Yesterday I though I'd try to create a comic, using AI.
I used ChatGPT to write me a comic script (3 pages organized in frames). I gave it 3 different requests but in all cases the script kind of sucked. But ok, this is something I can always edit myself.
For the rest, it seems to me there are at least 2 major challenges:
1) keep the same style throughout the frames.
2) keep the same heroes, the same appearnce (face, hair, clothes etc.)
I though I'd use midjourney.
First I created separately two different heroines (different hair, different face), looking "at the camera" so that their face is fully visible.
I put the two faces in one picture and uploaded the picture in Midjourney.
Then I copied the link and used it in the prompt, together with my description. Here it is:
Daniel Clowes detailed comic frame, navy blue, melancholic. Exterior wide shot of a decaying building in Greece. Two girls around 20 years old stand in front of the decaying entrance. One has curly black hair, above the shoulders, the other one has straight brownl hair, above the shoulders. Both wear shirts and skirts. We see their backs from afar.
It seemed hard to get what I expected. To give an example, eventhough I ask explicetely to see the back of the girls, they stand face to the camera. Also, I would like them not to be the protagonist of the frame, but rather the buliding to be the protagonist.
So, here are my questions.
a) has anyone tried something similar (creating a comic consisting of multiple frames) with fair results?
b) is there some other tool more suitable for creating consequtive frames of the same story?
c) how can you keep the same heroes througout the mutliple frames?
d) have you figured out what makes it more likely for ai to ignore part of the prompt? ("wearing skirts" and "we see their backs from afar", in my example.)
About c) I realize that when I feed a picture of the heores, this kind of overrides the rest of the description. They become so important that they stand out in the frame, even when I need them to be a detail in the picture.
Thanks
I used ChatGPT to write me a comic script (3 pages organized in frames). I gave it 3 different requests but in all cases the script kind of sucked. But ok, this is something I can always edit myself.
For the rest, it seems to me there are at least 2 major challenges:
1) keep the same style throughout the frames.
2) keep the same heroes, the same appearnce (face, hair, clothes etc.)
I though I'd use midjourney.
First I created separately two different heroines (different hair, different face), looking "at the camera" so that their face is fully visible.
I put the two faces in one picture and uploaded the picture in Midjourney.
Then I copied the link and used it in the prompt, together with my description. Here it is:
Daniel Clowes detailed comic frame, navy blue, melancholic. Exterior wide shot of a decaying building in Greece. Two girls around 20 years old stand in front of the decaying entrance. One has curly black hair, above the shoulders, the other one has straight brownl hair, above the shoulders. Both wear shirts and skirts. We see their backs from afar.
It seemed hard to get what I expected. To give an example, eventhough I ask explicetely to see the back of the girls, they stand face to the camera. Also, I would like them not to be the protagonist of the frame, but rather the buliding to be the protagonist.
So, here are my questions.
a) has anyone tried something similar (creating a comic consisting of multiple frames) with fair results?
b) is there some other tool more suitable for creating consequtive frames of the same story?
c) how can you keep the same heroes througout the mutliple frames?
d) have you figured out what makes it more likely for ai to ignore part of the prompt? ("wearing skirts" and "we see their backs from afar", in my example.)
About c) I realize that when I feed a picture of the heores, this kind of overrides the rest of the description. They become so important that they stand out in the frame, even when I need them to be a detail in the picture.
Thanks
