So today's Michael J Fox's birthday, and I thought I could pay homage to him on my Facebook timeline by posting an image of him and Christopher LLoyd from the 2015 version of Hill Valley as depicted in the movie "Back To the Future 2" that I hand crafted using Stable Diffusion.

So I tried this prompt, against v1.5 "photograph of the fictional Hill Valley from back to the Future in 2015, 35mm, sharp, ultra quality, sharp focus, tack sharp, clear, 8K UHD"

Having prior to this prompt had no luck in describing a full scene - thinking if I could start from a workable image - I can insert in MJ and Chris afterwards, but for the life of me, I cannot get it to do this or nothing resembling the scene from BTTF2.

Is there a way to get SD to recognize prompts for fictional locations from movies or tv shows without detailing absolutely everything for it from scratch and without training it specifically on pictures from those movies or tv shows?

Any 'prompt engineering' recommendations would greatly be appreciated. Thanks in advance!



You must log in to answer this question.

Browse other questions tagged .