GM! What makes #midjourney5 so special is the quality of the outputs when using long prompting. A mini thread🧵👇🏻
This prompt is over 250 words. All the parameters can be changed and replaced to create combinations and generate so diverse images. Prompt has to be structured:
👇🏻
Type of shot / scene / character / action ( repeat if there’s others) / light/ quality output description/ camera style/ type of rendering
Each section of the prompt is divided in subsections that allow you to deep dive in the prompt description for exemple: character> gender/color of skin/ hair/eyes/ mouth / look/ … I personally detail each subsections.
Conclusion: the effects of long prompting allows you to be very precise on the description and get accurate outputs of what you aim for.
Be curious on your environment and make descriptions precisely is a good exercise to progress in prompts making. Less generations better… twitter.com/i/web/status/1…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Following a great thread by @nickfloats about /describe command which is just mind blowing in analyzing an image… just realized that is pure image to text command!!!😱
First input the image reference👇🏻
2/ the text output! The MJ 🤖offers 4 descriptions… 1-2-3-4… choose one of the four to create a different output which is “inspired by the input image+the bot description”👇🏻Here the four results…
Then as for the /imagine command choose the upscale you prefer or make variations…👇🏻et voilà!
Just a little 🧵about consistency in characters in #midjourney. What I’m experiencing since V5… and it is just my opinion, there’s 3 ways to achieve this goal! 👇🏻
1/ import a referential photograph and insert it at the beginning of the prompt <generated reference > output
2/had the name of the character you want into the prompt and specify an action in relation with the character in reference. Here Louise Maurisset a woman skateboarder