Parameters
«Parameters are akin to the artist's brush, each stroke fine-tuning the final masterpiece.»
Aspect Ratio (--aspect or --ar)
This parameter changes the aspect ratio of a generation, allowing users to create images that fit specific dimensions.
- --ar 1:1 (default, 1024 × 1024, Facebook profile picture, Picture for Instagram-Feed, LinkedIn profile picture, Kanalicon YouTube)
- --ar 1.41:1 (for standard paper sizes A0 to A10)
- --ar 1.85:1 (movies, widescreen)
- --ar 2:1 ( 1536 × 768, title image LinkedIn)
- --ar 2:3 (686 × 1030, Portrait, Pinterest Standard-Pin)
- --ar 3:2 (1344 × 896, Format of Spiegelreflexkameras, landscapes, street photography, travel photography)
- --ar 4:1 (2176 × 544, Banner ads, Panoramic views, landscapes)
- --ar 4:3 (1232 × 928, Format of Digitalcamera, Video in 90s, Portrait, Landscapes) !!!
- --ar 4:5 (default, 824 × 1030, Portrait Plus, prints, Instagram Post Portrait)!!!
- --ar 5:4 (1200 × 960, Studio photography, architecture, landscape) !!
- --ar 16:9 (1456 × 816, Standard for TVs, monitors,, portable devices, Breitbildfotografie, Facebook-Titelbild, YouTube Thumbnails)
- --ar 9:16 (576 × 1030, mobile Video production, TikTok, Instagram Stories)
- --ar 21:9 (2560 × 1080, ultrawide, panoramic monitors)
--ar 1:1 (default)
--ar 2:1
--ar 2:3
--ar 3:2
--ar 4:1
--ar 4:3
--ar 4:5
--ar 5:4
--ar 16:9
--ar 9:16
Use external AI upscaler tools to increase resolution:
- ON1 Resize AI
- Gigapixel AI
- Let's enhance
Model Version (--version or --v)
This parameter allows users to use a different version of the Midjourney algorithm. The current algorithm (V4) is the default setting.
- --v 1
- --v 2
- --v 3
- --v 4
- --v 5
- --v 5.1
- --v 5.2
- --v 6
- --niji 5 (anime-style images)
- --v6.1
StyleRef (--sref)
- goal: apply artistic overall style from a reference image to a new image
- it applies colors, media, overall asthetics
- it keeps consistent visual signature
- Style Weight
- use style weight (--sw)
- default style weight: --sw 100
- range: 0 to 1 000
- tip: test it out, to find the sweet spot, but keep it low between 25 and 90)
- Style Versions
- there are four style versions (--sv)
- Version 1 to 4
- --sv 1
- --sv 2
- --sv 3
- --sv 4
- prompt --sv + number (--sv 1)
- default: Version 4
- Style Random
- You can also use --sref random to apply a random style to your images.
- Style Codes
- how are they made? --sref random
- they have unique numbers like -sref 1234567
- they can be blended by --sref URL1 URL2 URL3 (--sref 1234567 7654321)
- weight them by adding double colons (::):
- --sref 1234567::1 7654321::2
- to see overall look of a style code, use a moodboard prompt:
- /imagine multi-panel mood board,, split into multiple images, including lifestyle images and color swatches --ar 5:6 --sref 1234567
Stylize (--stylize or --s <0–1000)
The --stylize argument makes generations more beautiful, more consistent, more detailed, more interesting and changing the cropping /composition of the image. It affects the subject, the details, the cropping and the composition.
- default: 100
- range: 0–1 000
- personal preference between 90 and 250
Lower stylize values generates simpler and more minimalistic pictures
Personalization (--p)
- Personalization takes into account your choices when Ranking Pairs.
- By default, --p applies your own personalization, but you can use the personalization of other users by adding their shortcode to the parameter: --p shortcode.
- --p can be used with Midjourney Version 6.
Style (--style)
This parameter fine-tunes the aesthetic of some Midjourney Model Versions. Adding a style parameter can help you create more photo-realistic images, cinematic scenes, or cuter characters.
- --style raw (--v 5.2; --v 6): it allows you to generate images with longer prompts, more natural pictures, more predictable outcomes and less refined images.
4 Parameters Only for --niji 5
- --style cute: for generating adorable characters, subjects, and environments.
- --style expressive: for creating more sophisticated illustrations that emphasize words in your prompt.
- --style scenic: to generate beautiful backgrounds and cinematic characters in fantastical surroundings.
- --style original: it applies the original aesthetic styling for Niji Model Version 5; it was set as default for creations before May 26th, 2023.
Chaos (--chaos <0–100>)
This parameter adjusts the variability of the results. A higher value will produce more unusual and unexpected generations, adding an element of surprise to the image creation process.
- default: 0
- range: 0–100
- prefered: 10–25
The higher the chaos value, the more unusual and unexpected result.
/imagine close-up shot of an eagle looking directly at the camera --chaos 10
/imagine close-up shot of an eagle looking directly at the camera --chaos 75
Quality (--quality or --q <.25, .5, 1 or 2>)
This parameter determines the time the system spends crafting your image. It alters the level of detail in the image.
- default: 1
- range: .25, .5, 1, 2
The higher the quality value, the more details.
The lower the quality value, the simplier the image and the faster it is produced. Lower quality will be used for:
- Abstract Art: These types of images often rely on the broad strokes and vague shapes, where too many details might muddle the overall impact.
- Cartoon Illustrations: For these fun and playful images, keeping things simple and not too intricate might better serve the overall aesthetic.
- Minimalistic Designs: When it’s all about clean lines and an uncluttered look, lower –quality can help keep things neat and uncomplicated.
- Sketch-like Images: If you’re aiming for a rough, hand-drawn feel, a lower quality setting could help mimic the look of a quick sketch.
Repeat (--repeat or --r <1–40>)
This parameter allows users to create multiple jobs from a single prompt, which is useful for quickly rerunning a job multiple times.
--r 3: It produces 3 times 4 pictures (equals 12 pictures at a time)
Seed (--seed <0–4294967295>)
This parameter allows users to specify the seed number used to generate the initial image grids. The same seed number and prompt will produce similar ending images, providing a level of consistency when needed.
- to limit randomness and to get consistency
- to use the same starting point for your prompt
No (--no text)
The No parameter tells the Midjourney Bot what not to include in your image.
--no accepts multiple words separated with commas:
--no item1, item2, item3, item4
Sample prompt: still life guache painting --no fruit, vase, table cloth
Weird ()
This parameter introduces quirky and offbeat qualities to your generated images, resulting in unique and unexpected outcomes.
- default: 0
- range: 0–3000
Image Weight (--iw <.5 to 2>)
This parameter allows user to weight an uploaded image relative to the imaginative text.
- default: 1
- range: to 2
Tile (--tile <v3 or v5>)
This parameter allows user to create repeating seamless patterns.
Use seamless texture checker to check the quality:
