Midjourney. For each uploaded image, right-clicking with the mouse allows you to copy the unique image address. This means you need to copy and paste the address for all 10 images one by one. Although it's a tedious task, it's extremely important for creating variations while ensuring character consistency.

Now, we'll use the /imagine command again to create a variation. In the basic prompt, instead of multiple expressions and poses, we're going to put in real actions or facial expressions. This time, let's try creating a character by putting in jumping with smiling face.

Like this, prepare the variations you need according to the fairytale script. But the images that Midjourney generates all have a background color. So now, we need to use a compositing tool to create only a PNG character image. You can simply blow off the background to transparent with a solution like remove.bg.

Creating Background Images

Once all the characters are prepared in this way, what's left is to create a backdrop. In fact, there's no big problem when writing a prompt for a backdrop as long as the style matches. However, always remember to generate it with a first person view so that the characters don't feel out of place standing in it.

Background Prompt Example: Pixar styled, horizontal path, zoo, first person view, tiger in the cage, bright and sunny day, ice cream truck --ar 16:9

Editing Fairytale Video by Infusing All Ingredients into VideoStew

Once all the image resources are prepared, now all we have to do is dunk them into VideoStew and mix them well. First, put the script on the backdrop via Wizard Mode.

The method of using VideoStew is explained in detail in the Starting with Wizard Mode Guide Document and in the videos on the VideoStew YouTube Channel because the article is getting too long.

<iframe src="https://www.youtube.com/embed/9kZrcLaJC2g" allowfullscreen=""></iframe>

Go to Article
๐Ÿ“ฃ Engaging Students with VideoStew: Real-Life Examples and Practical Guide Digital Transformation of Education Using AIIn today's educational landscape, digital transformation is no longer optional but a necessity.Especially with the i...
Engaging Students with VideoStew: Real-Life Examples and Practical Guide
Junwoo 2024-09-04
๐ŸŽ“ Short-Form Content Strategy for Media Companies: 3 Key Tips for Running a Successful YouTube Channel Recently, traditional newspapers and online news outlets have been paying attention to short-form content. This is a strategic move to adapt to the changing tre...
Short-Form Content Strategy for Media Companies: 3 Key Tips for Running a Successful YouTube Channel
Junwoo 2024-08-28
๐Ÿ—ž๏ธ [Product Update] Improved Performance for Reading Numbers and Units in TTS The functionality has been improved so that all AI voices in Korean can properly pronounce numbers and units.If some sentences are not read correctly, you can c...
Junwoo 2024-08-27
๐Ÿค” UX Details to Make All TTS Engines Properly Read Korean VideoStew offers a variety of AI voices. From Google Wavenet, Amazon Polly, KT AI Voice, Naver Clova, Azure, to ElevenLabs...As a result, a natural issue arises...
UX Details to Make All TTS Engines Properly Read Korean
Junwoo 2024-08-27
๐ŸŽ“ Short-Form Video Channel Management Strategies for E-commerce Businesses (YouTube Shopping, Naver Short Clips, Coupang Shorts) The New Challenge in Digital Commerce Marketing: Short-Form VideosRecently, many companies are struggling to attract new customers. This is a significant issue,...
Short-Form Video Channel Management Strategies for E-commerce Businesses (YouTube Shopping, Naver Short Clips, Coupang Shorts)
Junwoo 2024-08-22
๐Ÿ—ž๏ธ [Product Update] New Feature: View Detailed Music Copyright Information All the music provided by VideoStew can be distributed across multiple channels without any copyright worries.For YouTube, if you register your YouTube channel ...
[Product Update] New Feature: View Detailed Music Copyright Information
Junwoo 2024-08-06
๐Ÿ“– Video Content Strategy to Conquer Naver Short Clips and Coupang Shorts at Once These days, many people are taking an interest in and challenging themselves with short clips on Naver Shopping Live and Coupang Shorts.In this post, we will di...
Video Content Strategy to Conquer Naver Short Clips and Coupang Shorts at Once
Junwoo 2024-08-06
๐Ÿค” Practical GPT Utilization for Startup Marketers (We said GPT, but we only used Claude ๐Ÿ˜…) Hello, this is Jun from VideoStew.Lately, I've been writing a lot of behind-the-scenes posts. In this post, I will summarize how I, as a non-developer at a star...
Practical GPT Utilization for Startup Marketers (We said GPT, but we only used Claude ๐Ÿ˜…)
Junwoo 2024-07-22
[Stop]