Media & Brand

What are good skills to work on?:

1. Learning Descriptors
   1. https://docs.midjourney.com/docs/explore-prompting
   2. https://midlibrary.io/midguide/general-modifiers-in-midjourney-prompts
   3. https://sprinkleofai.com/writing-good-midjourney-prompts/
2. Learning Stylessome text
   1. https://midlibrary.io/styles
3. Learning how to photobash
4. Knowing the vocabulary of art

You can paste images into your prompt and use them in 2 main ways:

1. Using a single image as a reference point with text 
2. Blend multiple images together with(or without) text.

You can use --iw to adjust the importance of the image vs the text. Higher values mean the image has more impact. It ranges from 0-3 with the default value at 1.

• You can assign weights to each image if using multiple
• Prompt: URL1::2 URL2::0.75

• You can get closer to what you want by photobashing an image first. This means to do a rough overlay of its component parts.
• Imagine you want a specific dragon head on a specific car. You can photobash it and reference it to generate the image.

the Chaos parameter:

The --c parameter influences how different the 4 generated images are. High values produce more varied results, and lower values are more similar. The range is from 0-100 with the default being 0.

Character Reference:

You can create images referencing specific characters using –cref URL

• The character weight or –cw parameter sets the strength from 0-100. –cw 0 focuses on the face only, with higher values using face, hair clothing, etc.
• You can maintain a consistent style as well by adding in the –sref parameter, more on this later.

The NO parameter:

This parameter tells midjourney what to exclude.

The Quality parameter:

The –q parameter changes how much detail to give an image. It defaults to 1, but goes up to 2.

The Repeat parameter:

The –r parameter runs multiple jobs at once. Accepts between 2-4 for basic subscribers

Style reference:

Used by calling --sref URL

• You can use more than 1 image as a style reference at a time: –sref URL1 URL2 URL3
• You can also use a random midjourny style with –sref random

‍

Style Weight:

• se --sw to set the severity of a stylization when using an –sref. Accepts values from 0-1000 with 100 as the default
• When using multiple style references, you can assign them different weights --sref URL1::2 URL2::1 URL3::1

Style reference seeds:

• Whenever you do -sref random you will notice there is a number that comes up. You can use that number in future style references to prompt images in that specific style. 
• You can blend sref codes together: ie –sref 123 456
• You can blend an sref code with an image -sref 123 URL
• And of course you can add weight to sref codes –sref 123::2 456::1

Stylize:

the –s parameter influences how artistic the image is. Higher values are more artistic but may not adhere to the prompt as strongly. The default is 100 but it accepts 0-1000

Tile:

 the –tile parameter creates an image that can be repeated seamlessly to create a pattern for fabrics, wallpapers or banners.

illustration of space and stars --tile

Weird:

The –w parameter determines how unusual something is compared to previous midjourney generations. It accepts 0-3000 and the default is 0.

Adding Text:

You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.

Custom Zoom:

You can use double quotation marks (“) around words or phrases to specify text you want to have appear in your image.