Crie quadrinhos e mangás com IA: Midjourney, Photoshop e automação de IA (não requer habilidades de desenho)

1. Course overview: Welcome to this course. This course is divided into three main chapters. In Chapter one, we will cover the basics of Mid journey, including the key commands, parameters, and the new Mid Journey editor. This chapter is perfect for those who have never used Mid journey before. However, if you are already familiar with it, you can skip this section. After the first chapter, you will be ready to learn what I call the ultimate formula. This is where the real magic happens. In Chapter two, will introduce this formula discussing what we need to create in order to apply it effectively. In Chapter three, we will dive into the formula. This formula will turn mid journey into our own magaka or artist, making him takes everything we tell him into consideration when creating our images or comic panels. This includes aspects such as image composition, character poses, intense exhibitions, camera angles, art style, the scene where our story unfolds and the other things that we can do to create a new image. This formula will be used to create a new image and more. Then we will learn how to refine and polish the images mid journey created in Photoshop and how to enhance our workflow by creating both dump assets and pro assets. Finally, we will put everything together and you will learn how to create a full professional looking comic page from scratch using Photoshop, including the speed lines and speech balloons inside Clip Studio. In this course, you will have unlocked a powerful new skill, the ability to create visually stunning comics without needing to draw by hand, whether you are a storyteller, your writer, or someone who is always dreamed of showing the word your ideas visually. This course will empower you to bring your vision to life. Your story deserves to be seen and now and only now with the power of AI. It can be I can't wait to show you the ultimate formula, but before we jump into it, let's start first with Chapter one. Mid journey basics. Let's get started. 2. Why Midjourney: Y Mid journey. A lot of choices are available now when you want to generate AI images. But for me personally, I found Mid journey to be the best choice for making a comic book, and here is why. Reason number one, Mid journey built in enhancements, one of the standout pros of Mid journey is its ability to automatically beautify images, ensuring that your images are always visually appealing with minimal effort. This means you don't have to spend time tweaking every little detail. Reason number two, no need to spend a ton of money. Unlike some other image generation tools, mid journey doesn't require you to have a high end expensive computer setup, like in the case of stable diffusion, where you need to invest in costly gear to get the best results. But with mid journey, everything is handled online, you can create amazing images without breaking the bank on a new machine. So while it's true that stable diffusion offers more control over specific details and images, I think mid journey has the best of both worlds. It strikes a perfect balance between ease of use and control over the image, especially after the recent updates like style reference and character reference, which Go which we will talk about them in depth later on. In a nutshell mid journey lets you create stunning images fast. You can play around with different styles without getting stuck in all the technical stuff. This means you can focus on what truly matters, telling your story through visuals, while mid journey handles the magic behind the scenes. Now let's dive in and start the journey of learning mid journey. Yes, it's a journey within a journey. See you in the next one. 3. Character Design Introduction FREE PREVIEW: Mid Journey had released a game changing update introducing the character reference feature. This tool is perfect for storytellers looking to maintain consistent character appearances across different scenes in visual comic narratives. So if you are as excited as I am, let's jump right in. We have hit a critical milestone in AI comic creation by introducing character references. We are getting closer to our ultimate formula guide for creating compelling characters. First, let's talk about character design. Understanding your character's look and personality is crucial before letting mid journey work its magic. Ask yourself, how does my character look? Do I have any inspirations or ideas? Don't be discouraged if your character resembles an existing one? Inspiration from other works is a common practice and enriches your creations. For example, did you know that Mariam from Hunter Hunter was inspired by Freezer from Dugan Ball? The key is to ensure your character reflects your story and style, even if inspired by others. Infuse your personal touch to create something unique. And there are several ways to design your character in mid journey. One, using a text pro. This is the simplest method two, blending two character images, create a new design by combining elements from different characters. Third advanced design with Photoshop for more control and customization. Let's start with the first method. Using a text prompt, you have two ways to generate this prompt. The first way is by using Chad GPT. The second way is by using slash Describe. In the next lesson, we will start generating our text with the first way using Chad GPT, Cu there. 4. Midjourney Account Setup: To start using MD journey, you need a discord account. First, head over to discord.com. Once you are there, you can choose to open discord in your browser or download the app. But the browser is just fine. Next, register and verify your account. Once you are inside discode, click the plus icon on your left and join a server and paste the invitation link you see here. Hit join server and boom, you are in the channel. Now before we start using Mid journey, we need to subscribe to apelanF. Visit midjourney.com and log in using your verified discord account and choose a subscription plan that suits your needs. I had subscribed to a basic plan. You can see now after you subscribe, there are two options to generate images with Mid journey, one on Discord or two on the website itself, which is easier and have more appealing interface. But there is a catch. There is a very important advantage in discord that we will use in the AltuitFmula. Put that in mind for later. Now mid journey website has no limitation. Before that, you couldn't use the website before you have created at least 100 discord generations. But recently, Anyone can use the website now to know how much images you have created, simply go to Discord and open Mid journey boot, and then hit slash Info and click Enter. You can see how many images I have created in my account. Also, you can see how many fast time remaining here. I have 355 minutes, which roughly equals to 355 images that I can create. Once you set up your discode account and subscribe to mid journey plan, you are all set and ready to go. In the next lesson, we will get hands on and create our first image in mid journey. See you in the next one. 5. Basic Commands: All right, folks, let's dive into discord and open up mid journey channel. Go to any of these newbies rooms. You can see other people, generations and cool ideas. But if you decide to generate your images in these public rooms, caution, it will get messy very quickly. So to chat privately with Migo Nebot click on Mi Journey username on any thread, then send a message. Now you are chatting with Miourn directly with no distractions, but you should know that your generations will still be showing publicly in a random newbies rooms. Unless, of course, you have subscribed to a pro plan or mega plant where you can create images in stealth mode. When you type slash, you will get these options. These are called commands. Don't make it hard for yourself. You don't need to learn all of them at once. To make a comic book, the most important commands and or used ones, you can see here on the right. Image is to generate an image from text, then we have describe. It's the exact opposite. It gives you a text from an image. We will talk about this later and we have Berlin, which lets you mix two images and create another one from it with no text input. It's also great for experimenting. Then we have slash Info, which as we saw before will give you information about your subscription, including image count, then your subscription, including image count, right? Then we have settings to access settings. It's actually type slash settings and you can see bunch of mid journey settings. These will be applied by default to any generation you create. First here, we have a bunch of different versions, but you are only going to need to use the best two version. One is Version 6.1 Alpha. This is the latest and greatest from mid journey. It offers stunning realism and detail with improved color, accuracy, and texture quality. Then we have Mig six. This is ideal for anime style, manga style or webtoon. This version is ideal for creating our comic book. But what if you don't want an anime style? What if you want to create a visual story in say three D Pixel style, then you should use Version 6.1 instead because it's better at generating three D images. Like in this example, I have generated the same prompt, but with different versions, a magical first with glowing creatures, pixel style, you can see in the Niagi example, the image don't look like pixel style at all because Niagi is trained on two D anime style. Where in the second example, where we use Version 6.1, we instantly got a scene in pixel style. Choose whichever version you like for your own project. In this course, as we are going for more anime manga manhaa style, we will more likely to use Niagi version. Next, in settings, we move on to raw mood. We will experiment with this mood in depth in the ultimate formula. But in summary for now, raw mode, as you can see, is not activated, which means mid journey beautification mode is activated and will be applied to every image. If you don't activate raw mode, this means that you don't need to worry about typing a very detailed prompt or add many parameters. In this way, mid journey will create beautiful images with very few words. However, this beautification also means that your text prompt has less control over the text input. So eventually raw mood has advantages when you have a very detailed action or scene that you want to describe in detail and watch mid journey to follow it more. If you want to generate an image in style row you don't need to open settings every time you need to do so. Simply, you can activate Style row by typing dash style row after your text prompt. Now mid journey will focus more on your text. Let's take a look at an example that explains the difference with and without raw mood. Let's hit slash Imagine space. Then the prompt is a kid drawing of a cat. Then we write the same exact text prompt, but with style raw. Which do you think resembles a kid's drawing more with or without style row? Without raw mood, Mid journeys aesthetics gives us more of what we ask for and this is a bad thing. But in the amod images, even though it's all colored, it's clearly very simple drawing, we actually can expect a kid to draw these images. In this case, with amod, I think mid journey gave us a better representation of a kid drawing. There is without raw mood, these images are harder to draw except maybe for number one, look at the image number three. No way, this is a key drawing. In another example, my prompt here is a dramatic fight between two rival warriors, swords clashing and dash style raw. I'm using mid journey version six. Then I created the same text prompt without style. To be honest, I actually like the images without Star ro this time. They are more artistic and beautiful. However, there are some cases and tricky manga panels that you will need to activate Star o to get better results. We will see more examples later on. Now let's see these stylized options and they are actually related to style raw. Again, instead of the need to open the settings, every time you want to change the stylization, you can simply do that directly by typing