I'm going to show you how to use AI to
create action-packed comic books with multiple characters and exciting scenes.
But first, let's take a preview at what we're going to learn to create today. Jin: "How could you betray me June? You were like a daughter to me." June: "Like a daughter you stole
me from my family... I was happy..." Jin: "I saved you..." June: "You took everything from me!!! My home,
my family I won't let you cause any more pain...
"It all ends today!! Jin: "We almost had everything!!!" Jin: "You never understood....I did it all for you...."
June: "It's finally over..." 20 Years Ago... Unknown Woman: "Please take her with you... we have no
food to feed her..." Jin: "I think I'll call you... June" I hope you enjoyed that. To create Ai Comics there's five steps.
The scripting and the storyboard, generating character references, creating images of
those characters doing different actions, combining the characters and adding in
the backgrounds, and finally arranging the pictures into a finished novel. We'll
cover all the steps in this tutorial. And by the end of this video you'll be able
to make an AI comic from start to finish all by yourself. First, we need to come up
with a story. I start by writing down the script, which includes a dialog that
occurs between my characters and also some basic ideas for how the scenes will
play out. My comic has two characters. Jin who's a rugged old man and a retired
soldier, and June who his adoptive daughter. You can get inspiration for
this step from movies, TV shows, or books. Or use a AI model like Chat-GPT to give
you some hints. After we finish writing the script it's time for storyboarding.
In this stage, I took images from the internet to help me visualize what the
find pages in the comic will look like. Storyboarding is extremely important, not
only because it helps us visualize the comic panels and character interactions.
But because we'll also be using the photos here as reference for the AI to
generate the comic book images. Without the step it'll be too difficult to have
the AI produce the exact image and character motions that we want. I got
most of these photos from YouTube clips of movies or famous photos of people.
I used AI to create a few of the simpler poses, like the side profile of the girl
standing since that's a pretty basic image. Not all the images and script we
have at this stage will go into the final comic panels, but they give us a
good idea of what the final story will look like. Once the story board is finished
it's time to create reference images of our characters themselves. I used the
Leonardo Ai image generator to create my character references. One for Jin, who's
the rugged old man and one for June. I put down the settings I used here if you
want to try them out. Here's the deal. The more basic and generic your characters are,
the easier it will be for the AI to keep them looking consistent in your comic
panels. This process works best if you use characters generated by the AI
itself. Now comes the hard part where we'll create the pictures for the comic
panels. In order to do this we'll use the character reference images and also the
storyboard photos we collected earlier. I'm going to take this picture of Thanos
raising his fist and use it as a reference for Jin in the fight scene.
I'll use Leonardo AI to generate the images. Let's navigate to the image generation tab.
I'm going to enter the settings I showed earlier for the anime style
specific AI. You can also do this with other image styles like realistic photos
or Pixar style images. Using this image guidance icon, I can attach a character
reference. So in this case I'll upload my reference image of the old man. Now I
want him to be lifting his fist in an aggressive manner like this picture of
Thanos. So I'm going to use the image guidance feature again and this time
click through the menus to get to the pose to image reference, which allows me
to take the pose from a reference photo and map it into my AI generated
image. There's one more final image reference we need to use. The content
reference which lets us take the structure from one image and transfer it
onto another image. This helps us really nail down the exact character pose and
angle we want in the final comic panels. You'll need to play around with
these reference weights a little bit. For character and content reference I like
to keep them from mid to high. That seems to work the best. And for the pose
reference you need to increase the weight a little bit. Around 1.5 should
work fine. Then in the prompt bar let's roughly describe what the character
looks like "a rugged old man with blue hair wearing a black jacket. He's
throwing a punch." I'm going to add on that "he's glowing with electricity" to
give him a superpower look. It will take some experimentation to make this work.
The hands are probably the most inconsistent part and you'll get a lot
of deformed fingers. I had to generate a bunch of different samples to get one I
liked in terms of the pose, the hands, and also the character consistency.
Here's another example. This time I'm using a female character
reference, and I'll go with a pose reference image of a side profile shot
of her looking upwards. I'm going to include a Content reference image just
like I did previously. Because we're generating a different character, the
default art style that the AI model will generate for her will be a little
different than before. So what I'm going to do is actually attach the male
character as a style reference. This will make sure that the female and male
characters have matching visual art styles. For scenes with multiple
characters in them, you need to generate the individual characters separately.
This side profile of the girl will eventually be combined with the side
profile of the man, so that they're facing each other. Now you don't need to
use the pose and content references for every image. It's okay to use a character
and style reference and describe what you want the character to do in the prompt.
However, for more complex images it does help the AI a lot to give it
some more hints. All right, so I finished creating all the individual character
images. Now it's time to combine my characters together and add in the
backgrounds. I'm using Photoshop to combine the characters into the same
scene. But you can use any free editor like PhotoPea as long as it has
background removal. First I'll place the characters into the scene. Here's a
dialogue scene between the male and female characters.
Then I'll use the object detection tool to select the silhouette
of my subjects and separate them from their current backgrounds. I want them to
be talking in a dark alley, so I'll head back to Leonardo Ai and prompt for "a dark
alleyway at night", and also attach a style reference from my original
character image, since I want to make sure the background
has the same art style as my characters. With the background generated, I can
insert that behind my character silhouettes in Photoshop. One important
tip is that you want to adjust the brightness of the characters and the
backgrounds when you're combining them together. First, I'll reduce the
brightness of the background since that's not the focus of the scene and I
want the characters to stand out. Then I'll increase the brightness level for
each of the characters. This step is really, really
important if you're adding in your own backgrounds. That looks way better now.
Here's a comparison of a fight scene I made with before and after adjusting the
brightness. The scene looks way more dynamic and explosive with brighter characters.
You often also need to adjust the colors on the clothing a little bit.
So the female character has a white shirt underneath her jacket in this image.
But if you look at the other comic panels, she's wearing darker colors
underneath. So I'll adjust the color tones of the white T-shirt a little bit
bit to make it darker and have a more greenish hue, to make the images match better.
There's no perfect way to get the clothing consistent in all of your
images. You can try using specific prompts for clothing like "a khaki green
jacket with a maroon sweater underneath." But you might need to create a bunch of
images before you get something that you're satisfied with. Finally it's time
to start creating the actual comic strips. I'll be using Canva for this.
The hard part of creating the images is over and now we get to have some fun using
our previously planned storyboard to arrange the comic panels. Just upload
your images and crop and resize them to fit properly. Make sure you add in black
borders to make it look more like a comic. The great thing about Leonardo is
the high resolution images it creates which makes it easy to resize and expand
and crop the panels to exactly what I want. For the text bubbles, I just took
them from the Canva pre-made assets and then added in text boxes on top of them.
If you don't want to use Canva or you don't have the pre-made bubble assets,
you can just do a quick Google search to find them. Once everything is arranged
according to the storyboard, we get some really nice looking comics. AI has come a really long way.
There are still some limitations, but it's pretty amazing what we can do with
it now if we get creative. Let me know down in the comments what other types of
tutorials you want to see. If you want a deeper dive into the character
consistency feature in Leonardo Ai go and checkout this video over here.