Image Reference and Consistency By the end of this lesson, students should be able to: What –sref does. The –sw (style weight) parameter. Using –sref for series consistency. A verified –sref reference + consistent text prompt structure produces style-consistent output across a series – new subjects rendered in the consistent visual style of the reference image. This is one of the most powerful tools for visual identity consistency in Midjourney: establish one reference image as the style anchor, then generate all series content using that reference. Selecting effective style reference images. For –sref: images with a distinctive, clearly readable style work best – high stylization, clear aesthetic character, and consistent treatment across the reference image. Generic or visually inconsistent reference images produce weaker style transfer. Generate a test set of 3-5 images using the candidate –sref reference before committing it to production series use. Enroll to continue this lesson. The preview above shows the lesson objectives and opening lesson content. Enroll to view the full lesson, complete the practice work, and take the lesson quiz.
Lesson 2: Style Reference (-sref) for Visual Consistency
Lesson Objectives
Lesson Content
--sref [image URL] applies the visual style from a reference image to a new generation – the model extracts and applies the style (color treatment, rendering approach, texture quality, aesthetic character) while generating new content from the text prompt. Unlike image prompting (which guides the whole image), –sref specifically targets style transfer from the reference.--sw controls the strength of the style reference. --sw 0 – minimum style influence; --sw 100 (default) – standard style application; --sw 1000 – maximum style influence, which may override text prompt elements in favor of the reference style. Adjust based on how strongly the reference style should be applied to the new content.
