Team -
Since a raccoon's antics have blown away the Internet, I thought I could test this with a custom GPT and a picture of a whiskey-loving raccoon. So, with this prompt, I asked ChatGPT to give me a picture of that rascally bandit using:
"I need a cartoon image of a raccoon that looks photorealistic. The raccoon is drinking a whiskey cocktail."
The first image is what it produced. Nothing wrong with it, but it certainly lacks a bit of depth and character. So, then I changed my prompt a bit to trigger the new GPT:
"My requirement is to generate a cartoon image of a raccoon that looks photorealistic. The raccoon is drinking a whiskey cocktail."
In a flash, ChatGPT provided me with a new prompt ready for cut and paste (especially if you want to use a different engine like Gemini - as Igor pointed out):
"A photorealistic cartoon-style raccoon sitting at a small wooden table, holding a whiskey cocktail in a clear glass tumbler with large ice cubes. The raccoon has detailed, lifelike fur texture, expressive eyes, and subtle soft lighting that enhances realism while keeping a stylized, charming cartoon vibe. The background should be simple and warm-toned to keep focus on the raccoon. The whiskey should appear amber-colored and slightly reflective, with condensation on the glass. The raccoon is sipping the drink through a straw, posed naturally, giving a playful but realistic impression. Highly detailed, sharp, lifelike lighting, rich textures, soft depth of field."
I added soft cinematic lighting and a slightly mischievous expression, and the second image is what ChatGPT generated!
Nano Bannana (the Gemini tool shown by Igor) came up with a fun interpretation, too! That is the third one.
This stuff is fantastic and fun! I can't wait to try this on a more serious subject!