a then space and choose a number 0-1 thousand. The default is 100. It's the same as setting the style to media. This means every image you generate will use s 100 and the higher values, then that result in more stylization, but will follow your prompt less. If you think about it, low stylized is a close cousin to style raw. For example, space, zero, and style row are doing the same thing. They both aim to reduce realistic and direct interpretations of the prompt with minimal artistic influence. We will go and dip about them in the ultimate formula. Back to our settings, we have turbo mode, which generates your job very fast, but will cost twice as GPU hours compared to fast mode. Then we have relaxed mode. It's slower, but it won't consume GPU hours, but relaxed mode requires a higher subscription plan. If you are still confused about the settings menu, don't rarely would need to open it. Need to open it if you know what you're doing. If you understand, for example, that if you want to activate style raw, you can simply type dash style raw parameter after you tick scroll instead of opening the settings. All right. Now that we have created our first images inside Mid journey and learned some of the important commands and the settings menu, except variation of course. In the next lesson, we will create our first image, following what I call a basic comic panel. Tick spond and we will learn the upscale and variation tools as well. See you in the next one. 6. Create our first story type text prompt: Now we are ready to generate our first story scene or comic panel type image. To create an image, type slash Imagine then Hit space, a prompt will show up. Let's start with just a text prompt without any preferences or parameters. But be cautious. I'm not just going to give you a simple text prompt. Remember, this course is all about creating professional AI comics with Mid journey. So even though this is our first example of the course, I want to make sure that I'll show you a structured text prompt that truly showcases what you can expect from this course. A text prompt can be literally anything. A paragraph, a sentence, a word, a letter, or even an Imoge. Let's try something like a cute elephant using his trunk to playfully spray water with a whimsical, enchanted forest. Background. This text prompt follows a story prom formula. Most comic panels are structured into this triangle formula, one, an object or more, two an act, and three, a background. In our example, the object is the cute elephant. The act is that it's using its trunk to spray water and the background is the enchanted forest. Now the problem with this formula is that if you want to create, let's say, consistent comic panel is that your artist is an AI. Suppose you are working with a real person as a writer, and he is the Bengal cow of your story. First, he will ask you how this elephant looks like. Tell me its color, features, body type, the style of the drawing, or any insignificant notes about the elephant. Then he will ask you how you want the style. Is it simple? Is it the cartoon, perhaps three D pixel style, or maybe even anime. Not only that, but he also needs to know what type of story you have. Is it light hearted for children, or maybe your story is more dramatic and has deep themes, or is it a showing type or even horror where your cute elephant might be a secret creepy monster who transforms at night. Once you tell your Mangaka all these information, he will always have loose parameters, settings or whatever you wish to call them in his mind. In every panel he draws. Now, back to AI. It's true that we told mid journey by the text input that we want a cute elephant, but which cute elephant will it create and in which style? AI is trained using large datasets containing thousands or even millions of images. So unless you specify what style or shape you want, AI is going to choose the style and shape randomly, and random is a bad word for a comic book. But don't worry about all this now. We will make Mid journey act as if it's an actual Mangaka drawing your comic. We will make him take all your considerations in each panel. All this will be in the ultimate formula. But for now, we need to learn the basics of Mid journey. So now that we have created our first image inside Mid Journey, in the next lesson, I'm going to show you an amazing example. I saw while I was scrolling in one of these Newbies rooms, so see you in the next one. 7. Upscale and Variation buttons: While I was scrolling inside the Newbies room, I stumbled upon this cool example. Someone used Mid journey to create a comic book style illustration of Norse good Odin, and it's amazing example for using Mid journey in the best easy way. I mean, look here. He just used a simple text prompt without using any references or anything. He simply typed an illustration of Norse god Odin, wearing an ePatch in dark comic book style. When you generate an image in mid journey, it automatically gives you four options. You can pick whichever you like the most, or you can choose them all if you like. Then below the images, you will see these four U buttons. These are for upscale. You can upscale your image to download it or to remix it. Then we have four V buttons. These are for variations. Suppose you like an image from the grid. If you click on the corresponding V button, you will get similar varied image to the one you choose. Let's see together what this guy did. First, out of all these four images, he obviously like this one the most, so he upscaled it. But which upscale button did he click on to upscale this one? Well, it's simple. These images are in order like this. Top left is one, top right is two, and here is three, and here is four. This guy liked image number one. He clicked on you one button. Then he got this image upscale, but there is still a problem in this image. Can you spot it? The guy asked for the Norse god Odin with an eye patch. And while he got the good Odin, he the ePatch was still missing. This is where the variation buttons or V buttons come into play. That's what he did. After he upscaled Image one, he clicked on Variation Strong as showing you can also do the same thing. If you hit V one image on Image one, but pay attention. This is only if you set your variation to high variation mood inside your settings. To show you real quick, we are back to settings. If you have variations set to high, this means that you converted all your V buttons to very strong. So in our example, to add an ePatch on the good Odin face, you need a strong variation to give me journey more space for altering your image. What happens when you hit V one, you will get this remix prompt interface. If we go back to our settings, we will find that our remix button is activated. So what exactly is remix? It's simple as this. When remix mode is activated, when you click on variation on any variation button, this remix window appears allowing you to alter the text prom. Miji now will use this image as the starting point. Then he will compare it with the new refine ak new text prompt and see what is the difference between the original image and the new text prompt. You can even leave the text prompt just as it is, and this is what will happen now. Mijei will first look at our text prompt. He will see this an illustration of Norse god Odin. Then he will compare this text with the original image. Metaphorically, of course, he will say yes, this is just right. The image is an illustration of Norse god Odin. Same for in dark comic book style. He will go back to the image and will say yes. This image actually in a dark comic book style. So he will not change anything. But then only finally, when he gets to the part of the eye patch, he will say, oops, this image does not have an eye patch. So the text prompt doesn't match 100% the image that I created earlier. This is where the remix will fix your image and we refine your image even further. So whatever is missing from the remix prompt, he would try to add. Now, here is our final result or should I say our guy final result. Look, there is at least three out of the four images that have a God Odin wearing an eePatch. So with only two steps inside my journey, this guy just got the image he was asking for. In summary, he first created a prompt. Then he noticed that his favorite image is still missing an ePatch. He clicked a very strong with remix mode activated. But of course, he did not change the remix text prompt anyway. As he wanted new journey to look again at what is missing from the image by matching the text prompt with the image better. In the next lesson, we will take a quick move to learn some of the important parameters that we are going to need in the next chapter. See you in the next one. 8. Midjourney Parameters: Wondered how to create an image when you're not sure about the style or how to generate multiple images from a single prompt. Well, here are two fantastic ways to do just that. By adding double dash repeat after your prompt, followed by a number, mid journey will generate that prompt multiple times. For instance, if you use repeat four, the same prompt will be repeated four times. How many times you can repeat depends on your subscription. With the basic subscription, you can repeat up to four times. A standard subscription lets you go up to ten and with the pro or media subscription, you can repeat up to 40 times at once. As you can see, we get four different tasks from one prompt only. It's great for experimenting with multiple versions of the same prompt, but be mindful, it'll consume four times your GPU. For example, you can also use C brackets and commas for permutations. This lets you create several different prompts by changing certain parts of your prompt, but still just one prompt overall. Let's say you want to see how a warrior, a monster, a girl, or a samurai would look in the same scene. You can do this with just one prompt. Cool. This is just one of the many parameters I'll be covering in this lesson. Let's keep going. Next up is chaos. By adding chaos, followed by a number 1-100, you can control how much variation there is between the images in your grid. A lower number means less variation, while a higher number means more. For example, chaos zero gives us a more consistent image while chaos 100 makes the images in the grid quite different. You can really see how chaotic the images become with higher chaos values. However, be cautious. When using chaos 100, mid journey can completely deviate from your prompt. For instance, I asked for an enemy protagonist with a fierce expression in the style of my hero Academia, but none of the images quite match the prompt. Chaos is great for experimenting and creating unpredictable results, but it may not be ideal for all styles, especially if you're aiming for a specific look like a comic book. Maybe try using a lower chaos value like 20 to get results that are closer to the prompt. Here's the same prompt, but with Getc chaos 20, you can see the difference. You get four diverse images that still remain closer to the prompt, but with more variation than before. Moving on, we have a Bami weird. Adding Add weird followed by a number 1-3 thousand to your prompt will create increasingly bizarre results. For example, weird one isn't very strange, but weird 3,000 produces highly unusual images. Even at the highest weirdness value of 3,000, I can still see bamboo forests and characters holding katanas in three out of four images. Despite the weird results, they still follow the prompt better than chaos. The weirdness parameter is great if you want to create a unique scene pose or moment is definitely a fun tool to try out. Here's a look at the images, starting with the lowest weird value and gradually increasing it. You can see how the same prompt evolves as the weirdness score increases. Now, another powerful parameter is double dash stylized. By adding Apple stylized, followed by a number 1-100, you can control how much artistic styling mid journey applies to your image. A low stylized value will give you simple basic images that closely match the prompt. A higher value like 1,000 will make the image more artistic. Let's try a child's drawing of a smiling sun with different stylized values. As you can see, at multi stylized