- Systematiq AI
- Posts
- MIDJOURNEY's Master Plan + Random Insane Image Workflow
MIDJOURNEY's Master Plan + Random Insane Image Workflow
What will they do next?
Fam,
New week…new updates.
Let’s not mess around…we’ll get right into it.
Here’s what we’re looking at:
Midjourney Mastery Update
(18) Style Codes to Try
Style Code Workflow For Insane Images
Midjourney’s Ambitious Plans
Reviewing Krea’s Big Splash
Midjourney Office Hours
Let’s Boogie.
MIDJOURNEY MASTERY UPDATE:
Pricing for Midjourney Mastery has been increased to ($187)
But…because some of you missed it…
Here’s the Details:
BETA ACCESS INCLUDES:
(67) Video Tutorials || (14) Modules - Everything from signup-to-storyboarding || (36) PDF Guides || (4) Midjourney Cheat Sheets || Monthly Updates (New Videos + PDFs)
Become an Affiliate:
(18) MIDJOURNEY STYLE CODES TO TRY:
Alright, so I was down on Midjourney Style Codes…I might be back on them.
Spent an afternoon testing a bunch and found some gems.
If you missed it: FULL LINKEDIN POST.
IT’S REALLY ADDICTING:
I spent an afternoon hitting re-roll a million times.
(Trying to keep myself entertained in between baby naps and feeding).
What I Like Most About Style Random:
Unique Color Schemes/Compositions/Lighting
Blending of Art + Photography
Style translation from image to image is great
Seems to work better than SREF image-to-image
What I Like Least About Style Random:
It’s a slot machine.
There’s no rhyme or reason to what comes out.
You waste a lot of fast hours generating.
There’s no guarantee you get something you like.
Here’s the prompt I used for the test:
Prompt: street-style photo of a {lamborghini, samurai, luxurious apartment living room, Gorilla holding a baby gorilla, perfect Martini, Woman in an elegant dress, haggard Biker on a Harley Davidson, insane Mountain climber, beautiful Floral arrangement} --ar 4:5 --sref random
STYLE CODES:
SREF 2928055011: Vibrant colors, dynamic patterns, high energy.
SREF 4232867803: Lush greens, detailed foliage, tranquil.
SREF 3690843001: Muted tones, architectural forms, urban feel.
SREF 3178308583: Soft pastels, smooth textures, intimate.
SREF 24490047: Sepia tones, vintage textures, nostalgic.
If you want all (18) style codes… Here’s a FREE DOWNLOAD.
With the carousel + style codes and description.
STYLE CODE WORKFLOW:
The style codes have a very distinct look.
So, how do you get more unique?
By blending multiple style codes to get some wild stuff.
Let’s take a look…
Blending Style Codes
BLENDING STYLE CODES:
One style code is cool…but…
Multiple style codes will give you something special.
To do this you need to:
Create a prompt
Add “--sref random” as the parameter
Generate the image.
MJ will provide a “code” for that sref
Repeat this a few times to get some codes.
Example Input Prompt: Samurai —ar 4:5 --sref random
Example Output Prompt: Samurai —ar 4:5 --sref 3983467
Once you have the codes…
Create a prompt
Use images generated from different codes as SREFs
Play around
MULTIPLE SREF’s
Here’s some things to experiment with:
Blending Multiple Style Codes using SREF images
Playing around with Style Weight [--sw parameter]
Switch between Style Versions [--sv parameter]
Dial up the Stylize + Chaos parameter.
Example Prompt: street-style photo of a samurai --ar 4:5 --sref https://s.mj.run/Vl5tTRcCaHM https://s.mj.run/hAcov65yD2o https://s.mj.run/WlgYHDioTc0 --sw 300 --stylize 500
Style Codes + SW + SV
Even with a short prompt…you get some cool stuff.
But, if you start to build out your prompts…you can really get some dynamic imagery.
Example Prompt: street-style photo of a samurai, fighting another samurai with a katana, in the {bamboo forrest, cotton field} dynamic low angle action shot --chaos 5 --ar {4:5, 2:1} --sref https://s.mj.run/Vl5tTRcCaHM https://s.mj.run/hAcov65yD2o https://s.mj.run/WlgYHDioTc0 https://s.mj.run/be05pqtHxG0 https://s.mj.run/TxPexwhfTvI --sv 2 --sw 100 --stylize 500
Final Products
You can go a thousand directions with this.
I’m going to do a longer piece on it this week…stay tuned.
MIDJOURNEY’S AMBITIOUS PLANS
Midjourney said this office hours that caught my attention….
Their goal is to create a NEW MEDIUM.
Meaning…they don’t want to replicate photography or video…
They want to do something totally different.
Branded Worlds
BUILDING ENTIRE WORLDS
Think…”text-to-metaverse.”
You’ve seen MJ releases Character Reference, Style Reference, etc.
They create consistency in scenes and characters.
So, the goal is for you to build your own world…
And with 3D capabilities…you’ll be able to immerse in it.
What will branded experiences look like in 5-10 years?
The future of marketing might be in a new medium.
Think 3D immersive buying experiences in “Apple Vision Pro”.
This video on Twitter really opened my eyes to what’s possible.
A simulated buying experience for Tesla through Apple Vision Pro.
Hey @elonmusk, I have some ideas on purchasing a Cybertruck (or any Tesla) using the Apple Vision Pro.
— Brett Morris (@brettmorrisxyz)
7:01 PM • Jan 30, 2024
KREA LAUNCHES VIDEO CAPABILITIES:
Krea has jumped into the video space.
And their new video tool is now available to everyone.
You can try it for FREE HERE.
announcing Krea Video Open Beta.
our new video tool is now available to everyone.
see tutorial below 👇 x.com/i/web/status/1…
— KREA AI (@krea_ai)
9:11 AM • May 23, 2024
REVIEW:
I gave it a whirl and it’s definitely cool…but it’s mostly in the “just for fun category.”
Not seeing commercial use cases (yet).
Unless your brand/business has a more “abstract” aesthetic.
Currently, you can only add images and interpolate them together. The only level of control you have is the “speed” of the motion and the image inputs.
That being said, this is version 1…and I’m sure it will advance like the rest of the platform. If you haven’t played around with Krea, I would suggest spending an hour or two getting to know it. (Especially if you like sketching).
When I ran my tests I liked the results…but I wish they were sharper and more defined. Everything is very “dreamlike”.
KREA Test
MIDJOURNEY OFFICE HOURS [5/22]:
Slow week in terms of updates from them…but here’s what we got:
Web Updates
Working on streamlining Discord feature parity and mobile improvements
Spending a few weeks on general web improvements, mobile, onboarding, and missing features
Midjourney 6.1 Release
Improves image quality, coherence, body, face, and skin-related aspects
Image quality improvements expected to be very noticeable
Moving slowly ahead with personalization
Aiming to ship 6.1 with image quality improvements and personalization together
No specific time estimates were provided due to difficulty in estimating.
Have fun.
If you need help…DM me on Linkedin.
Until next time…
Rory