Creating an Advanced Prompt in Midjourney

[tmonower958]

Advanced prompts can contain one or more image URLs, multiple text phrases, and one or more parameters.


The entire prompt is now divided into three areas: image prompts, text prompts and parameters. To demonstrate the comparability of the results, we will leave the text prompt identical to our image above and now deal with the other two adjustment screws.

Image Prompts at Midjourney
As soon as an image is accessible online via a link, you can copy this link to the beginning of your prompt. This image (or several images) now serves as a reference for the image AI. Midjourney processes the referenced image and derives all the important information about what can be seen in the image. If, for example, faces jamaica whatsapp phone number are added, they will also be found in the resulting image. For our next experiment, we will take a well-known personality, Albert Einstein, to concretize the person in the image. We will use the following portrait from Bayerischer Rundfunk:

If we now place the image prompt in front of the text prompt (with a space in between), the result is the following prompt with the following result:


And now Albert Einstein is sitting in a rocket in space – amazing, isn't it? And that's not all, we can now refine the picture even further with parameters.

Parameters at Midjourney
Parameters are options that can be added to a prompt.


Parameters can completely change the aspect ratio of an image, the upscaler used and the image design. Here is an overview of the basic parameters:

aspect ratio

–aspect , or –ar changes the aspect ratio of an image.

styling

The –stylize [0–1,000] or –s [0–1,000] parameter affects how strongly Midjourney's aesthetic style is applied to prompts. Normal value: 100

chaos

–chaos [0–100] changes how varied the results will be. Higher values ​​produce more unusual and unexpected results. Normal value: 0.

exclusion

–no is a negative prompt, –no plant would try to remove plants from the image.

Upbeta

--upbeta uses an alternative beta upscaler when you select the U buttons. The results are closer to the original raster image. The upscaled image has significantly less additional detail .

uplight

–uplight uses an alternative “light” upscaler when the U buttons are selected. The results are closer to the original raster image. The upscaled image is less detailed and smoother.

For a complete list of all parameters, see the Midjourney manual: Midjourney Parameter List .

We don't want any clouds in the last picture with Albert Einstein, so we use the parameter -no clouds . We also want to get unusual pictures, which we generate using the chaos parameter. So we add the parameter and the value 50 (range 0-100): -chaos 50 .

We now add the two parameters to the end of our prompt (space between text prompt and parameter). And we get the following result:


As you can see, there are no more clouds in the image ( -no clouds) . The parameter -chaos 50 has resulted in unusual image suggestions being generated. However, you should note the following: