Differences
This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision Next revision | Previous revision | ||
| ai:midjourney:workflow [2026/03/18 21:48] – [Use Prompt Comments in the Prompt Bar] mh | ai:midjourney:workflow [2026/03/18 21:56] (current) – [Always Work From One “Seed Image”] mh | ||
|---|---|---|---|
| Line 88: | Line 88: | ||
| This creates image families instead of random outputs. | This creates image families instead of random outputs. | ||
| - | ==== Use Prompt Comments in the Prompt Bar ==== | + | <WRAP round info 60%> |
| + | **Don’t Upscale Too Early** | ||
| - | Midjourney lets you add text that doesn' | + | Upscaling everything creates clutter and consumes credits. |
| - | === Example: === | + | **Rule of thumb:** |
| - | <code text> | + | Grid → shortlist → upscale only the best |
| - | forest shrine at dusk, cinematic lighting --ar 3:2 | + | </WRAP> |
| - | // exploration phase | + | |
| - | forest shrine at dusk, cinematic lighting --ar 3:2 | ||
| - | // fairy version test | ||
| - | </ | ||
| - | This way the prompt itself becomes documentation when browsing later. | + | ==== Controlled Experimentation ==== |
| - | ==== Create Clear Project Themes ==== | + | Instead of random prompts you change one gene at a time. |
| - | Instead of random prompts, work in sessions. | + | Example Base prompt: |
| - | For example: | + | <code text> |
| + | ancient forest shrine | ||
| + | foggy forest | ||
| + | soft volumetric moonlight | ||
| + | cinematic fantasy concept art | ||
| + | wide angle composition | ||
| + | </ | ||
| - | Session 1 | + | Then test only lighting: |
| + | * **Test A:** '' | ||
| + | * **Test B:** '' | ||
| + | * **Test C:** '' | ||
| - | Fairy forest environments | ||
| + | Now you understand exactly **what changed** in the image. | ||
| + | ==== Use Prompt Comments in the Prompt Bar ==== | ||
| - | Session 2 | + | Midjourney lets you add text that doesn' |
| - | Steampunk airships | + | === Example: === |
| + | <code text> | ||
| + | forest shrine at dusk, cinematic lighting --ar 3:2 | ||
| + | // exploration phase | ||
| - | Session | + | forest shrine at dusk, cinematic lighting --ar 3:2 |
| + | // fairy version test | ||
| + | </ | ||
| - | Ancient ruins concept art | + | This way the prompt itself becomes documentation when browsing later. |
| - | The advantage: | ||
| - | When browsing generation history, images are clustered naturally by subject. | ||
| - | ==== Don’t Upscale Too Early ==== | ||
| - | |||
| - | Upscaling everything creates clutter. | ||
| - | |||
| - | Instead: | ||
| - | |||
| - | Rule of thumb: | ||
| - | |||
| - | Grid → shortlist → upscale only the best | ||
| - | |||
| - | |||
| - | This does two things: | ||
| - | |||
| - | reduces noise | ||
| - | |||
| - | keeps upscales meaningful | ||
| - | |||
| - | Upscaled images = serious candidates. | ||
| - | |||
| - | ==== Use the Editor Before Upscaling ==== | ||
| - | |||
| - | Since you're already experimenting with Editor / scene composition, | ||
| - | |||
| - | Ideal order: | ||
| - | |||
| - | Grid | ||
| - | ↓ | ||
| - | Select best | ||
| - | ↓ | ||
| - | Editor adjustments | ||
| - | ↓ | ||
| - | Upscale | ||
| - | |||
| - | |||
| - | Why? | ||
| - | |||
| - | Because editing before upscale: | ||
| - | |||
| - | keeps fast iteration | ||
| - | |||
| - | saves credits | ||
| - | |||
| - | avoids editing giant images later | ||
| ==== Maintain One “Hero Image” ==== | ==== Maintain One “Hero Image” ==== | ||
| - | For each exploration session, try to end with: | + | For each exploration session, try to end with **one hero image** |
| - | + | ||
| - | one hero image | + | |
| That’s the one you: | That’s the one you: | ||
| - | + | * refine | |
| - | refine | + | |
| - | + | | |
| - | upscale | + | |
| - | + | ||
| - | export | + | |
| - | + | ||
| - | archive | + | |
| If you create 50 images but only 3 heroes, your archive stays clean. | If you create 50 images but only 3 heroes, your archive stays clean. | ||
| Line 193: | Line 154: | ||
| ==== Weekly “Creative Garbage Collection” ==== | ==== Weekly “Creative Garbage Collection” ==== | ||
| - | Once a week: | + | Once a week open Likes and ask: |
| + | * still interesting? | ||
| + | * worth refining? | ||
| - | Open Likes and ask: | + | Remove likes from anything that no longer excites you. |
| - | still interesting? | + | This **keeps your shortlist alive**. |
| - | worth refining? | + | === A Simple MidJourney Workflow Blueprint === |
| - | Remove likes from anything that no longer excites you. | + | <code text> |
| - | + | ||
| - | This keeps your shortlist alive. | + | |
| - | + | ||
| - | A Simple MJ Workflow Blueprint | + | |
| 1. Exploration | 1. Exploration | ||
| | | ||
| Line 220: | Line 179: | ||
| 5. Final | 5. Final | ||
| hero image selected | hero image selected | ||
| + | </ | ||
| - | + | This keeps the system manageable even with thousands of generations. | |
| - | Inside MJ, this keeps the system manageable even with thousands of generations. | + | |
| - | + | ||
| - | Next Step (Very Important) | + | |
| - | + | ||
| - | Once the MJ side is clear, the real power move is the download and archive workflow. | + | |
| - | + | ||
| - | Because the real chaos begins when you have: | + | |
| - | + | ||
| - | MJ exports | + | |
| - | Photoshop edits | + | |
| - | Upscaled versions | + | |
| - | Prompt notes | + | |
| - | + | ||
| - | + | ||
| - | I can show you a very clean professional archive structure used by concept artists and AI creators that makes thousands of images easy to navigate. | + | |
| - | + | ||
| - | It includes: | + | |
| - | + | ||
| - | naming conventions | + | |
| - | + | ||
| - | folder structure | + | |
| - | + | ||
| - | prompt storage | + | |
| - | + | ||
| - | versioning | + | |
| - | + | ||
| - | post-processing pipeline | + | |
| - | + | ||
| - | And it pairs extremely well with Midjourney. | + | |
| - | + | ||
| - | If you'd like, we can build that system next. It becomes your personal AI image studio pipeline. 🎨 | + | |