zero, you get the most realistic child's drawing of the sun. But as the stylized value increases, mid journey adds more of its artistic touch veering away from the realistic look. Here's the difference as we go through the different stylized values. You can see that the higher the stylized value, the more beautiful the image becomes, but it's less true to the original prompt. For the final example, I've pushed the stylized value to the extreme to get the most accurate child's drawing of a smiling son. By combining stylized zero with stylized 1,000, this is what I got. It's hands down the most realistic child's drawing of the sun among all the other examples. This shows how the stylized parameter and other parameters like double dash weird and double dash chaos can affect the final result. You can experiment with different combinations of these parameters to achieve diverse and unique results. Speaking of combinations, let's try an idea. I'll combine all three parameters, Cal conta stylized zero for realistic drawing, chaos 20 for more variation, and with 300 for something a little different. What if this kid is a bit unusual and creates something unexpected for his teacher? Check this out. We get stunning unique drawings of the smiling sun, each one reflecting a different imagination. It's like stepping into a child's drawing class where each kid's drawing is a bit different, yet they're all based on the same prompt. Understanding mid journeys, aesthetic parameters, unlocks endless creative possibilities. Master these tools and you'll be able to push the boundaries of what you can create with mid journey. But if you want to adjust the aspect ratio of your images, you can do that too. Just add AR followed by the width and height ratio you want. For example, our one to one will give you a square image. R 16 to nine will give you a wide screen image and nine to 16 will create a portrait image like a YouTube short. Now, let's talk about speed modes, turbo fast and relaxed. These modes affect how quickly your images are generated. We've talked before about selecting them in the settings menu. But another way is by adding double dash relaxed after your prompt to use the slowest mode, which doesn't consume GPU time. It's only available with at least a standard plan. First is the default speed and Turbo renders the prom four times faster than the first mode, but it consumes twice as many GPU minutes. You can see how fast Turbo mode is. It only took about 12 seconds to fully render impressive. We've covered a lot of mid journey parameters today, but there are still more to explore. In the next chapter, we'll dive into the ultimate formula and discuss the differences between this code and the mid journey website. See you in the next lesson. 9. Vary region and other refinement options: Let's go over some of the quick refinement options available in medi on decode before we move on to the website. Let me show you how it works. First, I created this image. Don't worry about the SRIF and the image length for now. When I generated the image, I received four grid options. I liked image three and four, so I upscale them when you hover over one of the upscaled images. You will see two traditional upscale options. Don't be confused. The first upscale we have clicked on is not a true upscale. It simply allows Mid Yoni to let you pick one of the four images to view alone. Now you can refine them further and perform an actual upscale. We have these two upscale options, upscaled subtle, and upscaled creative. If you want to retain all the information in the image while upscaling, use the upscaled subtle. However, if your initial image has some distortion or unclear objects, then upscaled creative is the better option because A, I will regenerate the unclear pixels and potentially fix them. In my case, I am satisfied with the first image as it is, so I am choosing upscaled subtle. We also have very subtle and very strong, which we discussed previously with the example of the god Odin with the ePatch, if you remember. Additionally, you can zoom out to expand the image. Mid journey will reimagine the scene as if we move further back. This isn't always perfect, but you can customize the Zoom by opening the remix tab and adjusting the aspect ratio. You can even make the image square. Instead of zooming out and creating new pixel information in all directions, you can increase the image in one direction only. You can pan left to right, up or down, and you will get this remix option allowing you to add a new text input to help Mid n, feel the new areas that you want to expand. For example, if I want to see more of the elephant's body, I will click and write and leave the remix prompt as is. As you can see, this is our images. After we pan write, we can now see more of the elephant and also favorite your image or re open it on the website. Now let's talk about region. If you like most of the image but want to change just a part of it, I think the butterfly is cute. But what if we want to remove it? Then we simply click Very region. Now we have two drawing tools to select the target area that mid Yoni will work on. We have a squared tool or a lasso tool, similar to those found in Photoshop. Let's select the butterfly. In the prompt, we add dash, no. Followed by the item you want to remove, separated by a comma, if you want to remove multiple items. You can add dash dash, no anywhere after you text prompt or anywhere between other parameters. But make sure that every parameters are separated by a space. For my example here, I will type dash dash, no space butterfly and hit Enter. This will remove the butterfly from the image. If it doesn't work on the first try, you can rerun region until it's fixed. You can see now the butterfly is perfectly removed. Now, if you like any of these images, you can upscale it and refine them even further. Now, we are ready to move on to mid journey website where we will learn more advanced parameters and references. Before we start our ultimate formula and creating our comic book assets, see you in the next one. I 10. Midjourney website , The chapter 1 finale: Why Mid journey website is a fantastic alternative to using this code. If you have found this code complicated, Mid journey website offers some tools that make the process much simpler and more intuitive. Now we will dive into these features, but first, here is a quick note. As of now, you need to have generated at least 100 images on this code before you can use Mid journey website. Let's start with the basics. On the Mid journey website, you have a range of options that eliminate the need to remember complex commands. Under the Create tab, you can set your parameters without typing unnecessary commands like slash Image. Everything you need is available on a user friendly interface. So instead of typing dash dash before you adjust any parameter, simply by opening this setting step, you can adjust the aspect ratio stylization values like stylize, weirdness and variety, which is that if you are unsure about the parameter, just hover of it to see an explanation as shown, variety parameter is also known as chaos. So here you have the three aesthetic parameters and you can adjust them with the sliders right here, of course, you can reset them all back by clicking on this reset button. You can also still use dash dash in the creation tab itself, like you used to do in this code, whichever is more suitable to you and you can choose between standard or style role effortlessly. But this is not why you are moving to the website because most importantly, image prompting, style and character references are much easier in the website. These in specific, we are going to use a lot in our ultimate formula. One of the biggest advantages is the ease of image prompting. You can upload an image or simply drag and drop one to use as a reference, either from your device or from any image on the mid journey website. Compared to this cord where you have to copy the link of the image and paste it in the prompt. And to create an image, simply type your prompt without slash Imagine and hit Enter. For example, type a pirate on a huge ship and hit Enter to see your image being created. You can easily re run a previous job by simply re entering the prompt and getting different results. You can also copy options like copying the prompt, job ID, image URL, and more are available with just a click. You can customize the interface to suit your needs, enable or disable various tools to keep your workspace organized. Tools like repaint and reframe and even more versatility. Repaint is equivalent to region. You select an area in the image and modify it. And then we have a frame tool. You can change the aspect ratio and preview changes in real time. I can adjust the image with precise control. I can make it a wide screen, and I can I can choose to either put the image at the start or the end to expand the image in any direction I want, giving me a visual representation of what it will look like exactly, showing how much space will be left around the image. This couldn't be performed in this court. Look what I did in this example in two steps. I have reframed twice to create a unique composition of the image. I expanded the image to show the full body of the girl and the elephant. To create a unique scene and composition with simple steps. These tools will come in handy when we start to create assets for our comic book. If I like an image and want something similar, there is a search tool that finds similar images created by others, whether in style, composition, or overall tone, it's really handy. You can see how it's easy to bounce around and create images on the mid journey website with just a few clicks. Instead of remembering a bunch of parameters and typing them into this code. But still, I recommend using this code sometimes, especially if you are generating on your phone. The website isn't as mobile optimized yet and used in text in this code can be more efficient. If you are paying attention to what I'm doing right now, I found one of the similar images that I liked, then I dragged it to my creation tab and used the image as a style reference. My text prompt was a girl in front of a cake shop and I wanted this image to be in a certain style. Remember, when I said that A, I is trained on millions of data and images. Unless I specify this style, mid journey will give the image in a random style. So there are two ways of specifying the style, either as a text prompt, just as we saw before, where we can write at the end of the prompt, something like in the style of Disney or in the style of Demonus layer enemy. Or you can type bunch of words that describes the style of the image. This one way, but the other more focused way without additional text is using a style reference. This will make mid journey replicate the style of an image. You give it and provide your image in that style. This is just a glimpse of what we are going to learn in the next chapter. I think we have talked enough about mid journey basics. Oh my water relief. I won't lie to you. I found making Chapter two a bit challenging because it's about basic mid journey lessons that many of you might already know. But here is the exciting part. We are now moving to where things will get spicy real quick. We will push Mid journey to its limits and create our comic book assets. I will share my little secrets and techniques to take Mid journey to the next level. The next chapter is the real beginning of our course, and I'm so excited to share everything I have learned with you. Buckle up and get ready for the real magic to happen in the next one. 11. The ultimate formula introduction: We should know before we start the ultimate formula. We have talked about how to create images from text, including some of the parameters and options that will alter the image. But text prompting, including the aesthetic parameters is only one of the inputs you can give mid journey to get the result. There are other three key inputs that will help us create consistent common panels in our story. The first is reference image and its image weight parameter. Second, the style reference and its style weight parameter, and finally, third, the character reference and its character weight parameters. We will talk about how each of these inputs affect the output image in terms of composition, style, character pose, and more, and how they affect the image when combined. After that, we move to the second thing we need to know how to get to the desired results. If our first attempt didn't work, and we will discover how mid journey sees an image. In terms of the composition or style or everything to learn from it, how to write the perfect text prompt. Even if we are bad at text prompting. All with the power of the magic discord command slash describe and then refine our image even further to get our desired image for our story. Once we learn the other three inputs and learn how to use a reference image and get text prompting insights from mid journey, we will be ready to start the ultimate formula. The ultimate formula is the formula in which we take several steps inside mid journey to create a bunch of organized consistent images for our story and make them ready for using in Photoshop to make our comic book. In the next lesson, we are diving into the three image inputs and seeing how the shape of our final image. This is going to be awesome for creating all kinds of scenes and poses no matter what direction or idea you have in mind for your story, can't wait to see you there. 