Product and material studies
Pair a product silhouette with another image's material, lighting or color language to explore a campaign direction.
Midjourney
A blend, not a collage
Midjourney V7 Blend reads the core elements across several references, then reorganizes subject, composition, materials, color and mood as one new image. It is not a simple overlay; it turns separate visual ingredients into a fresh creative direction.
It extracts the references' core visual ideas and generates a new image, so it is better for direction-finding than pixel-perfect replication.
Core value
What does Blend actually combine?
Blend goes beyond color mixing. Visual information from each reference shapes the new image's structure and character.
- Subject and silhouette
- Sets the object and structure people recognize first.
- Material and palette
- Carries metal, glass, fabric or illustration texture into the new image.
- Composition and light
- Influences framing, scale, contrast and visual emphasis.
- Mood and world
- Brings weather, period, place and narrative tone into one direction.
Best for
When should you use Midjourney V7 Blend?
Use Blend when the direction already exists across several images and you want a new result that carries their shared visual DNA.
Editorial portrait concepts
Combine a pose or styling reference with a different studio mood to discover a fresh portrait treatment.
Environment worldbuilding
Mix architecture, landscape and atmosphere references into a coherent setting for stories or brand worlds.
Illustration style discovery
Bring together composition, palette and medium references to find a visual language worth developing.
Flux Art
How to get a stronger blend
Start with two references that have clear roles, control the framing, then let V7 discover an unexpected but coherent connection between them.
Blend images with V7- Give every image one jobStart with two clear references: decide which supplies the subject or structure and which supplies the style, material or mood.
- Align the framingChoose references with compatible crops and use an output shape close to their aspect ratio so the composition has less work to reinterpret.
- Judge the shared ideaLook for a convincing new direction rather than an exact copy, then keep the best result and refine it in the image workspace.
Quick pre-upload check
- Images are clear and key subjects are not heavily cropped
- Every reference has a distinct, intentional role
- You have permission to use every uploaded image
Blend recipes
Reusable blend brief templates
Blend is driven mainly by the images, not a long text prompt. Use these templates to choose references and define the result; switch to V7 Imagine with multiple image prompts when words must guide the blend.
01Launch concept
Product × material
[Reference A: product silhouette and camera angle] + [Reference B: material, palette and lighting]. Goal: keep A recognizable, borrow B's finish and mood, and create one polished [campaign use] visual.
02Brand portrait
Portrait × editorial mood
[Reference A: pose and framing] + [Reference B: wardrobe, color story and studio light]. Goal: create a new editorial portrait with a coherent [brand mood], not a literal collage.
03Worldbuilding
Place × atmosphere
[Reference A: architecture or landscape structure] + [Reference B: weather, time of day and color]. Goal: build a believable [story world] scene with layered depth and a unified atmosphere.
Compare related AI image models
Midjourney V7 Blend FAQ
What is Midjourney V7 Blend?
Midjourney V7 Blend is a reference-led way to combine multiple images. It reads their subjects, composition, materials, colors and atmosphere, then reorganizes those ideas into a new image.
How is Blend different from making a collage?
A collage keeps the original images and arranges them together, while Blend reinterprets their visual elements. The result is usually neither reference A nor reference B, but a new creative direction.
How do I use Midjourney V7 Blend in Flux Art?
Open the image workspace from this page, upload at least two reference images, choose an output shape and start the blend. You can complete the workflow directly in Flux Art without using Discord.
How many reference images should I use?
Start with two images so you can see what each one contributes. Add more only after the direction is clear, and follow the current upload limit shown in the Flux Art workspace.
How should I assign roles to two reference images?
Let one image provide the subject, silhouette or composition, and let the other provide material, color, lighting or atmosphere. For example, use a product photo for its shape and an architectural image for its surface and light.
Do the reference images need the same aspect ratio?
No, but similar crops and aspect ratios usually produce a more stable composition. A close-up portrait and an ultra-wide landscape ask the model to reinterpret much more of the framing.
Do I need a long prompt for Blend?
No. Blend is driven mainly by the reference images, and the blend briefs on this page help you choose images and define a goal. When text must control specific content, use Midjourney V7 Imagine with multiple image prompts instead.
Can I use Blend for product images?
Yes. It is useful for exploring product materials, lighting, color, backgrounds and campaign atmosphere. If packaging text, logos or small product details must stay exact, review the result and refine it before publishing.
Can I use Blend for portraits?
Yes. You can combine pose, styling, framing and studio mood to explore editorial or brand portrait directions. Blend should not be treated as a tool for preserving one person's face exactly.
Is Blend useful for environments and illustrations?
Yes. Architecture, landscape, weather, palette and artistic medium can come from different images, making Blend useful for story settings, brand worlds and new illustration directions.
Will reference image A always be preserved more than image B?
No. The A and B labels help you plan each image's role; they do not set a fixed influence. If one element matters most, use a clear reference with a simple subject and minimal background distractions.
What should I do if the blend looks cluttered or loses the subject?
Reduce the blend to two images with distinct roles and use references with compatible framing. Replacing a busy image is often more effective than adding more references to correct the result.
Should I use Blend, V7 Imagine or Nano Banana?
Choose Blend to discover a new direction from several images, V7 Imagine when text and images should guide the scene together, and Nano Banana or GPT Image 2 when you need more controlled edits to an existing image.
Can I use Blend images in commercial projects?
Blend results can support brand concepts, advertising drafts and product visual development, subject to the current Flux Art and model-provider terms. Make sure you also have permission to use any reference images, likenesses and brand assets you upload.