Midjourney is a text-to-image generative artificial intelligence system. This means that it is capable of generating images starting from text instructions. The only way to interact with the Midjourney Bot is on Discord and to do that, you have to enter the appropriate commands at the prompt.
There are several requests that can be submitted to Midjourney, but one of the most famous is the command“/imagine” which allows to generate four different image attempts starting from a short text description.
While the template isn’t perfect, Midjourney is a powerful tool for unlocking creativity and imagination. For this, the most important thing is to dedicate the proper attention to prompt engineering techniques to obtain the desired result in the shortest possible time.
1. What are the parameters on Midjourney and how do they work
First of all, we must consider that the Midjourney Bot works best with simple sentences that describe exactly what you want to see, a bit in a “telegram” style and therefore it is good to try to be concise and essential in the choice of words.
Specifically, the parameters are options added to the end of a prompt to change the way an image is generated. They can change the aspect ratio, the style, the visual noise, but also the versions of the artificial intelligence model that alimenta Midjourney.
The more details you add, the more material the bot will have to work with. Surely you need to arm yourself with a little patience and get ready to repeat the process several times to get the desired results. With a little practice, the process will become easier.
2. Change the look and quality of images
Using the parameter “–aspect” o “–ar” it’s possible change the aspect ratio of the image which will be created. The proportions refer to ratio of width to height of an image and are usually represented as one pair of numbers separated by colonssuch as 7:4 or 4:3.
The default is 1:1 and –aspect accepts only integers (139:100 instead of 1.39:1)
The most common are:
- –aspect 1:1 Default aspect ratio
- –aspect 5:4 Common frame and print ratio
- –aspect 3:2 Common in printed photography
- –aspect 7:4 Close to HD TV screens and smartphone screens
Example of prompts: /imagine prompt colorful narcissus –ar 5:4
By adjusting the parameter “–quality” o “–q”, you can change the time it takes to generate an image. Set one image quality higher takes longer to process, but produces finer detail. However, the quality setting does not affect the resolution of the output.
The default quality is 1, which is the base quality, but you can set it to .25 (quarter quality) or .5 (half quality). Higher settings aren’t always better. Sometimes “less is more”: small values can make abstract images more beautiful, while higher values are perfect for bringing light to the details of buildings and architecture.
Example of prompts: /imagine prompt detailed peony –q .25
The parameter “— stylize” o “–s” it affects the intensity of colors, compositions and artistic forms. A low stylization value produces images that look more like the prompt, while higher values produce images that are more artistic and less related to the prompt. The default value for stylization is “–s 100”, but you can set it to any value between 0 and 1000.
Example of prompts: /imagine prompt illustrated strawberries –s 100
3. The parameters for the creative “control” of the images
The parameter “–chaos” o “–c” it has an impact on the diversity of the image grids initial. A high chaos value produces more unusual and unpredictable results, while lower chaos values produce more consistent and repeatable compositions. You can specify a chaos value between 0 and 100 using the “–chaos” command, where 0 is the default value and 50 is already a large value.
Example of prompts: imagine/ prompt pineapple owl hybrid –c 10
The parameter “–no” can be used to eliminate particular details or imperfections from the generated images. For example, using “–no text” try to delete any text present in the image. However, this may not always produce the desired result.
Example of prompts: /imagine prompt exotic shorthair –no red-haired
To interrupt a job mid-processing, you can add the parameter “–stop” at the prompt. This produces a softer, less detailed image, which can sometimes help. For example, add the parameter “–stop 50” at the prompt terminates the generation process when it is 50% complete. As a result, the image will be quite blurry. –stop accepts values: 10–100. The default is 100.
Example of prompts: /imagine prompt splatter art painting of almonds –stop 90
The Midjourney bot uses a number “–seed” o “–sameseed” per create a field of visual noise, such as television static, as a starting point for generating initial picture grids. The “–seeds” are randomly generated for each image producing a wide variety of options each time a prompt is used. But they can be specified with integers between 0 and 429496729. Using the same seed number and prompt will produce similar final images.
Example of prompts: /imagine prompt eagle jug –seed 123
To identify the number of –seeds of a creation on Discord, just react with an emoji in the shape of an envelope.
4. Create repeating patterns and explore Midjourney releases
The parameter “–tile” generates images that can be used as repeating tiles for continuous patterns such as fabrics, backgrounds and textures. It works with template versions 1, 2, 3 and 5 and only generates a single tile. You need a template creation tool like Seamless Pattern Checker for repetition.
Prompt example: /imagine prompt mandarin fish –v 5 –tile
Midjourney frequently updates the template for improve efficiency, consistency and quality. The newer one is the default template, but you can use others (1, 2, 3, 4, and 5) by specifying the number after the parameter “–version” o “–v” or with the command /settings selecting one of the versions.
Midjourney V5.1 o “–v 5.1” is the latest and most advanced, it has a strong default aesthetic, making it easy to use with simple text prompts. Very consistent, excels at accurately interpreting natural language prompts, produces fewer artifacts and unwanted edges, has greater image sharpness, and supports advanced features such as pattern repeating.
Il modello Niji o “–niji 5” was born out of a collaboration between Midjourney and Spellbrush to produce Japanese-style anime and illustrations. It’s great to use in dynamic and action shots and character-focused compositions.
Example of prompts: /imagine prompt dalmatian playing –v 4
to know more: Artificial Intelligence: what it is and what it can do for us