So i tested around a lot to get GPT4 to write the best photo prompts possible and here is what i came up with. Some bits i grabbed from the Net and other parts i wrote myself. All the Information about Stable Diffusion,Dreambooth, midjourney is from Wikipedia or the midjourney site. The Example Prompts are all from Prompts i made with GPT4. I lay the emphasis on Photography right from the beginning. Its rather lenghty but i got the best Results by following these steps:
Prompt 1: We are going to create Images that look like photographs with a Diffusion Model. I am going to send you some information about diffusion, okay?
Prompt 2-6: Insert Stable Diffusion.txt, dreambooth.txt, how midjourney works.txt, midjourney parameters.txt, midjourney v5.txt into the prompt one after the other. Do it in the exact order.
Prompt 7: Do you feel like you now understand Midjourney, the diffusion model and dreambooth?
Prompt 8: Now i will send you some example prompts used to make photographs in Midjourney V5.
Prompt 9: Insert midjourney photo prompt examples.txt into the prompt.
Prompt 10: Now i want you to write prompts. Use a detailed and descriptive language to describe your image prompts. Use detailed descriptions from photography like cameras used,camera settings, lenses, focal lenght, angle of view, shutter speed and other terms from photography. Please write a prompt to create <INSERT YOUR FIRST PROMPT HERE>