12. The second Input - Image prompt: Is an image reference or image prompt? An image reference or image prompt is part of the prompt that influences a job's composition, style, and colors. You cannot start generating an image with only one image link. You must either add multiple image links or a single image link along with a text prompt. For example, you can just copy an image link, paste it in the creation tab and press Enter. You will get this error. You need to add some text prompt, even if it's just a single letter. Okay, so where to add image links in our prompt. You can add it literally anywhere between or after your text prompt, but it must be before your parameters. Ideally, most of the time, you will add the image link at the start of the prompt before your text. Of course, all this in case of using this code. If you are using mid journey, you don't have to worry about it. You can just paste your image link and it will automatically be activated as an image prompt aka image reference. You might now say, Well, thank you very much, but you didn't answer the big question, what is image prompt and how the effect the output image. Well, thank you for reminding me about what this course is all about. I can answer this question in one full hour, maybe more or I can give a simple answer without overwhelming us with unnecessary information. We only want to get good at creating image for our comic book. Based on my countless trials and errors with using image prompt, aka image reference, either by itself or with other inputs, like style reference or character reference. I will give you the very simple answer for this question that you need to remember before you start creating any image for your story. Based on my experience, image prompt at the highest image weight influences the final image as follows. 70% composition, 25% style and color, and 5% character. Image weight is a parameter which is IW. And it is used to adjust the importance of the image versus text portion of your prompt. The default image weight is one, the highest value is three, and you can add a very low value such as IW space, 0.25. Also, you can't add image weight parameter without an image link, which makes sense because these two belongs together. What is composition, the prompt position in our comic book. Is the arrangement of elements within a scene to create a visually appealing and effective narrative. This includes the placement of characters, backgrounds and other visual elements to guide the viewer's eye, convey emotions, and enhance the storytelling. Good composition helps to focus attention on important parts of the scene and create a sense of balance and flow. So you should ask yourself as a storyteller and a comic creator. What do you want your viewer to see in the next panel scene of your story? If you really want to see how image prompt affects the composition of the image and how it works in the prompt at all, I should answer this very basic question first, which is how may journey or any other AI tool create an image. It's pretty fascinating actually. Let's break it down. AI image generation starts with something called noise. Think of noise as a canvas covered in static like an old TV with no signal, the AI then takes this noisy canvas and bit by bit transforms it into a recognizable image. Using the data and millions of images we have talked before. It's like watching a foggy picture slowly come into focus. Now here is where it gets interesting. When you add an image prompt, you are essentially giving the AI a head start. You are saying, Hey, start with this image instead of a random noise. This influences the final result significantly. Great news is that I can literally show you the process in the early moments as my journey start creating the image. I mean the noise part. We can use this by one of these two parameters, stop or dash dash video, and I will choose dash dash video, which will give us a video showing the initial image grid being created. To see the video, we must use this code. First, go to this code and type any prompt you want, but add at the end video. Now, after the job is finished, go over to reaction and press on the envelope icon. If you can't see it, then simply type envelope and you can favor it. A. Once you press on the icon, the video link will appear. Now I will show you several examples of videos of the AI creating image grid from noise in real time. We will show you these videos with and without image prompt. To see how image prompt significantly influences the starting noise and guiding AI to create an image in a very similar composition. Now, before we see those examples, first remember, most of the time, we are going to use image prompting for composition. This is the main use and we are going to use the highest image weight. So if you are serious about a specific composition in your mind, IW space three is a must. In the next lesson, we will look at examples with and without image prompts. This will help you see how they affect both the composition and the style of the final image. See you there. 13. Image prompt examples: Let's see this example. This is a detailed text prompt without a reference image. The prompt is a screenshot from the game Dark Souls, a bone character standing in front of ancient runes, covered with moss and vines, a full body shot of the third person view character holding a sword or shield facing the camera with dead trees and falling leaves on the ground. In the style of the unreal engine, a dark fantasy setting, don't get overwhelmed by this long text. We can break it down as follows, a screenshot from the game Dark Souls. In the style of the unreal engine, a dark fantasy setting, these sentences specify the style of the image. Then a bond character standing holding a sword or shield. This is our object in front of ancient runes covered with moss and vines, dead trees, and falling leaves on the ground. This is our background and finally, a full body shot of the third person view character facing the camera. This part helps guide mid journey towards a certain composition by typing that we want a full body shot. Third person view will instantly make Mid journey, give us straight composition from a third person game. This detailed text prompt, despite being complicated, actually has it all to give you an image that you described very well as a text without the need to use any other image inputs. I actually didn't need to use hAGIBT or spend time thinking of this prompt. I simply give mid journey an image and use the slash describe Command. This is in my opinion, the most underrated feature in mid journey. As you can see, it gave me four full description of the image and how mid journey breaks it down into text so that I have a rich description of the same image. This is our result. It perfectly aligns with what we type. But what if we want more control, more guidance on the image composition, for instance, I don't want my character to be standing on the right or left. What if I want a similar composition to this image, then I'm going to use the exact text prompt. But I will add this image as an image prompt to guide mid journey towards a similar composition, but pay attention to what I said earlier. It's true that an image prompt greatly helps with composition by as high as 70%. Unfortunately, though the image prompt will also affect the style and colors of the final image. What I mean by that is, let's say your comic book is in the style of a web ton or cold manga and you like the image composition that you want to use as image input to get a similar composition. But this image you're going to use is in a different style or color than your comic book. That means it will affect your style by 25%. But we may get workable images, but you will have to use a lower image weight than three because a realistic image input set to image weight three will most of the time create an image in a realistic style. So in conclusion, it is very important to use an image prompt that is very similar to the style of your comic book. Now look at our final image with using image prompt. We significantly guide the grid journey to create a very similar composition to our image link. And I will show you now both videos of the image being created with and without image input to see how the image prompt affects the initial noise. In the next lesson, we will dive into the second type of image input, style reference, we will explore how it shapes the final image and then move on to the most crucial input for our comic book. Character reference, the fun part is just getting started, so get ready for some exciting stuff. See you there. 14. Style reference: All right, let's talk style references. You can use images as style references in your prompt to influence the style or aesthetics of images you want Mid journey to make. To add style references to a prompt, use the S ref parameter with the image link URL after you text prompt. Style references only apply to prompts that include text. They cannot be used with image only prompts. You can use multiple style references and just like image prompt, weight, IW, that ranges 0-3. You can also adjust the style weight by adding SW and the number 0-1 thousand. This W refers to style weight. You can also uses ref random to apply a random style to your image. Did you know that actually some people try style random countless times? Because when you mess around with the style random feature, you get this awesome style reference code. Why is this cool? Because with that cool, you can slap the same style onto any image later on. No need to keep describing the style or using down reference images. It's like having a magic button for style consistency. While style references are cool, they aren't the MVP when it comes to keeping your candles comic panels consistent. And here is why when you need to nail a tricky composition that words alone can describe, image prompts are your best friend. Imagine you are trying to to set up a scene from an enemy or manga with a specific composition. Image prompts will guide the AI to get it just right. If you want your comic to have a consistent look, make sure your image prompts match or close to the style you are aiming for. And second reason is use text prompts to back up your style. Now, don't worry about style references just yet. You can use text prompts to reinforce the style you want. Describe the style, the mood, and the details you are going for in your prompts. This will help the AI match your comics vibe without the potential clash that style references might bring. This lead us to reason number three, avoiding style reference drama. Sometimes style references can create a bit of chaos, contradicting with your image prompts or text descriptions. To avoid that, focus on crafting the perfect image prompt and text prompts first. Sure. Style references are awesome, but they are not always necessary. Use them wisely and only when you need to tweak things. I mean, look at this