Creating artwork that captures the essence of the mid-journey stage can be challenging, but with the help of stable diffusion, it’s possible to achieve amazing results. This guide will provide examples of which words to use for different art styles and types to generate mid-journey like results in stable diffusion.
Additionally, it will also include tips and techniques on how to create the perfect prompt to produce artwork that captures the essence of the mid-journey style. By understanding the power of language and how to use it effectively, People can use stable diffusion to enhance their creative process and achieve stunning results.
Prompts and negative prompt Basic For environment and objects
Prompt in stable diffusion means that it will generate the stuff that you have written in its prompt box, and negative prompt is the opposite of prompt, meaning that you will write the thing that you don’t want in your image.
So how can you get Midjourney-like result using prompts and negative prompts? I will be going over what words to use and what to look for in certain styles.
When you are creating an object or environment, it is best to stick to the normal prompt, but in stable diffusion 2.1, you need to be more specific about what you want, like if you just write “Victorian fantasy hall in the morning,” it will give a very messy and blurry image, but what I mean by specific is that you need to specify that you want the image to be highly detailed, sharp, and insanely intricate.
Example:
Fantasy land with castle on the top of mountain , 8k, detailed, sharp, vivid, soft glow, artstation, (Midjourney style) hyperrealism, highly detailed, insanely detailed, intricate, cinematic lighting, depth of field, god ray, glow, glare, cinematic: 1.5, dynamic lighting, behance, sharp focus, concept art, dreamlike: 1.5, surrealism: 1.5, spike painting, radiant light: 1.5, an ultrafine detailed: 1.9, bright colors: 1.5, illustration, concept art, key visual
As you can see in the example above, I have written what I want and also specified how much detail I want and what kind of lighting and style I want in my image.
But many times you will still get weird results, and this is where negative prompt comes in. As I said previously, negative prompt is the thing that you don’t want in your image, like a blurry image or a low-detail image, so simply write that in negative prompt.
Example:
bad, (ugly:1.6), overexposed, low contrast, cut off, tiling, watermark blurry:1.5, deformed, weird colors, mutated color, muted color, photo realistic, fused, less detail, lowres, out of frame, worst quality, low quality, normal quality, displaced object
Prompts and negative prompt Basic For Humans and characters
This is not much different from the prompt for environment, but there is some difference in some negative prompts.
For prompts for humans and characters, you need words that often get the best face or anatomy for characters, so what type of word do you need? Something like just a normal word that you would use in a prompt; the real magic will happen in the negative prompt.
Prompt Example:
Killua zoldyck transcending, 8k, detailed, sharp, vivid, soft glow, artstation, (Midjourney style) hyperrealism, highly detailed, insanely detailed, intricate, cinematic lighting, depth of field, god ray, glow, glare, cinematic: 1.5, dynamic lighting, behance, sharp focus, concept art, dreamlike: 1.5, surrealism: 1.5, spike painting, radiant light: 1.5, an ultrafine detailed: 1.9, bright colors: 1.5, illustration, concept art, anime, key visual
Because you will get some messy stuff and artifacts in characters’ and faces, you will need to add some words that will remove messy and undesirable stuff.
Negative Prompt Example:
bad, (ugly:1.6), (low contrast: 1.5), cut off, tiling, watermark blurry:1.5, deformed, weird colors, mutated color, muted color, photo realistic, fused, less detail, lowres, out of frame, worst quality, low quality, normal quality, displaced object, bad face, black and white, low saturation, ugly proportion, 3d render, flat, flat reflation, deformed face,
