How to Combine Images in Midjourney [TUTORIAL] | Tokenized AI

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
in this mid Journey tutorial for beginners I'll show you everything that you need to know in order to use image prompts to combine two or more images to create something entirely new including how to upload images the correct syntax that you should be using and how you can add text prompts to change them even more hi it's Christian heiron from tokenized AI where I show you how to navigate the exciting New World of AI software if you see value in this video make sure you give it a thumbs up it really makes a huge difference all of the links and everything I mentioned in this video you can find Linked In the description box below and while you're there make sure you smash that subscribe button [Music] thank you all right welcome back I'm going to show you now how to combine two images into something completely new in mid-journey now before we do this since the way my journey treats the image prompts has changed fundamentally in version four we're going to have to switch to version 4. technically you could just add the algorithm modifier at the very end of your prompt but I think for most of us this is going to be the easier way to do it so I've now switched to version 4. and we're going to start off by uploading a couple of images so I prepared a few things so an image of a Viking an image of Abraham Lincoln a Shaolin Warrior and a winter landscape in the mountains I'm gonna upload all four of these into Discord so that we can then get the links to those Earls because that's what we're going to need to add into the image prompts so I'm going to start off first of all by taking the copying the link to the winter landscape and I'm going to start typing my imagin prompt and we're going to add that first link as the first image prompt now we need a second image prompt and for this I'm going to use the Shaolin Warrior I'm going to copy that Earl as well and I'll add this as the second image prompt now I'm not going to add anything else now I could do that but let's first see what happens if I just combine these two and whether it actually gives me something that makes sense so let's see and voila there's our first set of images let's have a look and as you can see what it's done is it's understood let me just quickly move my image here but it's understood that the first image had basically a castle and like mountains and a winter landscape and it also understood that the second image had a Shaolin Warrior and what it's basically done is it's taking those two concepts and combined them into one and I've not given any additional direction to How it Should interpret these instead it's just gone ahead and done something right and I think in this particular case it's actually used the it looks a little bit like it's taking literally like a combination of the Styles between the landscape and the Sheldon Warrior now I'm not a particularly big fan of this style so what I'm going to do now is we're going to redo that whole process but I'm going to add some text prompts to it just to give some Direction now let me close this first the nice thing is now I can technically I don't need to recopy the Earls that I pasted into the command earlier now I can simply copy this first section because these are just basically Earl shortened or the an Earl shortener from mid-journey and if I click on this you'll see that it's actually the original image right so what I'm going to do now is I'm going to type imagine I'm going to paste those two image prompts again but this time I'm going to add some text prompts so first of all I'm a fairly big fan of using cartoon style because it just it's slightly exaggerated but it tends to spit out some really really um hyper realistic uh imagery and we want this to be in the fantasy Style and I'm going to just add realistic I'm not sure what the realistic is necessary in this particular case but let's see what happens let's wait that for that to generate all right and let's look at what it's given us now this is completely different already right now the interesting thing about this is since I did add some text prompt the fact that I added cartoon style and into the prompt I think that basically tells the algorithm that I want something more in a cartoonish style but what it also does is that most cartoons tend to be very much focused on characters so that's why I think in this case it has opted to put more emphasis onto the character and less onto the actual castle or any mountains which is completely different than if we look at the previous one where it had sort of a balanced approach right so you can already influence the weight of the two images without actually using image weight in that sense and I think these are really really great this is really the exact kind of style that I was looking for and yeah but we can actually change this even more so what I'm going to do this time is I'm going to again I'm going to copy this whole thing I'm just going to make this very slight change to our prompt imagine paste this in here and what I'm going to do now is I'm going to add a little bit of information that will say something about the type of composition that I'm looking for and in this particular case I actually don't want this to be a female Shaolin Warrior because for some reason it's been interpreting it as a female I want a male warrior that is um marching towards uh Castle I could all say a castle it doesn't really matter um and that's it let's see what happens wait for that to generate again okay now let's have a look and again it's completely different now so what it's clearly understood is that I actually wanted this type of composition you have a warrior that is marching towards a castle that is exactly the type of composition I wanted it has maintained largely the cartoonish style although it's not as focused on the character as in the previous set but it's still using as you can see the colors of the Shaolin Warrior you could obviously argue that this doesn't look much like a Shaolin Warrior anymore but of course you could change your text prompt to incorporate that so if you wanted it to really explicitly look like a Shaolin Warrior well then you would just add that little piece of information that word into your prompt and it would produce that sort of an image now um I'm gonna move on forward to the next set of images because that's also where it becomes very interesting let me just quickly go grab my Abe Lincoln um to be honest I have not yet fully understood whether there is a difference between putting the image prompt or a particular image prompt first and the other one's second I've not been able to determine whether there's any uh weight in the order of them but um you could obviously experiment around but it's very difficult to actually