Image Generation / Midjourney
Categorize and explain Midjourney parameters by use case — turns parameter confusion into structured knowledge.
Why This Prompt Exists
Midjourney has dozens of parameters. Most users memorize 2-3 and ignore the rest — missing capabilities that would improve their output significantly.
You get:
- using default aspect ratio for everything (wrong for most use cases)
- stylize values that don’t match your subject (too chaotic or too flat)
- no chaos adjustment (same output every time, no exploration)
- ignoring negative prompting (–no) when you need it most
- no parameter experimentation because you don’t know what exists
But parameters have specific purposes:
- –ar (aspect ratio): controls output dimensions for different use cases
- –c (chaos): controls output diversity across generations
- –s (stylize): controls artistic interpretation vs. literal following
- –no (negative prompt): excludes unwanted elements
- –iw (image weight): controls reference image influence
- –seed: reproduces specific results
Without classification, users never discover what’s possible.
This prompt categorizes and explains Midjourney parameters by use case.
The Prompt
Assume the role of a Midjourney technical educator who explains parameters. Your task is to categorize and explain Midjourney parameters by their function. Generate: 1. PARAMETER CATEGORIES | Category | Parameters | Purpose | |----------|------------|---------| | Output Format | --ar, --w, --h | Control dimensions and shape | | Style Control | --s, --stylize, --style | Control artistic interpretation | | Variation | --c, --chaos, --seed | Control reproducibility and diversity | | Content Control | --no, --iw | Exclude or weight elements | | Quality | --q, --quality | Trade-off between speed and detail | | Version | --v, --version | Specify model version | | Advanced | --tile, --video, --repeat | Specialized functions | 2. DETAILED PARAMETER TABLE | Parameter | Values | Default | Best For | Avoid When | |-----------|--------|---------|----------|------------| | --ar | X:Y (e.g., 16:9, 4:3, 1:1) | 1:1 | Social posts, presentations, print | Square-optimized templates | | --c | 0-100 | 0 | Exploring variations, multiple options | Consistent branding | | --s | 0-1000 | 100 | Artistic interpretations (higher) | Photorealistic (lower) | | --no | comma-separated list | none | Removing unwanted elements | Overly restrictive prompts | | --iw | 0-2 | 1 | Strong reference influence (2) | First attempt (use 0.5-1) | | --seed | integer | random | Reproducing specific outputs | Exploration | | --q | 0.25, 0.5, 1 | 1 | Final renders (1) | Drafts and testing | 3. USE CASE RECOMMENDATIONS | Use Case | Recommended Parameters | Rationale | |----------|----------------------|-----------| | Social media graphic | --ar 16:9, --s 250 | Wide format, moderate stylization | | Product mockup | --ar 1:1, --s 50, --q 1 | Square, low stylization, high quality | | Logo exploration | --ar 1:1, --c 50, --s 100 | Square format, high variety, neutral style | | Character concept | --ar 2:3, --c 80, --s 300 | Portrait orientation, high variety, artistic | 4. PARAMETER COMBINATION RULES - Compatible: --ar + --s + --c (work together) - Incompatible: --v 5 + --style raw (version-specific) - Order matters: parameters go at the end of the prompt - Syntax: `--parameter value` (space after double dash) 5. QUICK REFERENCE CARD (one-line for each parameter) - `--ar 16:9`: widescreen - `--c 50`: moderate variation - `--s 250`: noticeable stylization - `--no text,logo`: exclude text and logos - `--iw 1.5`: stronger image reference - `--seed 1234`: reproducible result INPUTS: Your experience level: [BEGINNER / INTERMEDIATE / ADVANCED] Use case (optional): [E.G., "Logo design", "Product photography", "Social media"] Specific parameter questions (optional): [E.G., "What does --chaos actually do?"] Midjourney version: [V6 / V7] RULES: - Default values matter — know what changes when you add a parameter - Higher --stylize doesn't mean better — it means more interpretation - --chaos controls diversity between runs, not within an image - --iw > 1 works for V6, not for earlier versions - --seed is essential for reproducible workflows - Negative prompting (--no) is underused — most users should use it more
How To Use It
- Default values matter — know what changes when you add a parameter.
- Higher –stylize doesn’t mean better — it means more interpretation.
- –chaos controls diversity between runs, not within a single image.
- –iw > 1 works for V6, not for earlier versions.
- –seed is essential for reproducible workflows (brand assets, series).
- Negative prompting (–no) is underused — most users should use it more.
Example Input
Your experience level:
“INTERMEDIATE — I know basic parameters but not advanced combinations”
Use case:
“Social media graphics for LinkedIn (carousel format)”
Midjourney version:
“V6”
Why It Works
Most Midjourney users learn parameters by copying from others — not by understanding what each parameter actually does.
This framework improves outcomes by forcing:
- parameter categorization (grouping by function)
- value range documentation (what numbers mean)
- use case recommendations (when to use which)
- compatibility rules (what works together)
- quick reference (one-line memory aids)
Failure modes this prevents:
- Default aspect ratio for everything (wrong for most outputs)
- Chaos set to zero (all outputs look identical)
- Stylize too high for photography (artifacts, distortion)
- No negative prompts (unwanted elements in every generation)
This improves on: Trial-and-error parameter learning. Structured classification builds mental models.
Related to: MJ-03 (Aspect Ratio) for –ar specifics; MJ-04 (Chaos) for –c tuning.
Build Better AI Systems
Subscribe for advanced prompt engineering, AI coding tools, debugging frameworks, and practical strategies for developers and engineers.