Negative prompts in Midjourney, a generative AI tool, provide several benefits, such as accuracy, efficiency, and creativity. Here's how to use them and some tips for refining the results.
Using Negative Prompts
Add the “-no” parameter right after your description in the initial prompt; this allows you to exclude specific elements from the generated image.
You can add multiple negative prompts to eliminate various items.
Examples of use:
A street with the negative prompt “no people” to exclude people.
Play on the weight of the elements.
If you want to make an item appear in a smaller quantity, you can adjust that too!
Change the word weight in your prompt from -0.5 to 1. For example, “a meadow of flowers: 1 tree: -0.5".
This influences the value given to certain elements of the image; you will find yourself (with the example just above) with a meadow and a few trees.
It's a way of including an element without making it central.
Tips for successful prompts
To get the best results with negative prompts, focus first on what you want to appear in your image and not what you don't want to see.
When you notice that some unwanted elements still appear, don't hesitate to enrich your prompt with synonyms or additional details.
Be careful that your negative prompts do not contradict your main description.
If you request a busy city scene but add a negative prompt for "no people," this can create confusion for the model.
In this case, it would be more appropriate to specify "a bustling city with streets crowded with cars and buses, but without pedestrians." The most effective is to use the negative prompt “no people.”.
Warning: Too many instructions or elements can make the generated image unnatural!
Experiment and share your feedback.
Feel free to experiment with different prompt combinations and share your findings with the community. Your experience can help refine the capabilities of AI and enrich the collective understanding of how it works.
By following these tips, you can maximize the effectiveness of your negative prompts and fully explore the creative potential of AI image generation.