Consistent Character Across Shots

Consistent Character Across Shots

Keeping one face and one wardrobe across a series, using reference parameters instead of hope.

midjourney image character consistency reference

What it is for

One face across a series does not happen by describing the face harder. This uses reference parameters to lock the character and wardrobe, so the variation lands in the pose and the setting rather than in who the person is.

The skill

Consistency comes from references and fixed vocabulary, not from describing the
person more carefully each time.

Establish once:
1. Generate the character alone, front-facing, neutral light, plain background.
   Upscale the best result. This is your reference image.
2. Write a FIXED description block — age, build, hair, eyes, wardrobe — and
   never rephrase it. Rewording it is what changes the face.

For every subsequent shot:
- Pass the reference with `--cref <url>` and set `--cw` deliberately:
  `--cw 100` holds face and clothing, `--cw 0` holds the face only. Use 0 when
  the wardrobe should change and you want the face to survive it.
- Keep the fixed block verbatim, and add ONLY the scene: location, action,
  light, camera.
- Keep `--ar` and the style reference `--sref` constant across the whole set.

When a face drifts:
- Do not add words like "same face" — they do nothing. Lower `--cw`, or
  regenerate from the reference and re-upscale.
- Check that the scene description is not fighting the reference: "in shadow"
  and a bright reference will pull the face toward a different one.

How to install it

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from

Skills app Original