Midjourney V6 Consistent Style Tutorial + Best Prompting Tips
Turn AI into your ultimate creative collaborator
Midjourney V6 vs Everything
MJ V6 is better because of the following reasons:
Natural Language Coherence
Prompt Adherence
Photorealism
Prompt setup:
Set the main scene + style
Describe the details
Describe the setting
Things to avoid:
Style buzzwords like 8k, vray, HDR, Photorealistic, Award Winning etc
Prompts with comma-separated keywords
—Basic V6 Prompt—
A man and a woman sitting next to eachother at a wooden table in a quaint coffee shop --ar 16:9 --style raw --v 6.0
—Detailed V6 Prompt—
A man and a woman sitting next to each other at a wooden table in a quaint coffee shop. On the right is a middle-eastern woman in her 30's 𝙬𝙞𝙩𝙝 𝙡𝙤𝙣𝙜 𝙘𝙪𝙧𝙡𝙮 𝙝𝙖𝙞𝙧, she 𝙞𝙨 𝙬𝙚𝙖𝙧𝙞𝙣𝙜 𝙖 𝙜𝙧𝙚𝙚𝙣 𝙨𝙬𝙚𝙖𝙩𝙚𝙧. She is holding a cup of coffee and smiling. On the left is 𝘾𝙖𝙪𝙘𝙖𝙨𝙞𝙖𝙣 man in his 30's 𝙬𝙞𝙩𝙝 𝙜𝙡𝙖𝙨𝙨𝙚𝙨 𝙖𝙣𝙙 𝙖 𝙨𝙝𝙤𝙧𝙩 𝙗𝙚𝙖𝙧𝙙, he is 𝙬𝙚𝙖𝙧𝙞𝙣𝙜 𝙖 𝙗𝙚𝙞𝙜𝙚 𝙨𝙬𝙚𝙖𝙩𝙚𝙧. His is talking with his face turned towards the right. They are in a warm and rustic cafe interior with bright incandescent light bulbs on the wall. Filmstock: CineStill 800T --ar 16:9 --style raw --v 6.0
Midjourney V6 has listened to each and every detail except for the woman who should be holding the cup, but I think we can cut some slack there…
Try it out!
Midjourney V6 Consistent Style Tutorial
This is my favourite way to get Consistent Styles
(while having full control)
Full tutorial 👇
Midjourney released the first test algorithms for 'Consistent Styles' today.
They are calling it the feature "Style References".
They work similarly to image prompts where you give a URL to one or more images that 'describe' the consistent style you want to work over.
How to use Style References:
- After your prompt type --sref along with a URL to an image --sref urlA
- Midjourney will look at the image URL as a 'style references' and try to make something that 'matches' its style or aesthetic
- This works for both V6 and Niji V6 (it does not work with V5 etc)
Advanced Settings
- Use multiple urls to multiple images like this --sref urlA urlB urlC
- Set relative weights of styles like this --sref urlA::2 urlB::3 urlC::5
- Set the total strength of the stylization via --sw 100 (100 is default, 0 is off, 1000 is maximum)
This is so much more powerful than style tuner and gives you finer control over the final image generation style.
I am super happy with this update!
Are you ready to chart new territories in the realm of AI? 🗺️
Have fun!
If you need help…DM me on Linkedin.
OR respond to this email with “HELP”…and we’ll chat.
Until next time…