Midjourney for image direction

What it is good at

• Best-in-class aesthetic quality for editorial and campaign imagery.
• Style references let a whole set share one look.
• Excellent at light, material and atmosphere.

Where it falls short

• Poor at exact products, text and logos.
• Needs iteration; first results are rarely final.
• Learning curve on parameters.

How to use it for creative work

1. 01

   Write like a photographer

   Subject, action, setting, light, lens, mood. Six short clauses beat one long sentence.

2. 02

   Fix one variable at a time

   Change the light or the lens, never both, or you cannot tell what worked.

3. 03

   Build a style reference

   Once one image is right, use it as the style reference for the rest of the set.

4. 04

   Keep a prompt log

   Save the exact prompt with each keeper. Your own log becomes a house style within a month.

5. 05

   Finish elsewhere

   Crop, colour and retouch outside the tool. Generation is direction, not delivery.

Ready-to-use prompts

Full-length editorial image of [subject], standing in [setting], overcast north light, 85mm lens, shallow depth of field, muted warm palette, film grain, calm and unposed --ar 4:5

Extreme close-up of [material] catching low side light, visible texture and weave, neutral background, macro lens, editorial still life --ar 1:1

Wide campaign frame of [subject] in [location], early morning haze, long shadows, 35mm lens, restrained colour, documentary feel, negative space on the right for type --ar 16:9

[Approved image as style reference]. Same visual world, same light and palette. New frame: [describe only the new action and angle].

Contact sheet of nine related frames exploring [theme], one consistent palette and light, mixed close-ups and wides, editorial research board feel --ar 3:2

Questions people ask

Can it render my actual product?

Not accurately. Use it for mood and direction, and an image editor for the real product.

How do I keep one model across shots?

Use a character reference from an approved frame and repeat the same description each time; expect small drift and choose the closest frames.