Hi Midjourney Enthusiasts! I was joking by telling you to steal my list — all of the things that are in this list are freely shared on Google Docs for you to find and peruse. Let's be clear I did not create these PDFs, all I am doing is providing you with a compiled list of PDF files in one place to make viewing and downloading them easy for you.

The 15 files in this list combine to create an online book called Midjourney: The book of SREF styles. (The authors of the content are credited below.)

The first and second chapters provide a concise and useful analysis of the first 40 "random" styles available to Midjourney users. The following chapters present SREF codes for Drawing and Painting styles, Colorful and Monochromatic Sketch styles, 3D and Digital Drawing styles, Photographic, Mixed Media, Printing, Lithographic, and some Additional styles.

In each chapter the styles are presented as a list displaying the list number and dots of the colors dominating the style.

None
screen captured from the Chapter 1 PDF by the author

Then there is a list of thumbnails of each style. Here are thumbnails of the first eight styles.

None
screen captured from the Chapter 1 PDF by the author

These thumbnails seem to be showing how each style creates a unique looking cat.

And then there is a list of tags (keywords) that you can use to order and filter the list.

None
screen captured from the Chapter 1 PDF by the author

When testing each style the testers used a consistent set of prompts. Here are the prompts that were used.

prompt 1:

{flower, boy, house} --v 6.0 --ar 2:3 --sref 00000000 --s 250 --seed 123 

prompt 2:

{flower, boy, house} --v 6.0 --ar 5:7 --sref 00000000 --s 250 --seed 123 

prompt 3:

landscape --v 6.0 --ar 2:1 --sref 00000000 --s 250 --seed 123

If you aren't used to looking at Midjourney prompts I'll interpret these prompts for you. The codes preceded by two dashes are called parameters. The v parameter sets the version of the software to use. The ar parameter sets the "aspect ratio" (relative dimension sizes) of the image. A 1:1 (1 to 1) aspect ratio would produce a square image. The s parameter stands for "stylize" (how strongly to apply the style to the image — can be values from 0 to 1000). The seed parameter gives the server a point to start from (using the same seed value in each prompt insures that the server will "start" the image generation from the same point.).

The {} brackets are called a permutation. The values of the permutation are separated by commas. In this instance, prompt 1 and prompt 2 would each be run 3 times, once for each value in the permutation (once with the text "flower", once with the text "boy", and once more with the text "house").

The sref parameter is where the style reference number is put in the prompt. After the sref, the 0's would be replaced with the number of the style. The style numbers can run from 0 to 4000000000 (4 billion).

So if I were to translate the first permutation of prompt 1 into English it would be something like "create a flower using version 6.0 of the software, make the ratio of the width to the height 2 to 3 (like 200 pixels wide by 300 pixels tall), apply style number 00000000, apply the style at approximately 1/4 of full strength, and use 123 as the seed value."

Samples of the images created by the style are included in PDF. Here are examples for style number 8 (sref 8).

None
screen captured from the Chapter 1 PDF by the author

In the image above prompt 2 was used with an aspect ratio of 5:7 (5 to 7). The words in the rounded rectangle are the tags mentioned before. The white outlines separate the permutations. The first group of 4 is for the word "flower", the second group is for "boy", and the third group is for "house".

None
screen captured from the Chapter 1 PDF by the author

This image uses prompt 1 with an aspect ratio of 2:3.

None
screen captured from the Chapter 1 PDF by the author

Finally this last image shows prompt 3 with the text "landscape". Its 2:1 aspect ratio demonstrates that it is twice as wide as it is tall.

For comparison I have included two more images of pages for style number 20 (sref 20).

None
screen captured from the Chapter 1 PDF by the author

This image shows the responses for prompt 2 (5:7 aspect ratio). Notice the tags.

None
screen captured from the Chapter 1 PDF by the author

And this image shows the landscape image created with sref 20.

You can tell that creating these images took a substantial amount of time by the creators of the files.

Credit for the Content in the Files

The content of this online book is owned by the authors of the resources, Ran Liron, Daniel Trabelsi, Inbal Weisman, Ron David Ben Ishay, Nataly Shafir, and Gina Dawidowicz. They put in the time to research and collect the style codes in the PDFs

The index provides some contact information for these authors:

Here's the list.

The List of Links to View/Download all of the SREF PDFs

Each link will take you to Google Docs to view the PDF in your browser. You should be able to click an icon button to download the file to your computer. Each individual PDF has keywords and tags to let you search inside the file.

Individual PDF files for Midjourney: The book of SREF styles

If you found this article to be useful, please let me know by responding to this article. You can also clap up to 50 times if you want to. My other articles are all available from my profile.

Thanks for reading and Peace to you. David

None

This story is published on Generative AI. Connect with us on LinkedIn and follow Zeniteq to stay in the loop with the latest AI stories.

Subscribe to our newsletter and YouTube channel to stay updated with the latest news and updates on generative AI. Let's shape the future of AI together!

None