Midjourney mami - open source ur process

example prompt to get you started making images like this - need to add in more detail for ur ideas. My final prompts usually end up being 5-7 lines long in discord.

28 years old South Asian female wearing {insert clothing description} laying her head down on a table in an empty room, {insert detailed description of the room}, in the style of {insert photographer(s) / director(s)}, 35mm film, pastel colors, kodak, cinematic lighting, late afternoon, golden light, {insert type of camera lens} --ar 3:2 --v 4


how many times did you have to render/prompt to get it to these images?

Depends how good the original image is for sure - but I usually create Variations at least 4 times, and up to even 15x if im obsessed with getting an image perfect (like the very top one here was about 13 variation pushes).

Midjourney will slightly render up every time you create a variation so the more you do it before hitting “Upscale” for the final render, the better the end result usually is



1 Like

Is it usually just an intuitive decision if one of the variations are worth pursuing as the main one?

In v3 of midjourney - every time you choose “Variation” you are getting 4 very different compositions, so the choice is more impactful. But with v4, its not really a broad choice like that. The 4 renders you get are super close to the original image, so you’re really looking for nuanced differences (e.g. I like to look for the faces that have evolved more because that will give a better final render)