determine what is driving the algorithm right so um oh yeah just as a note right if you are not familiar with how to uh grab what I'm doing here I'm basically opening the image that I uploaded I click on open original and I right click the image to copy the image address right that's how I get the URL anyway so I've post pasted these two into my prompt and I'm going to leave it at that I'm not going to add any additional information and we'll see what it does when I combine a scene and a or let's put it this way more a concept and the portrait of a famous person see what happens all right so before we have a closer look let's just have another look at the original Source images so the first image that we added was a photograph an old photograph of Abe Lincoln and the other image was this like a digital art painting of a Viking Warrior and what mid-journey does when you combine them is he basically gives you a Viking Warrior with a face or at least traits of the face of Abraham Lincoln um I'm not sure in I think in the photo there were some elements that suggested what kind of clothes Abraham Lincoln was wearing and I do think that this has been in basic implemented into well it's used the information and given us kind of a weird mix here but it looks quite quite amazing to be honest let me try to zoom in I mean this is shockingly good now we're gonna add some extra information on this because I just want to show you how we can use this so instead of I'm not going to reference Abraham Lincoln in any way instead I'm going to just add to the prompt strong man defending the White House cartoon style and I'm gonna add a dash of cyberpunk because I want more of a modernish uh look to it and let's see what happens all right I think I mean these are just have a look at this isn't this just amazing let me see if I can get a bigger version of it just look what it's done so it's understood that we wanted to have sort of a protector standing in front of the White House and it's done that it's also used elements of the color scheme that is associated with cyberpunk and and it's also implemented this whole scene in the cartoon style that we set remember in the previous version of the image where we left out the text prompt um the style was primarily driven by a blend of the two Source images in this case we've told it well that source that blend looks quite good but we actually want to have it in a cartoon style and that's what it's done here I mean this is I mean this is just amazing right so you've got this super strong bulky guy and he has the traits of like slight traits of an Abraham Lincoln and um I don't know whether this actually it almost looks like this is like cyborg like um elements in his face anyway it's futuristic which is just awesome now um we could technically this has nothing to do with combining images but if you wanted to you could also um do a remix now remixing only works in version four and you need to kind of activate it and let's see what happens though I'm gonna take let me see which variation I'm going to take I'm going to take I'm going to take version two and what I'm quickly going to add is strong man defending the White House holding a red balloon and yeah that's try this one more time okay and so let's have a real quick look remixing takes a bit longer for generating but if we open it up we can see that it has kind of understood the red balloon element but it's not really implemented in the way that I would have expected but the point being is that you can blend these two images together add a prompt to get the composition you like and then you can continue to build upon a particular variation that you really really like now all of this only works in version four and just to make prove a point I'm going to switch back to version three and show you what happens if I take this prompt that first one and add these two image prompts in version 3. okay and let's have a real quick look as you can see this did not work the way we would have hoped and the reason why it does not work is because in version three the way image prompts work is that the images are used more as an aspirational Target does not mean that it will explicitly use the composition and the content of the image instead it uses the rest of your prompt and we didn't add any other text prompt to drive the image towards well more to look more similar to the style and the overall I can't really explain it to be honest but the overall style of the image prompts that you've given it and you can control how strongly that is done with image sorry with image awaits right um now clearly this does not work that we had hoped and that's why you need to use this in version 4 which is much better added clearly but in version 4 you could technically even add I'm just going to switch back real quick you could technically add more than just one image prompt oh sorry more than two image pumps you could actually use three four five it doesn't really matter just be aware that the more you add the more complex it's going to get and the less you're going to see um or the less you're going to recognize of each image in your final generated image it's much easier to do this with two and potentially with a third one some of the cases that I've experimented with I've first built something with two images and then I've taken one of the new images and I've basically Blended that with another image so in total I've used three but the prompt itself is just a generated image with another image prompt so this was how you use two images to combine them into something completely new in mid-journey I hope you found this very useful and I think you're going to have a lot of fun doing all sorts of shenanigans with this function all right I'll see you in the next video and take care bye I really appreciate you taking the time to watch the full length of this video if you found it useful and enjoyed it it would mean the world to me if you hit the like button and if you'd like to get notified the next time I release a new video remember to hit the Subscribe button just under the screen I'll see you next time and take care bye [Music]
Info
Channel: Tokenized AI by Christian Heidorn
Views: 101,628
Rating: undefined out of 5
Keywords: midjourney combine images, midjourney combine prompts, midjourney combine 2 images, midjourney combine faces, midjourney, midjourney v4, midjourney combine photos, how to combine images in midjourney, midjourney blend two images, midjourney image prompts, midjourney ai tutorial, midjourney ai, midjourney tutorial, mid journey, mid journey tutorial, midjourney image prompt, midjourney prompts, midjourney reference image, midjourney how to use, midjourney image to image
Id: PC3XtBNQAXY
Channel Id: undefined
Length: 16min 7sec (967 seconds)
Published: Wed Dec 14 2022
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.