Amazing Art Made EASY!! | Mastering Midjourney (2024)

version six is so impressive but the difference between a mid-journey user who knows what they're doing and a mid-journey user who struggles to make good use out of the feature set is profound mid Journey has been changing a lot and just the other day they posted the Alpha version of their brand new web interface you can only access it if you've created more than 5,000 images in mid Journey so if you are a mid-journey veteran this is the perfect time to catch up and learn the features and techniques that let you create the best images possible on this tool and if you're brand new well there is still no better day than today to get started because mid journey is more powerful than it's ever been before and this video will be the perfect guy to make sure you can take full advantage of its amazing capabilities so let's see what we're working with this is mid Journey's brand new interface on alpha. Mid journey.com and I'm loving it it's such a fun and Slick way to generate images up here you have the play to put your prompts and typee anything in generate it and you can watch as it comes together unlike Discord you're not scrolling through tons and tons of things everything is in this very clean interface that fills your whole screen and you can really see what's going on what you've generated and how it all came together but the truth is the biggest changes to Mid Journey over the past couple months have nothing to do with the interface instead there has been a huge shift in the most fundamental part of creating images as you know if you want to create amazing images you have to write a great prompt writing prompts has actually changed a lot as new versions of mid Journey has been released there was a time when the best thing you could do was write something basic and then just list a series of tags that improve the quality in version six prompts like this still work but there's something off about the results and we can definitely do better version 4 prompts relied on a neat little feature called multi- prompting and this was the first time you really got fine control over the images you were generating in mid Journey what you do is write descriptions of your subject give that prompt the highest weight and add these little characters to create a multi- prompt so that you could blend that prompt with secondary descriptions of your style adding a number let you give it a lower weight and adding more prompts let you specify the quality and mood of your image without detracting from your highest weighted prompt the subject and if there were artifacts or weirdness you would also be able to add negative prompts to get rid of them again all this still works in version 6 but it's clunky a lot of work and multi- promps can really confuse the model nowadays if you see these what used to work as style hints can now lead to visible photography equipment or visible paint brushes or pencils when you really just wanted to see something photographic or artistic so we evolve again and today this is the format I use I write between two and three full sentences describing the visual nature of my scene then I list tags that push the style mood and vibe in the direction I want to go in this part the full senten is I make sure I hit four checkboxes I need to describe the style of image I want or at least the medium it was created in I describe the subject or main focus of my scene I clearly describe the background and I mentioned the lighting and color conditions this much tale written in plain English is important now because mid journey is finally able to understand the relationships between objects lighting and color if you don't explain what is being worn versus what's being held versus what's been suspended in an ice block mid Journey will have no idea where to put things and how they work together so by writing these sentences to describe the image you can clarify a lot of things mid Journey might make guesses at and in so doing get a lot closer to what you were really after older versions of mid Journey really couldn't do this at all so I consider this an upgrade even if it's a little bit more work tags are similar in version six they have evolved quite a bit you no longer need tags to indicate that images are HD award-winning or high quality these concepts are finally built into the weights of the model and so you can treat them as basically assumed what version six needs are pushes in the right Direction you use tags so that you can adjust the feeling of an image without changing those sentences you wrote If you ever used the old style of prompting you know tags are great mostly because they're so easy to rewrite if you don't like the impact purple Ambiance had on your image you can swap it out change it or delete it Al together and if you're listening to me talking and shaking your head because writing these sentences or coming up with tags seems like a lot of work to you I get it and I totally agree with you the truth is I hardly ever write problems from scratch anymore because instead I use chat GPT to do it for me if you don't know I created one of the most popular gpts in AI art generation it's called the gbit Tre art designer you might have already seen it because in the first couple days when the GPT store came out it actually climbed to the top 12 trending out of more than 3 million gpts and today it's still on the front page admittedly a little further down in the dolly section basically now I and more than 100,000 of you use it to automatically write mid-journey prompts in the exact format I just described to use it you do need to be a chat GPT plus subscriber but watch how it makes this so easy instead of actually writing a prompt you can describe your idea to my GPT like you're having a conversation no matter how much thought you've given to the idea so far you could get as specific as saying dogss wearing capes on their birthday and you get prompts in my format that create these images but you could also be as vague as saying just make me artwork that I would hang on my wall and it comes up with this diverse set of images not to mention the four prompts you get are in code blocks so it makes it so easy to click copy and paste them right into mid journey and as you can see they are perfectly formatted and on top of that all four prompts are automatically generated as Dolly images as well and my GPT brainstorms even more ideas for what else you might want to generate next all based on what you initially sent it simple or complex having done this for you is a game Cher because it lets you take chat gpt's idea and workshop it instead of having to start at the very beginning obviously you can Workshop it by changing the sentences or rewriting the tags I think you get how this works now that you understand the format you can adjust the subject and leave everything the same or keep the subject and change the background or keep both but change the tags to make the image horror style or maybe artistic or just fun and happy once you're comfortable playing around with a promps the next thing you'll need to understand is all the tools that come packaged with mid Journey over the past few months mid Journey has been giving you more and more options to control the end result of your image and just the other day the updated mid Journey so that those tools apply to version 6 as well which is so amazing come take a look at these in the new UI you're probably familiar with parameters there are only a few parameters that are fundamental to how mid Journey interprets prompts and now in the new UI they put them front and center click this little button you get sliders for each one I'll run through each of these very quickly and edit in the results to visually show you how they affect the image this is the aspect ratio and it does what you think it changes the width and height of the four Images you're about to generate the style Factor here lets mid Journey's engine take over and creatively beautifies the image the higher the value is the higher the impact this is powered by mid Journey's rating system so it's based on a somewhat averaged out idea of a beautiful aesthetic the weird factor is sort of the opposite by using it you're telling mid Journey how far you're allowed to push the boundaries of the accepted aesthetic the higher it gets the more quirks mid Journey will introduce this is the chaos Factor it sounds the same but it's a little bit different it controls how different the four Images should be from one another if it's low mid Journey will stick to one basic interpretation of your prompt the higher you set this value the more mid Journey will do to interpret your prompts in four different ways the mode option here lets you turn off the style tuning so if you know what you're doing raw mode might let you have a little more control over the output based on the prompt but it's also more likely to mess up some of the details than in standard notably standard will be better at eyes fingers Horizons and Reflections so again stick to standard unless you really know what you're doing version lets you switch to any of the older versions or other models which right now is just the nii models which were trained for creating animated or cartoon style images so this is what that does speed lets you speed up or slow down your Generations which depending on your mid Journey plan can make the GPU minutes cost per image go up or down but this has no actual impact on the quality of the image you get at the end so yeah these are the most important parameters you might also know there are a few more parameters than what are available here you can read up on those in the documentation by clicking view help documentation command parameters and tools then open up parameters if I was going to suggest one I'd have you read up on what D- node does and also about Das Dash quality these are all cool but what I think you're really going to care about is the options you have when you've already generated an image on the web interface you can click any one of your own images and all of the options show up in the bottom right of course you love to see variations and upscales these are the bread and M of mid Journey but the key features here that really open up the door into what's possible are pan and zoom being able to look left and look right and up and down as well as being able to zoom out gives you precise control over something that used to be totally lacking from mid journey and that is the composition of your final image not only can you generate an image but now you can explore that entire world especially since in Discord you aren't constrained to two options for an amount to zoom so by clicking custom Zoom you can zoom out by 1.35x and then pan to the right and then zoom out by 1.8x more doing this lets me put my character exactly in the part of the image where I want him and with remix mode on something you can access by typing SL settings and Discord you're able to change the prompt as you go so while I'm zooming out I can make adjustments to the scene and if you get how this works the flexibility is amazing just watch as I zoom out here while this happy little swan turns into a Halloween nightmare now those are cool and really open up the possibilities but not nearly as much as this last feature that really turns you into a wizard and by that I mean has actually turned me into a wizard if you watch my other videos you might be able to guess I'm talking about my new favorite feature very region this is the feature that lets you erase a part of an image and regenerate just that portion and with remix mode on well that means you can change the prompt for the part you're regenerating this feature of course lets you generate something and clean up any part of it you didn't love you can fix up text that ended up being spelled incorrectly remove something give a cat sunglasses or if you have a prefer option set up for your favorite character you can erase whoever it is you just generated and put your favorite character in that scene instead I have a whole video teaching you how to do that and version six is so powerful it makes this so much easier I hope now that we're near the end of this video you feel more confident about writing amazing prompts and combining mid Journey's tools to create great images mid Journey has slowly been turning into the feature Rich art tool I always dreamed it would become please go ahead and try out my GPT I promise it makes everything I just taught you so much easier to get into and try before we go I do want to highlight one more part of the mid Journey documentation which is the commands mostly these are quality of life things right now they only work in Discord but they can really speed up your workflow if you're creating lots of art the truth is if you want to be a mid-journey pro you should really be familiar with every option parameter tool and command and you know what I think I think the best way to start is by watching this video next see you there

Channel: Glibatree

Views: 10,553

Keywords: GPT-3, AI Demo, AI Story Telling, AI Art, Midjourney, ChatGPT

Id: 8JyYPFCUGRc

Length: 13min 38sec (818 seconds)

Published: Sun Feb 04 2024