Midjourney also has a few extra settings. “Simplicity,” “Clean lines,” “Minimal colors,” “Negative space,” “Minimal still life.”īut really, you’d start off with the WHAT first, describe the scene and the details then add the style keywords.“Emotive eyes,” “Intense gazes,” “Contemplative mood,” “Expressive gestures,” “Stylized poses”.“Majestic mountains,” “Lush forests,” “Glittering lakes,” “Desert dunes,” “Golden sunsets.”.“Dreamlike,” “Surreal landscapes,” “Mystical creatures,” “Twisted reality,” “Surreal still life.”.“Vibrant colors,” “Geometric shapes,” “Abstract patterns,” “Movement and flow,” “Texture and layers.”.Here are some basics keywords where I didn’t even add a subject! If you don’t pick ANY style, it will guess, and it may not guess what you want. Experimenting with different keywords and styles can lead to surprising and unique AI generated images and art. You can have a realistic close-up portrait, but if you do a big scene with lots of detail, it’ll get kind of close but not perfect.įor best AI art results, it’s important to provide clear, descriptive text-to-image prompts and to adjust the settings in the AI tool to your desired level of customization and control.
#Rendered imagepack free cities full#
You can use full sentences or not, but the more complex, the trickier it gets. WordPress is combining my doubledashes to an m-dash).Īt the basic level, just describe what you want to see. (You need to add 2 short dashes – – and the spacing has to be right. The sizes need to come at the very end, after the whole prompt, with two dash, you can copy/paste these: The default size is square, but you can get a landscape or portrait (tall or wide) with these. You’ll probably want to test sizes and ratios too. But you can also make logos, sketches, tattoo or comic book art, children’s books… the sky is the limit and the results are insane. Mostly likely, you’ll want either an artistic, illustrative style or a photorealistic style.
Then add styles until it’s closer to what you’ve imagined. Keep things simple in the beginning, just describe what you want to see. So these prompts will help you get what you want faster, but they’re just a starting point for you to develop “ethical” Midjourney prompts that are more unique. Note: it’s dodgy, but not illegal, to copy an artist’s style directly by using their name (for living artists at least) but you could also just paste in 2 pieces of artwork and reproduce something much closer to their originals. I’ll go through soon and make examples of some of these so you can see the differences, but here’s a handy reference guide. What are the best text-to-image prompt keywords for unique types of art and photographic styles? For best AI art results? How about architecture, fantasy art, character renders and logo design? I’ve been playing with Midjourney and other text to image generators built around stable diffusion and I have my own prompts to get the best quality images, but recently I let chatGPT3 help me brainstorm some new ones… so here’s a huge list! Scroll all the way down for new prompts, V5 examples and best midjourney prompts for stunning results. Update! Version 5 just came out, and it’s more photographic and less artistic.
0 kommentar(er)
