Revolutionize Your AI Image Prompts: Discovering Midjourney's Describe Command

have you ever struggled to write an AI image prompt or describe an image's style my journey has you covered with their described command in Discord in today's video we'll explore this Innovative tool and show you how it can revolutionize the way you make art in April 2023 mid Journey introduced their game changing tool describe instead of you describing what you want to see to the AI you show the AI an image and it tells you what it sees it's sort of a reverse prompt it's an image to text no more guesswork or struggling to come up with the perfect prompt and the tool is so easy to use in Discord type slash describe into the message bar then drag an image into the box or click to upload a file click return mid-journey gives you four text prompts that describe the image choose the ones that best fit and click the buttons beneath the command to generate images based on the prompt before generating the images mid-journey gives you a chance to edit the prompt it's sort of like remix you can remove unnecessary or incorrect information or add details that the AI might have missed or that you'd like to include by default The Prompt was including dash dash B5 but this may have changed check the version before generating the image my test photo is one I took in the Colorado mountains mid-journey has given me four suggested prompts and I'm going to choose this one to run I'm not so certain about the Metropolis but I'm gonna let it run anyway and here's the best result it's fine there's no sign of a city the colors are a little bit off and it didn't quite capture the composition but it's not a bad stab at the subject matter the describe tool isn't going to replicate the original image the AI is still generating creative interpretations mixing The Prompt with a seed but the new image can capture the essence of the original The Prompt gets you in the right ballpark but why might you want this tool well for one it helps you pinpoint an image's art style and maybe point you to a relevant artist or two [Music] and the tool helps you think like an AI what do you need to prompt for to get the result or maybe you just need a creative nudge in the right direction so how well does the tool work I've been trying to get a piano to fly but mid-journey kind of struggles with this concept I found an image online that's sort of what I want let's see what mid Journey says a prompt should be for a flying piano well there's some interesting ideas here robotic expressionism and surreal robotics those are two that I definitely will use in the future webcam photography and Cloud Core are also good choices but what is monumentalism and is it really iconic civil rights imagery so the individual terms may be a little bit odd but when I run the prompts do I get flying pianos well no I get some pretty nice Chrome plated pianos but their feet are firmly on the ground and I think sometimes my journey is just making up words it's a simple flower right the aic's Mana Punk pigeon core Prairie core Cottage Punk and whatever Pantone Punk is I scoff but run the prompts and the images come out just fine maybe these Punk core words add a bit of aesthetic spice let's try an image made with version 3 of mid-journey The Prompt doesn't really reflect the image let's see if describe can fill in some details Mitch uni came up with a lot of Artistic Styles like Peter morbacher and Luke fields and Chinese fashion designer guoope I love the handy little links that take you out to image searches of the artist's work I'm not sure the prompts are quite capturing the image but I can see the inspiration so maybe that's what described Us best it's a Vibe creator let's see what happens when I add an image prompt to the suggested prompt ah that gives a much closer feel but if you don't get prompts you'd like the first time try running the describe tool again with the same image you'll get different results it would be great to have a reroll button right there to generate four more prompts but I wonder about this response is my journey talking to me and here's another editorial comment from the bot I wonder what mid-journey is trying to say but are these prompts Universal does one AI talk to another let's see if describe can capture some of that mid-journey magic sauce when I roll the prompts in Leonardo or blue willow mid-journey does a good job of capturing the vibe if not the detail the Blue Willow results were more colorful illustrations but this may be because Blue Willow doesn't seem to know or reproduce artists Styles as well as the other AI but let's flip the script and take an image created in Leonardo AI can midterly replicate the style the image is more mid-journey than Leonardo but the prompt captures many of the features describe also works in both mid-journey and niji journey is there a difference the niji prompts are different but not particularly anime and the results aren't that dissimilar but Nietzsche 5 was just released and that does make a big difference let's put the AI to a real test with a difficult image from the Amadeus movie poster it's with graphic elements can the AI recognize the elements or even the image itself well it didn't recognize the poster as such though it did flag the image as an album cover the AI is struggling with the shapes and the subject at one point the AI is seeing elephants overall it missed on the Styles but I've still ended up with some awesome new styles to experiment with there are a few quirks with this new tool that will probably resolve themselves a prompt comes to you as a message from the mid-journey bot this message disappears if you generate an image based on suggested prompts then you're fine but if you wait too long you'll have to run the describe tool again and get different results also some users are reporting banned words showing up in prompts oops mid-journey's subscribe tool is a fun and easy way to create awesome AI generated image prompts there are a few bumps in the early testing but the potential is huge describe is a powerful addition to your creative toolbox letting you see into the mind of an AI and generate prompts with ease let us know how you're using the describe tool in the comments below if this video was helpful remember to like the video and subscribe to the channel this is Jen making the photo let's make something amazing together [Music]

Channel: Making the Photo

Views: 7,247

Keywords: Making the Photo, Jenn Mishra, Wits End Photography, Digital art, AI-generated art, AI image, digital photography, AI art, AI photography, best AI image prompts, How to prompt in Midjourney, Describe commmand, Midjourney Describe, image2text, img2text, Midjourney AI, Midjourney art, Midjourney AI tutorial, How to use Midjourney AI, Midjourney tutorial, Midjourney prompts, Midjourney tips, How does Midjourney work, Using Midjourney, AI Midjourney, prompt craft, canva

Id: 7DZnK1l_YtU

Length: 7min 1sec (421 seconds)

Published: Wed Apr 05 2023