image. You can see that we get a similar type shot. Nearly the girl is standing in the same bows as the girl in the reference image. And this red flower turned to a butterfly at the same place with the same color. So clearly, style reference also affects the composition of the image, but less than image input. Style reference is the least important input for making our consistent comic book, but still is a very powerful tool in our arsenal that we can use if things not work quite well without it, and it's great to know it's there and even try it in combination with other inputs. But I just want to tell you that most of the time, you might not need style reference. What a ride, everyone. We have reached the peak of our course. I'm super excited because now the real adventure begins. We have covered all essential mid journey commands, parameters, and three different inputs. Next, we will explore the final image input, character reference. Then after that, we will be officially ready and well equipped with the knowledge to start to learn the ultimate formula. With various examples and tackle complex scenarios to create consistent professional comic panels and turn your ideas into stunning polished works of art. See you in the next one. 15. Character design introduction: Mid Journey had released a game changing update, introducing the character reference feature. This tool is perfect for storytellers looking to maintain consistent character appearances across different scenes in visual comic narratives. So if you are as excited as I am, let's jump right in. We have hit a critical milestone in AI comic creation by introducing character references. We are getting closer to our ultimate formula guide for creating compelling characters. First, let's talk about character design. Understanding your character's look and personality is crucial before letting mid journey work its magic. Ask yourself, how does my character look? Do I have any inspirations or ideas and don't be discouraged if your character resembles an existing one? Inspiration from other works is a common practice and enriches your creations. For example, did you know that Mariam from Hunter X Hunter was inspired by Freezer from Dragon Ball? The key is to ensure your character reflects your story and style, even if inspired by others. Infuse your personal touch to create something unique. There are several ways to design your character in mid journey. One, using a text pro. This is the simplest method. To blending two character images, create a new design by combining elements from different characters. Third, advanced design with Photoshop for more control and customization. Let's start with the first method. Using a text prompt. You have two ways to generate these prompts. The first way is by using Chad GPT. The second way is by using slash Describe. In the next lesson, we will start generating our text with the first way using Chad GPT, see you there. 16. Charater design Method 1 with Chatgpt: Told HGPT to give me various character design in the style of Bleach the Molisl my hero Academia, and Naruto. I told him to make sure that each character must be creative and not a replicate. Then I got these results. Now I told them that I want each character design to use a mid journey as text prompt and ensure that you include the character appearance, and that the characters are in full body and on white background. This help us isolate the characters a white background. In this tip, you can tell me Johnny any note about your character personality or even if it is inspired by existing characters. But now as I only want to see random character designs that mid journey will give us, let's copy number four, Sura Noto inspired. Let's see what this prompt to give us. Let's paste it in Mid journey creation tab. Actually, I will generate the same prompt twice, first with the latest mid journey version 6.1, then with Anime version Niji six. You can go wrong with either because you have included the style of the image you want very well. You will also get an anime look images. Even with Version 6.1. While the images are being created, let's go back to Chat Gibt and let's also try this character, the demonus layer inspired character. Let's copy it and paste it here and let's also create it with two different versions as well. Let's see our examples. This is the Narrato inspired text with Version 6.1. I really like the character, yet it's similar to Narrato somehow because this character is blonde as well. Mid journey have predicted what a blonde character from Narrato Anime would look like, and he is Narrato himself. To get a different character, you might want to change the hair color or other appearance descriptions that doesn't exactly match and existing character narrato. But I will stick with this character for now. Still, there is a problem with this image. As you can see, the character is not isolated for creating our character sheet and multiple poses. It is preferred to have a character isolated, a white background. Even that we said standing a white background, we still got these wolf spirits behind them. I can see why if we look back at the prompt, this is actually what Mid Journey gave us with Wolf fang hanging around his neck and the character is accompanied by ghostly spirit walls with swirling wind effects. I think this is very cool image. Mid Journey not only gave us the character, but also his ability. This is a perfect character for a shining type enemy. If you want this character and it just as is, you can create a character sheet for it by isolating the background using Photoshop. There are actually some free AI tools online that can remove background and isolate the subject. Here I'm actually going to try the same prompt, but without the extra ghostly wolf, my instinct tells me that the result character, despite it will more likely be isolated a white background. I don't think we would get the same cool shown in character design like this one. I forgot something for creating a full character image using portrait Asper ratio, aka R nine by six will give us better full body character. Wow, check out our result image. This is Naruto inspired character again. Generated from text only, we have used the same text prompt without the wolf's description and wind. You can see the character now is perfectly isolated, even though I forgot to remove everything that triggers the wolf background. If you see something you want to fix, go to repaint. If the arrows in the image are small enough and you're afraid to consume too much GPU in mid journey, this can be fixed easily in Photoshop, and I will show you how to do that in Chapter three. But this is way easier than drawing a character design from scratch. You can can see how AI already give us a boost in creating several character designs even with shown inabilities in few seconds. Now we are inside repaint window again. The text prompt doesn't have to be perfect. I will just remove most of the long text prompt. As I want mid journey to focus more on repaint, I will only leave full body character design of soda in Osuka in the style of this should work fine. We can even remove in the style of Naruto because mid journey will analyze the whole image and will do great job at regenerating the selected area. So your text prompt here doesn't need to be specific, especially in our case because we don't want to add a specific object. For example, I have selected his hand place. Mid journey knows that in this place, we should see this character's hand. But what if our character is holding knife? In this case, you must very clearly write that in your repaint prompt. But in my case, I will leave it like this. I imagine you should fix the hand and legs. Now this is so great. The hands and legs are fixed, despite maybe an extra leg finger that can be used. Fixed in Photoshop, I think we can download this image actually to use it for later. This is what I got with the same text prop, but in Niji mode example, I really like it. But you can see we didn't get the full body shot as we asked for. One of the reason is that we didn't change the aspect ratio. Sometimes we get a full body like in the previous job with version 6.1 with the default square aspect ratio. Now we can regenerate the same prompt and change the aspect ratio. Or if we like our character in this example and we want to get the full body, we can try reframe tool and then adjust the aspect ratio. Let's actually do that. Let's adjust the aspect ratio to nine by 16. Then put the image at the dalmo start because we want to generate the bottom part of the character and hit submit. Right away, we get stunning reframe with a full character shot with more of our character wolf spirits. I really like this image and particularly this one I like the most. It looks like the blue flames were coming out of the wolves and had tear up some of his. This is really creative and impressive mean we can fix this burn if we want, but I will leave it. I think this is an amazing character design concept art that we got just by using Cha gibt and Mid journey. As always, you can isolate the character from the background in Photoshop, or we have second way to vary our image by using the remix variation. First, if you don't see the remix variation here, then go to more options and activate remix option. Then we can remix subtle or strong and to remove the background, remix subtle is not enough. Because you want to give Mid journey more space to change the picture. Remix means that mid journey will work from this image and we tweak it. Remember our first ever example about the god Orden with the ePatch. We used variations strong in remix mode. The difference between vary options and remix options despite the work exactly the same. In remix mode, we get to adjust the text prompt to tell me journey to vary our image. But following our new text prompt, right here in this example, I want to vary the image strong to move the background, including the wolves. So in order to do that, we need to change the text word that triggered the wolves. That means options. In our case doesn't work. We have to choose remix strong so that we can adjust our text SP to remove all the words that triggered the background. You can see remix mode is activated in creation tab. I'm going to remove with wolfangs hanging around his neck. The character is accompanied by ghostly spirit wolves with winds around him. I will also remove wind driven elements around them and start. Now all the words that triggers the background is removed. Now I will start the remix job and now here's our result. We get a similar character design, but without the wolves, yes, there is still some wind, but this is just traces. Because of our original remix image, we can easily remove them. The only thing bothering me is that character's proportion is not perfect in my opinion. His leg seems very short and there is one of the examples where we didn't even get the full body. This is because I forgot again to change the aspec ratio. Please always remember if you want to create a full body shot character design in mid journey, it is advised to use a space nine by 16 to get that portrait aspect ratio so that you avoid mistakes and refine jobs later. And here is our demons layer inspired character. Again, I can believe how amazing this first generation of our character is. You can see some minor defects like her missing hand. But again, the subtle description that she's holding a Katana imbued with ice really give us sick showing character design right away and to fix her missing hand, let's try repaint tool here. If you are in this code, repaint is the same as very region. Now we can select the areas we want mid journey to fix or regenerate. In the first try, I'm going to leave the prompt just as is and to see the other result in the same repaint window. In the second try, I'm just going to remove all the text prompt and type her hand is moving water. Leaving the text prompt in repaint window is also a great way to fix things and you don't need to retype the full text prompt that generated the first image. Why? Because mid journey already can see the full picture and knows the style and everything. By only selecting a part of the image, you can type what you want to fix and you don't have to rewrite