Most of these tricks work with BlueWillow, Stable Diffusion, and many other models.
Image generated using BlueWillow
Midjourney is all over the rage right now, people are going crazy over it and trying different prompts to create beautiful artwork but If you don't know it already, Midjourney is an AI Image generation tool that allows you to generate dynamic images based on the text prompt that you provide, they call it a text-to-image tool.
Midjourney is surely a wonderful platform but there are some tricks you might not be knowing that will help you to get desired results with Midjourney and I am going to share some of those tricks that I personally use.
1. The Seed Parameter
Are you trying to create a digital magazine, write a book, or write a comic book and want to create different character images using Midjourney but all the images are different even after giving the same prompt? Well then fear not, Midjourney got you covered with seed parameter. The seed parameter allows us to create variations of the same character with different expressions as long as you are giving the same prompt (with additional details) and the same seed value.
Example Prompts:
Initial Prompt: Image of a 23-year-old male wearing a white shirt with blue eyes and blond hair — seed 30
Image Courtesy: Midjourney
Variation 1: laughing Image of a 23-year-old male wearing a white shirt with blue eyes and blond hair — seed 30
Image Courtesy: Midjourney
Variation 2: crying Image of a 23-year-old male wearing a white shirt with blue eyes and blond hair — seed 30
Image Courtesy: Midjourney
Note: the seed can contain values ranging from 0 to 4294967295.
2. Aspect Ratio Parameter:
Most of you might be knowing this already and it’s a no-brainer, this parameter allows us to customize the aspect ratio of the images generated by Midjourney. Default value: 1:1
Example Prompt: A high-tech lab in the future where some scientists are researching the humanoid robot — ar 3:2
Image Courtesy Midjourney
Valid values of the aspect Ratio based on the Midjourney model.
Image Courtesy: https://docs.midjourney.com/docs/aspect-ratios
3. Multi Prompts:
Midjourney is quite smart but it does not understand contexts, say for example you want to generate an image of a Dog with flames coming out of him or a fiery dog picture and you gave a prompt saying Hot Dog, even though there is space between Hot and Dog, Midjourney will consider it as a common prompt and it will generate images of HotDogs like this:
Prompt: Hot Dog
Image Courtesy: Midjourney
to avoid this problem we can use the multi-prompts operator provided by Midjourney i.e. double colon (::) like this:
Prompt: Hot:: dog
Image Courtesy: Midjourney
Midjourney also allows us to give weightage to words like this:
Prompt: Hot:: 2 dog
In this case, weightage 2 will be given to the word Dog and the word Hot will get the default weightage of 1.
4. The describe command:
Not sure what prompts to give but have a reference image, then you can simply ask Midjourney to describe the image using /describe command, it will give you a description of the image, i.e. what it sees in the image.
Mojrouney Describe command Demo
5. Getting the Seed and Job Details:
We talked about the seed parameter but you forgot to provide the seed, you have generated the right image using your prompt and would want to use it, then worry not, Midjourney has got your covered, you can simply add the Envelope emoji using Discord’s reaction and it will take you to Midjourney Bot Channel and it will show you the Job ID and Seed Number used for the image.
Like this:
Adding the envelope reaction
Modjourney Job Details
Bonus tip:
All the tricks mentioned above are quite useful and will help you a lot but here is a bonus tip. If you are generating images and you are not happy with the result then you can try changing the image generation model.
Midjourney supports various models like MJ1 (Midjourney 1), MJ2, MJ3, MJ4 (Default Model), and MJ5. You can tweak with models and try Niji Mode on and off to generate different results. You might see a little time difference while experimenting with different models and settings but it’s worth the wait to get that perfect image.
You can modify these settings using the /settings option of Midjourney and toggle various options.
Midjourney Settings
Conclusion:
- Midjourney is an extremely powerful AI image-generation tool that generates high-quality images based on the provided prompts.
- Midjourney provides many customization settings and various parameters allowing us to get better results.
I hope you enjoyed reading this article, You may follow me on Twitter, LinkedIn, or right here on Medium where I am active and share similar content.
