SOLVING Common Midjourney Problems + AlphaFold-3’s Crazy Potential

Fam,

Let’s get right into it.

Midjourney is gearing up for the launch of V6.1.

There’s no timetable yet…but based on their office hours convo…it’s getting close.

So, while things are slow on the dev front…let’s improve our skills.

Here’s what we’re rolling with today…

1. Midjourney Mastery Course Update.

2. Solving Common Midjourney Problems.

3. Film Stock Tokens to Try.

4. Cinematic Elements for Prompting.

5. Linkedin Live: Midjourney’s New Website.

6. AlphaFold-3’s Crazy Potential.

7. Midjourney Office Hours.

Let’s Boogie.

SOLVING COMMON MIDJOURNEY PROBLEMS:

I get a lot of questions…most of them can be solved within Midjourney.

So I put together a quick troubleshooting guide.

If you missed it: FULL LINKEDIN POST

PROBLEM: FULL BODY SHOTS

This is a big one…For w/e reason you have to jump through hoops to get full body shots in Midjourney.

However, there are a few tricks that can help.

Solutions:

• Start with 9:16 aspect ratio.

• Use “Change Aspect Ratio.”

• Use “Feet” in the text prompt.

Simply prompting “full body shot” doesn’t work often.

But oftentimes it’s not a “prompting problem”…It’s an “aspect ratio problem.”

The Thought Process:

A human body is taller than it is wide.

So use an aspect ratio...that is taller than wide. (Ex: 9:16, 4:5, 2:3, 1:3, etc)

Same theory applies to other objects.

9:16 vs 16:9 Aspect Ratio

Not making sense?

Let’s try this prompt…and look at the difference.

Prompt: editorial photography, full body shot, happy pregnant woman wearing a stunning white dress contrasting against a vibrant graffiti covered wall, bright mid-day lighting, Lecia m10-R, --ar {16:9, 9:16}

Use Change AR:

Maybe you don’t need your image in 9:16…you need/want it in 16:9…

Just use “Change AR” to expand the image.

So do this in 2 steps:

1. Generate image in 9:16

2. Change AR to 16:9

Simple as that.

Use Feet in Text Prompt:

When in doubt…add “shoes” into the prompt.

Midjourney wants to listen…and it will force the feet into the shot…it’s a little hack, but it works.

Prompt: editorial photography, full body shot, happy pregnant woman wearing a stunning white dress, white sneakers, contrasting against a vibrant graffiti covered wall, bright mid-day lighting, Lecia m10-R, --ar 16:9

SOLVING BLURRY FACES:

One of my biggest gripes with Midjourney is the further a “face” is from the frame…the worse it generates.

Fixing this isn’t as straightforward…but…

If you start in close..and then zoom out…

You’ll keep the facial details and resolution.

Process:

1. Generate closeup shot in 9:16 AR

2. Change AR to expand image.

3. Use “Zoom” to zoom out.

Prompt: editorial photography, closeup shot, traditional Scottish hunter in the Royal Forrests of Arran, captured with 100mm lens, greens and grays, dappled sunlight --ar 9:16

Think about building “inside out.”

There’s a few more good hacks in THE POST.

(36) FILM STOCK TOKENS TO TRY:

Film stocks are awesome tokens because…

One token can have multiple effects…and it can give your images a cool vintage dynamic.

Here’s (36) more “unique” ones to try - FREE DOWNLOAD BELOW 👇

LINKEDIN LIVE - NEW WEBSITE WALKTHROUGH:

Last week I went live with Drew Brucker to break down Midjourney’s new website.

Naturally, we talked about the website updates…but we ended up going deep on some other concepts.

If you want to nerd out on Midjourney for an hour…this is the spot.

CINEMATIC TOKENS FOR MIDJOURNEY:

Midjourney knows photography…it also knows cinema.

Cinematic cameras have a unique visual signature and they can help your prompts.

Here’s a few to try:

• Panavision Panaflex Millennium XL2: Anamorphic, bokeh, lens flares

• Lomography LomoKino: Lo-fi, grainy, saturated analog aesthetic

• Leica M10-D: Sharp, contrasty, excellent color rendering

• Canon ME20F-SH: clean and noise-free in near-darkness

• 16mm Krasnogorsk-3: Vintage, grainy, nostalgic Russian

• Bolex H16: 16mm aesthetic with imperfections

• Blackmagic URSA Mini Pro 12K: Extremely high resolution, detailed

• RED Ranger Monstro 8K VV: High resolution, wide dynamic range, aggressive, crisp digital look

Prompt: cinematic still, insane rock climber scaling Mount Kilimanjaro, captured on [cinematic camera] --ar 16:9 --stylize 90

Going to do a full breakdown on cinematic elements this week…

Stay tuned.

ALPHAFOLD-3’S CRAZY POTENTIAL:

This is one of the cooler aspects of Ai…

The potential to significantly alter our existence (in a positive way).

The AlphaFold 3 model can predict how biological material works together and interact. (Like Proteins and DNA)

Understanding how these pieces fit together helps scientists figure out how our bodies work…and how to predict biological interactions.

MEANING…scientists can test ideas and theories sooo much faster.

This tool could potentially aid in:

• Faster drug discovery and development

• Deeper understanding of diseases

• Personalized, targeted medicine

• Improved crops and agricultural practices

• Development of sustainable, bio-inspired materials

• Breakthroughs in fundamental biological research

• Time and cost savings in scientific experiments

This tool has the potential to win a Nobel Prize.

But only if we choose to use it in the “right way.”

Just as there’s potential for super-positive outcomes…this also makes it easier for people to use it in the wrong way…I’m sure you can extrapolate what I mean.

This video does a good job of breaking it down.

MIDJOURNEY OFFICE HOURS (5/8):

Upcoming Features for Rooms:

• Adding more features to rooms to encourage usage

• Private rooms

• Recent rooms

• Allow people to watch rooms without an account

• If no accounts in a room, users can only watch, not create

• Allow adding friends to a room for voice chat while creating

Personalization and Midjourney Version 6.1

• Personalization training was unstable, producing inconsistent results

• New version of personalization is showing promise

• Will be tested by guides and mods before deployment

• Version 6.1 focuses on image quality, coherence, and text rendering

• May decide to make 1% of things worse to improve the rest

Midjourney Version 7

• Much smarter with more data and better prompt understanding

• Improved handling of multiple objects

• Faster with bigger batches and better aesthetics

• 30-40% of V7 features likely to be included in 6.1

• Significant image quality improvements expected

Development Process

• Many people working on 6.1 and personalization

• Doing more things than usual, which slows progress

• Balancing act between doing too many and too few things

• Ongoing R&D for 3D, video, and real-time models

• Progress is slow but steady

Personalization Rationale

• Current aesthetics limited by trying to appeal to everyone

• Model tends to favor 38-year-old male preferences

• Risk of pushing out other demographics

• Personalization allows tailoring to individual preferences

Have fun.

If you need help…DM me on Linkedin.

Until next time…

RF