the full problem. Now the job is submitted and I will use same repaint text prompt, but with Version 96 instead of 6.1. To show you, it really doesn't matter at this point. Repaint. Prompt can be very simple and straightforward to only describe the part that you want to fix, whether it is in anime style or even a realistic image. Here are our repaint results. We can see the hand got fixed in both prompt versions. Now we have our Demons layer inspired full body character design concept art that we can use to create a character sheet. We will do that in the ultimate formula. Before we move to the next method in designing our character in mid journey, let's see what's happening here. I like this character, but he's showing up as a boy. What if I want him to look more mature or muscular? Or if our character ages in our story to adjust, we can use the reframe tool and make the image taller. A nine by 16 ratio doesn't seem enough, so I will try a one by two or maybe even a one by three. No, that's too much. I think one by two ratio feels just right, and I'm moving him up to create more space in the lower area, and I'll click Submit. This is just the first step. We have only reframed the image to make it taller, creating more space for mid journey to work with. Now I will do a repaint and select the lower area if you. You think about it, we gave mid journey three inputs to fix our body proportions. First, our remake sticks prompt, and second, the part of the image as information to help mid journey fill in the selected areas we want to fix. And third, the extra space in the lower area that we have selected. This encourages mid journey to expand the character's lower part. Our refined character will have longer legs, an extended lower part and a more mature appearance. I think this looks stunning compared the before and after the body adjustments, all we did was reframe to increase the space below the character. Then in another step, we used repaint, giving me Johnny new area to fill. Here are some more examples I have created with H the same method using Tubiti and mid journey. I will use some of these images in our next method. Finally, in this example, this girl doesn't look like a high school girl as her body is a bit childish. Let me demonstrate very quick on how fixed it. And that's it. We have designed our character's first concept art in mid journey, letting Kat creatively describe our character and give us the text prompt. This method is great either if you don't have a specific design in mind or for more control, you can also provide HBT with more details about your character. If certain characters or elements inspired your creation, sharing those will help generate a better text prompt and more fitting to your story. 17. Charater design Method 2 and 3: Second way is using the slash describe command in this code. Provide MD journey with an image of an existing character you like. Type slash describe and upload an image from your device or use a link. Let's use this image. Let's copy the image link and paste it here. You will see you will have four different text prompt descriptions for our image. MID Journey will give you four text prompts to create your character. There is a similar tool on Mid Journey's website, but I prefer the disc code version. Here are some examples of prompts I generated with slash Describe and the resulting characters. Moving to the second method of character design inside Mid journey, which is pending images. This doesn't require text prompts. You can blend two character images to create something new. Like mere whim, the difference here is that Mid journey acts as our artist using these images to inspire an entirely new character. Check out this example. I blended an elephant image with a monster to create a unique creature that transforms at night. Here is our result. This method lets your imagination worldwide, offering countless possibilities for experimentation. But if Aba you are serious about character design and want something truly original, there is a more advanced method using Photoshop. It requires some refinement and drawing, which I use for my own character design. Since I wanted more control over my character's design, like the body, face, and expressions, I'm going to show you the advanced character design method that I took for the next method. I also showed you how to create a new character and how I customize my character using Mid journey, chat GBT, and Photoshop. First, I needed to get the body type and clothing down first. Here is what I did. I started with the body design. Step one, I typed in the creation tab inside Mid journey. Imagine full body turnaround of an arrogant, angry high school man in the style of my academia, and I'm using the NIGISix version. You can see that I didn't specify that clothing as I wanted to give mid journey some freedom and see what he would came up with. This is the image I got a perfect full body shot. But unfortunately, it wasn't a full body turnaround. Instead, I got some random cut out upper body poses, but no worries. To get the full side and back views, we move to step two. Here is where I use the pan write feature with remix mode activated. Why pan right and lot left and why in remix mode? Well, when extending image journey image, the a boot follows patterns. For example, if I pan lift, I would likely get more of those random upper body shots. But by panning right where my character's full front body is, I am more likely to get a full body shot in the extra space. N Y remix mode, it's because I wanted to add view, side view to the prompt to guide mid journey further on what was missing and should be added. Then I got more full body images for my same character from different angles and check out the result. Did you notice something? The side and back views are nearly perfect and matching my front body shot? In terms of body dimensions and even the clothes, isn't this mind blowing? This is where planning and extending an image journey really excels for creating AI comic assets. We can instantly get a consistent character from different angles and poses without even specifying what our character is wearing. This is because mid journey follows the pattern in the image. He checks the text prompt and thinks, we have a front view of this character and there we mix prompt mentions, side and back views. There is a new area I should fill in. Then I probably should fill the side and back views of the same character. This is where some of the magic in mid journey happens and we will show you more examples of these techniques later. Now that I have got my characters front side and back view images, it's time to create the face. These were my trials until I found a face that really matched my vision as a starting boy. Well, first I typed character sheet, multiple facial expressions of Kasuki Bakugo and the style of Naruto on a white background. I tried mixing a character from my hero Academia with the style of another enemy like Naruto, hopefully to get a new character that still had Bakugos exhibitions. But the result was too similar to Bakugo, so I tried another way. On my second try, instead of naming Katsuki Bakugo, I described an arrogant young man. This description best fits my character when you first see him. I like this image, but the quality and clarity of the character's face weren't great, so I tried one more time. This is where things began to click. I use Cha JBT to create a text prompt by asking it for a full description of my enemy character. I told him my character's personality is similar to Katsuki Bakugo. But with that list, my character's hair is spiky like fire, mimicking, his fiery temper, this is the prompt I got, and this is what meat Journey gave me. Notice that I enabled dash dash style row in this example. Why? Because the more detailed and specific your text is, the better the results when you activate style row. It follows your text more closely as we discussed in an earlier chapter. This is where I am at now. I have got my characters body images and a decent starting point for the faces. The next step you guessed it is Photoshop. With Photoshop, I combine this body image with this one to create this and here is the same steps inside view. And then finally, I refine my character face further in mid journey to get my final result. This is a very close match to the character I envisioned when I started creating it, since I am only giving you a summary of my journey in designing my first character. In the next lesson, we will finally dive deeper into the ultimate formula and we will apply everything we have learned to create consistent AI comic assets for our story. Including the type of images you will need and how many reference images are required. Also, how to achieve different poses, expression and camera angles in Mid journey. We will also tackle any problems that pop up along the way. By the end of the ultimate formula chapter, we will be ready to assemble all the assets we have created inside Mid journey and start storyboarding our first professional AI comic page. See you in the next one. 18. The Dump Assets: The last lesson, we created these two images for our character. As you can see, these are the minimum character reference images you need before starting to generate countless images for your character. We have a front fold body shot on a white background and a side fold body shot on a white background. Now let me answer some questions you might have. One, why do we need full body shot images for character reference and two? Why do we need at least front and side angles of our character? And finally, three, why should our character reference images be on a white background? Let's start by answering the first question with an example of how not to use character reference images. Imagine using only a face shot as our character reference. Here is an image link, and I can either copy the link or drag it into a mid journey creation tab. Now let's say we want our character to strike a particular pose. First, we need a text prom that describes this pose. The best method without a doubt is to use slash Describe to minimize guesswork. You want mid journey to tell you how it interprets this image, including the pose, so you can create a solid text prom that accurately captures it. After using dash, describe and copying the link, we get four different text prompts that describe the image. A great tip is to look for words that are repeated in more than one prompt. These are often key to achieving a similar image. For instance, full body shot is mentioned multiple times. Let's copy that. We might also copy any email since it's relevant. We also see side view repeated in three of the four prompts, so we will use that as well. In fact, the fourth text prompt provides a clearer description, side view walking post. I will use that. Now about the background in the first prompt, it says no background, but simply typing no background in the prompt is not perfect because mid journey interprets every word individually, often leading to an unintended background. And instead, you can specify on a white background or blank background as the fourth prompt suggests. You can also use the negative prompt, dash, dash, no space background. Let's try this. With these adjustments, we now have a nearly perfect prompt for achieving our pose. Looking at our third text prompt, number three is the most detailed one. It gives us cool additional descriptions about our characters, clothing, hairstyle, shoes, and even the image style like line art, et cetera. However, since I want mid journey more precisely to achieve this pose, I will use this image as an image prompt. Also, don't forget to sit your character image as character reference, and we will increase our image prompt weight to three. This will guide the initial noise towards a very similar pose. By using this image as our image prompt, you will notice it's in the same style as our character image, both our enemy and simple line art. We don't need to add style references or type additional text prompts about the style. We are using our text prom to get a full body side view Walking post shot guided by this image prompt with the highest weight. But here's the catch. We are using a character reference image that