Your image prompt isn’t broken.
Your aspect ratio is.
🧵↓
Most people prompt like aspect ratio is an afterthought.
But it’s the hidden puppet master of perspective.
If your...
portraits look off...
cinematic shots fall flat…
landscapes feel cramped…
It’s probably the frame (not the idea).
Here’s the fix:
• 3:2 or 16:9 for landscapes and breadth
• 16:9 with negative space for ad-ready layouts
• Avoid ultra-wide for close-ups (space kills detail)
• 4:5 or 2:3 for vertical depth (portraits, tall objects)
• 1:4 to stack scenes vertically (like a visual story strip)
• 2:1 for clean side-by-sides (perfect for “before/after”)
• Prompt for composition: “rule of thirds,” “left third,” etc.
• Ultra-wide (21:9) to imply hidden space, mystery, tension
• Instructions like, “subject on left, empty right” to guide focal balance
With the new system, it breaks previous --sref codes, so you no longer can just plug-in --sref #### for legacy codes. You will have to use the --sv4 parameter when using any style code generated before 6.16.25.
With that, I wanted to drop some killer legacy Midjourney style codes. These slap.
One massively underrated way to create realistic images in Midjourney (or any AI image gen tool for that matter)?
Film.
By using a specific film in prompts, you can:
→ elevate mood, location, scene, colors, lighting
→ achieve targeted styles—no long-tail keywords needed
→ give a unique look/feel while matching perfectly to a desired era
In this pocket guide, I'm going to share
❑ 35+ film types you can use (color, b&w, creative)
❑ 30 keywords/tokens you can use without specifying a specific film