been messing around w ai image generators for a couple months now and idk if it’s just me but getting consistent, realistic human figures feels weirdly hard. midjourney, sd, leonardo, even a few smaller apps all kinda freak out on super normal words or random contexts. like i put the word bed in one prompt and the whole thing went sideways lol. anatomy also gets funky sometimes, even when i copy a prompt that worked before. i’ve tested other tools on the side like domoai while comparing outputs but the same pattern kinda shows up across the big models too.
anyone else running into this? curious if there are tricks or certain prompt styles that help keep things stable.
[link] [comments] https://www.reddit.com/r/midjourney/comments/1p5el2d/trouble_generating_realistic_humans_without/