isn't a full body shot. Let's see what this prompt will do. Please don't laugh. Well, that's it for today. I hope you enjoyed my course and see you in the next time. Just kidding. This bad example actually demonstrates how the character reference image can affect our output image composition as well. As you can clearly see our distorted character appearance, even though we use this image with the highest weight to guide the composition. But using only a face shot as a reference has ruined our character's body. Now, let's use the same exact prompt. But this time, we will use our full body side viewshot as the character reference and remove the old character reference image and set our new image as character reference. Here are our results, much better dimensions, and we get this exact pose. But with our character, quick tip. In Ijimud I got even better character body accuracy. However, for more complex poses, you should know that sometimes a character image reference can lock in your character's pose, even if you guide it with both a text prompt and an image prompt. If you are having trouble achieving the desired pose, say you want your character running, you might want to remove the image prompt because our image prompt character is working, not running. In this case, you could rely solely on your character reference image to depict your character running. If you want to see your character running in a side view, then you should use the character reference with the full body side view. I actually made a mistake in the first two examples. Fortunately, I got a decent result, but let's fix it now. Our character reference image and image prompt were both portrait images not square. It would be better to use AR nine by 16 to help with the character body accuracy. Let's do that. Let's see this example. Oops, as you can see, the result isn't too great. Why? Because our character reference interfered with the text prompt in terms of composition. We specified running in the text prompt, but our character reference image was standing. So you get a mix of both to fix this, first use the character weight parameter CW because when you don't use CW, mid journey sits the character reference weight to the default, which is 100. A 100 character weight means mid journey will take both your character's face and clothing as references, giving you very similar clothing to your character as well. But the downside is that this also locks in the pose. By using the lowest character weight, let's say CW space zero. Mid journey will give us a consistent phase, but not the clothing. This will free our character's pose. To free our character's pose, we must lower the character weight parameter 0-70. But what about our character's clothing? We want consistency. Remember, because if we use very low character weight, we won't get the same clothes. To fix this, let's go back to our disc code. Hit slash describe and copy the image of your character's full body. Why? Because we want to use mid journeys description of our character's clothing in the text prompt to maintain consistency. Here is Mid Journey describing our image. What we care about is the character's clothing. We will only pick the words that describes our character's clothing, wearing a black jacket and pants, comma, white shoes. Let's copy and paste that into our text prompt and don't forget to include your text prompt before any parameters and press Enter. This should be much better. If you don't want the characters hand in their pocket, you can describe the hand pose in your text prompt. Or if you want different results for running poses, you can always rerun the job and use style row for different weird values. But I don't recommend using chaos values because chaos will produce very random results and deviate from our composition and text prompt. By just adding style row while sitting the character weight to zero and describing our character's clothing in the text prompt, we get much better running poses. You can always fix small imperfections using either the repaint tool or Photoshop and you can even extend the image if needed. This is an example I created earlier for my own comic. The text prompt was a high school boy is walking in a side view in the style of Japanese animation. It's a full body shot with no background. I also added with an angry expression on his furious face. My text prompt here was more refined and polished. Even though I use the same exact image prompt and character reference, I got a better result with a more accurate facial expression. Adding subtle descriptions of your character's mood in your text will guide the output more towards the desired image. Now, what if we want our character in a different angle? As you can see here, I got my character angry in a front view. I use these two image references. Using two full body image references from different angles will help create more variation and dynamic poses for your character. By the way, all the images we have created without an image prompt are called dump assets. Why? Because we didn't use an image prompt to direct M journey more specifically into a pose. Instead, we let my journey generate a random image based on our text prompt and character reference. So let me show you how I created more dump assets that will enhance your AI comic creation workflow and innocently give you more options when designing your story. Keep in mind that if you plan to change your character's clothing in later scenes, you should create dump assets for your character in those other outfits. Now, let's create more dump assets. I will generate multiple random facial expressions for my character. We will need one or two very good quality face images for reference and an image prompt like this. I use KatsukiPakuku's character, face sheet. You don't need KatsukiPakuku specifically. We choose this image to only trigger a certain initial pose. Don't worry. You won't get KatsukiPakuku at all. This just lets me journey. No, you want several faces with different expressions by directing the initial noise as we learned before. Now I will show you several examples and many dump assets I have created with this method. Again, you can even create more varying images by using reframe tool as this. You can either extend the image by changing the aspect ratio to let Mjoury create more faces, or we can edit the text prompt by removing the character reference images. And even the image prompt to give Mi Johnny more freedom to get more variations, don't worry, you will still get your same character because M Jony has enough information about your character face and style from the rest of the image. In this way, you can create countless assets for your character's face from different angles that you can instantly access and blend into any body shot image to fill in your comic panels and move your story forward. You can also create more dump assets, full body shots for your character. In this example, I have chosen three full body images as character reference. You can play around with text prompts, character weight, and style to get various full body poses for your character. In this example, let me actually try to set the character weight back to 70 for more poses, freedom while still retaining some of the clothes. I will tie back in wearing white shirt, black jacket, pants, comma white shoes. And this is the result. There are poses that are flawless and we can crop it just as is with Photoshop. But if there are flaws in any image, we can always repaint, and if the face is incorrect, for example, we can use any face from our dump assets and blend it with any body pose from dump assets. Really, you have countless possibilities and I will actually do this right now. Thank you. As you can see, we have learned how to create our dump assets for full body and how to blend them to create countless possibilities. Now, what if we want to create more directed image for our story? What if your character is doing a pose in your mind that you want to visualize with mid journey. In the next lesson, we will learn how to take image prompting to the next level to get our desired pose. See you there. 19. The Dump Assets 2: Let's create more assays of a new character following the ultimate formula. Step one, creating a front fold body character design on a white background. I had choose to design this character with only a text prompt that I got from Chad GBT. Let's call him Nishida and step two, upscale the image and save it to our character folder. Step three, extend the image to left or right, press on repaint and change the aspect ratio. Then edit the prompt to full body turnout, side view, then submit. Here, I got a very good backview of Nishida. However, as I wanted to continue creating side view of him, something weird happened. During recording my next lesson, mid June had released this stunning update. The new image editor say goodbye to reframe, repaint Very region and Zoom because these all are combined into a single unified interface the new image editor. In my opinion, this will make refining our image way more seamless and is indeed a huge step forward. Let's try the new feature by creating a side and backshot of Nishida. Click on Editor and this window will open. You can change the aspect ratio of the canvas by pressing on any of these default. Or you can drag this little line to increase your canvas manually. What amazes me is that we can blaze our image anywhere inside the canvas. Previously in reframe interface, if you remember, we only could have put our image at the start, center, or end. Now we can place our image anywhere inside the canvas. And here are the repaint brush to select the areas you want to regenerate. As we only want side view and Baview, let's just click on the default aspect ratio 16 by nine and place our image at the start. Nothing crazy here. We'll also add to the prompt back view comma side view, and I will change the full body shot to full body turn now and submit and boom. I think I got a perfect result first image. This image has the best side view shot I ever got of Nishida. Also, I noticed we get different hand gestures and both are nearly flawless. Could this be because of the newest update? Anyway, I will upscale the image and download it. Let's move on to the next step, step four, refine our character using Photoshop. But before refining our character, there is an extra step to make before it. Let's call it step 3.5 because as you see, Nishida is wearing a mask for some reason. So to refine our character reference images, we need to see Nishida's face, right? Unless you are planning to hide his face for the entire layer of your story. Like Kakashi, what if Nishida would reveal his face somewhere in the plot? So let's be prepared for that. First, I created this image with image prompt earlier and I really like the hair. The face, however, is not clear. So I blended this guy's face with this hair to create my customized character face. Let me show you how I blended them in Photoshop. U Now we have a decent Nishida phase. Let's put our new face. On our character, full body images. With our new front nicht face, I was able to easily generate more faces of him by using it in mid journey as character reference. And I got this amazing side view, face shot. So now we have front and side face images that we can use as character reference. Now step four, let's refine our full body character images thirst before we blend our new faces on the full body shots. Uh U Step five is to create dumb assets for our character. Generate multiple random poses, expressions and angle shots for your character. Here are the dumb assets I created for Nishida literally within 2 minutes. You can pause at any image to see the text prompt parameters and the character or style reference is used. We can at least pick ten images already and edit them inside Mid journey or Photoshop and accessible assets for our comic book. That's it for Nishida. Next, we will create characters with background scenes and we will learn how to compose our characters on any background. See you in the next one. 20. The Pro assets Creation Level 1: Let's create pro assets for AI comic panels. For our cause, this lesson is the epics of our journey that's cherry on top. The moment you have been waiting for, name