Scenario 11: Miscellaneous
This chapter covers some of my favorite scenes that don't fit other categories. After writing this though, I realized they're all great for creating physical goods!
Stickersβ
Not sure if others also like sticking stickers on laptops like I do. While sites like AliExpress are great, my tastes are pretty niche and hard to satisfy there. So I tried having AI generate some custom stickers, using Stormtroopers and Darth Vader as examples:
Prompt | Explanation | |
---|---|---|
What is the type? | sticker | Making a sticker |
What is the subject? | Stormtrooper, cute | Cute chibi Stormtrooper |
What is the environment? | null | Not critical, let AI decide |
What is the composition? | null | Not critical, let AI decide |
What is the shooting medium? | null | Not critical, let AI decide |
What is the style? | vector | Added vector since I'm thinking of printing |
Parameters | null | No parameters |
The results have nice white borders and vector art style suitable for printing (great for AliExpress sellers!):
Papercuttingβ
I really like layered papercutting art but it's quite difficult to make manually. So I tried with AI:
Prompt | Explanation | |
---|---|---|
What is the type? | layered paper art | Layered paper shadowbox |
What is the subject? | Death Star, shadowbox, volumetric lighting, | The Death Star, but I don't just want a flat shadow, I want a 3D shadowbox with lighting like products I've seen on AliExpress |
What is the environment? | null | Not critical, let AI decide |
What is the composition? | null | Not critical, let AI decide |
What is the shooting medium? | null | Not critical, let AI decide |
What is the style? | null | Not critical, let AI decide |
Parameters | null | No parameters |
The initial results were decent but the lit sphere in frame 3 didn't look right. Frames 2 and 4 were closer to the layered 2D paper look I wanted, so I added "diorama" to enhance the flat shadowbox style.
voilΓ ~ The addition really improved it (right 4 images) - so beautiful, hope to find on AliExpress someday:
Postage Stampsβ
Not sure if people still know about postage stamps in 2023, but I still collect them as a hobby. Thought it'd be fun to generate some retro Star Wars designs:
Prompt | Explanation | |
---|---|---|
What is the type? | vintage postage stamp | Vintage stamp |
What is the subject? | Stormtrooper, red ink, | Stormtrooper and Darth Vader again. Specified red ink color. |
What is the environment? | null | Not critical, let AI decide |
What is the composition? | null | Not critical, let AI decide |
What is the shooting medium? | null | Not critical, let AI decide |
What is the style? | line engraving, intaglio | Traditional stamps used plate printing, often line engraved and etched. So I added those terms - one for line engraving, one for etching. |
Parameters | null | No parameters |
The final images are so artistic:
Postersβ
Finally, let's look at posters:
Prompt | Explanation | |
---|---|---|
What is the type? | movie poster | |
What is the subject? | Star War | |
What is the environment? | null | Not critical, let AI decide |
What is the composition? | null | Not critical, let AI decide |
What is the shooting medium? | null | Not critical, let AI decide |
What is the style? | studio ghibli, retro anime | I wanted a retro Studio Ghibli anime style |
Parameters | null | No parameters |
Posters don't have to just be movie posters - you can also make advertising posters by changing the type:
Prompt | Explanation | |
---|---|---|
What is the type? | advertising poster | |
What is the subject? | Darth Vader points his finger at the viewer | |
What is the environment? | null | Not critical, let AI decide |
What is the composition? | null | Not critical, let AI decide |
What is the shooting medium? | null | Not critical, let AI decide |
What is the style? | studio ghibli, retro anime | Same as above |
Parameters | null | No parameters |
Tip 17: Reconstructing Promptsβ
Simplest way is just asking the creator.
If you can't, Midjourney has an image-to-prompt feature. First enter /describe:
Click describe, then upload image and hit enter.
Midjourney will return possible prompts - click the 1,2,3,4 buttons under the image to directly send prompts to the bot for image generation. On the right are images generated from prompt #1 and #4. It seems decent for reverse engineering composition, subject, style of original images, but don't expect identical results: