Creating a consistent character in Midjourney V5.2 is challenging, especially for the characters in the photographic style.
There have been some published methods on this topic, and most require multiple rerolls to collect enough images to blend them in a prompt. That requires significant effort, experimentation, and kissing goodbye to your precious “fast hours.”
In contrast, the success rate for creating consistent characters in the illustration style (comics, painting, cartoon, etc.) appears to be higher than that of photography. There’s also information on how to make a consistent character sheet in an illustration format with different views, and so on.
I want to show you a new approach to the problem of creating consistent characters in the photograph format. It’s not a foolproof solution, but it has a lot of advantages and is worth a shot.
It does not overpower or replace other published methods of creating consistent characters in photography. However, you can quickly get a different look for your character (see above featured image) without rerolling multiple times to collect enough images as in other methods.
I assume you know [Pan] and how to use /remix. If all of these steps are unfamiliar to you, please read the following stories: (1) Midjourney’s new Pan Feature: 5 Things to Know (for an overview of how to use [Pan]) (2) Midjourney Pan Feature: 8 use cases (for a conceptual understanding of how [Pan] works — especially for the Use Case ???2 where [Pan] forbids altering the original image when panning — but I’ll show you how to break that rule later).
OK, let’s take a look at what the new method has to offer.
An Overview of the Symmetrical Method