it as you wish because now we are going to create AI comic panels at three different levels. Level one, easy. At this level, the comic panel will future only the character without a background scene. It may include a fixed background, such as color gradients, speed lines or action lines. Level two medium. Here, the comic panel will include both the character and a random background scene. This can be achieved either by using character and style references or by combining all type of image references. The pros of this method is minimal photoshop work compared to Level three, and the character blends seamlessly with the background. However, the cause is less control over the final image and the results can be inconsistent. Level three hard, despite the name, level three isn't particularly difficult. It just involves an additional step. In this level, you will create the character and the background separately inside mid journey, then blend them in Photoshop. The pros of this method is greater control over composition, style consistency, and character continuity. And it's more likely to achieve a perfect result on the first try in mid journey, reducing the need for further extensive editing. In fact, level three might actually be easier than Level two, as it often delivers better results with less reliance on editing. I will guide you on which level to choose based on the type of comic panel you are planning to create. How do we get our character in a desired pose or create a desired background? These are your best friends for creating pro assets One Photoshop. No real skill needed. Even rough edits will do, I will explain later to slash describe, we will use this command most of the time. Be an inspiration image for character poses or backgrounds and for character reference images for body and face, as we have learned to create before. That's it. With these four friends, you will likely be able to achieve any comic panel you want for your story. Let's start with level one. The first example, character only comic panel. This is Yusuke. What if I want him to do a hand pose in meeting fire? It's simple. Step one, we will do a low key composition for our panel. It doesn't have to be perfect. In fact, it can be very rough. Let's see, we have this image of Yusuke standing. What if I want him to emit fire? Here's a silly photoshop composition where I have put around them hand here and assemble to defire even though this image has poor photoshop work, it has enough information to use as an image prompt. The purpose of this method is to give me journey an image prompt that will trigger the initial noise of our exact comic panel we imagine. This leads us to step two. Let's copy our composed image link and open me journey in this code. Use slash Describe and paste the image link. We get four nearly perfect text prompts. Let's pick number two. Step three, paste our text prompt, and then drag our image into the creation tab and I will activate it as image and style reference. To show you how strong and stunning this method is, insulating almost perfect results, I won't even use a character reference, but I will increase the image weight to two for more direction. You can increase image weight to the maximum value of three and I'm using NijiSik and see the result. Unbelievable. We give Mid journey a low key edited image from Photoshop and he created a perfect comic panel that match our vision. This is why I create my AI bro comic panels in two steps. First, for only the character, then second, for the background. But why? Because giving Mid journey our low key composition image for our character on a white background makes mid journey focus only on the specific pose, expression, angle, and type of shot. This allows us to get the best description of our character in the comic panel and also help mid journey generate the perfect character image. We already got perfect results without using character reference, but I'm also going to show you the result if we use our composed image as image, style, and character reference. Another result by adding style raw. As you see, you will most likely get perfect results. This is why making a low key composed image can literally achieve any character pose or scene you imagine. Let's move to our second example. Look at this. I really like the character pose, angle and type of shot. But what if we want use K doing the same thing? Again, we will follow the same example. Let's follow our ultimate formula again. One low key Photoshop composition work. We need to replace this guy's face with use K space. But when I searched in my dump assets for use K space, I didn't find a good low angle face image of him. So I simply went to one of my previous dump assets I created for use K and I just added to the ticks prompt, low angle and from blow. These two help direct the pose angle. You can even increase the ticks to eight for low angle to direct mid journey further, creating multiple phases for our character in a low angle and check this out. We got a perfect result. Back to low key editing. Let's select our phase with the object selection tool, free transform and roughly place our new pace Remember the two D fire image. Let's select our file, right click and choose Layer via Copy. We move it to our main image and I use the eraser tool to remove this unwanted white fell. Now with the fire layer selected, I will make it smaller and roughly place it here in place of this ball. Now, all we have to do is isolate the body from the background using the object and quick selection tools. Once we finished, right click and layer via copy. Now we have the body face and fire in separate layers. We can remove the background layer and we will end up with a pretty solid composed image that is ready to use in mid journey. Step two, use Describe to describe the image we just made. Then we copy one of the prompts. Step three, paste the ticks prompt and drag our image. Again, I will only activate the image as image and style reference without character reference. But this time, I will increase image weight up to max value at three and I'm still using Nijisix then submit. And I will generate more variations with the same prompt. First variation is only by adding style role. Second time by adding a character reference image, make sure to use a full body image as character reference because the composition image we want to achieve is not a close up. Suppose your desired composition is upper shot only. In that case, we can use an upper shot as character reference. But this time, we must use a full body image, and let's reduce the character weight to 50 to give more freedom to the pose. However, we can use the default character weight if you want the same hair and clothing and submit. Check out the result. This is flawless. We get the same pose but better. Magarini basically refined and fixed our low key composed image. This gives us countless possibilities and ideas on how to execute our comic panel. And here is our result with character reference used. You can see that our character hair is not perfect. That's because we use lower character weight. In this case, we can always replace the face and hair with any image from our dump assets for more professional, consistent looking characters throughout our story. Moving to the third example, we can create a comic panel with more than one character using dump assets and blend more than one character using Photoshop. But what if we want more specific comic panel that has more than one character where each character is in a very specific post perspective and expression. In that case, we must use pro assets, and here is a great example to showcase it. Check this out. What if I want to achieve this composition? But with my characters, let's go. First thing is we can take this image as a whole and generate a text prompt by slash describe or use it as image prompt. Because when there is more than one character in one image, especially when they are in very different perspective or expression, mid journey will be very confused and this will be clear if you try to generate text prompt using described command or when you use that image as image prompt, the solution, very simple is to split each character in a separate image, and now copy the first image link and generate text prompts with describe command. Now mid journey will focus on generating a perfect text prompt for each character. Now, let's create the first image in Mid journey. Here is our character. I use front and back images as character reference. You should add our character back view image because it's similar to the composition we want. Here is the image we are going to use as image prompt and I will use style row because our composition is very detailed. I need mid journey to pay attention to the text prompt for better following my character, facial expression and pose, I'm using Image way three, and this is what I got. Stunning, I got the exact image I wanted. I will actually dilute two images because I will do some simple refine work to get the best character pose and facial expressions from these two images into one single image, as you see. Now moving to second image. Here is the prompt I got. I use this image as image prompt with the highest weight, and this is my character image as reference. I only used her Bview image because as you see the image prompt, we want. We don't see that guy's face at all, only his backside. And the fact that he's running, this might give us some clue. We only want a back view image as a character reference. And it should be enough. Again, I'm using a text prompt that I got from Slash Describe, but I changed it slightly to describe my new character. I have generated also text prompt for this girl so that I know how mid journey both interpret this girl look and clothes. She is a woman wearing a white button up shirt and a navy blue skirt. I don't have to write what she's wearing because I'm using the default character weight 100, which also keep the clothing consistent. But I did that anyway. If you wanted to free the character pose and decided to lower character weight to something like 70 or 50 or maybe zero, then writing our character clothing in text prompt is a must. Look at the result image. Perfect pose, but her shoes is missing. Let's easily fix that. Inside mid journey editor. Inside the editor, sometimes I like to remove any image reference, especially character reference because mid journey has enough information about your image style and character. And we'll generate the selected areas. However, if your character has a very specific type of clothing or hair that you want to regenerate, then you might want to keep your character reference image inside the remix text prompt, and this is it. We have this image and this image. It's time for some Photoshop. I think the boy is missing some of his head. Let's fix that quickly in mid journey. Now, let's open that image inside Clip Studio. If you don't know Clip Studio, it's an art studio program dedicated for animation, manga, and web tools. But in our course, we are just going to use it for creating speed lines, focus lines, pitch, and thinking bubbles. Then import those back into Photoshop. To create focus lines, go over to this little icon called figure or use the shortcut and you will see these different types of focus and speed lines. I'm going to use focus lines. Let's try scattered focus lines. It lets you create a circle around the focus object. In this case, it's the girl. We will draw circle around her. Once you created the focus line, if you want to adjust the focus line, you must first select the focus line layer. Then try different values of our line properties. Now moving to the second focus line, which is dark focus line. Instead of drawing a circle, this let you draw curved lines around your target. I think I never needed to use more than these two, but you can always try other focus lines. Meanwhile, I think we can get stunning results. With these two, the next important icon for making AI comic pages is this balloon icon with short cut T. Now, what are the types of balloon in comic panel? One narration balloon. We can use these rectangular balloons or slightly around it balloon with no tail for narration or commentary from the author. Two speech balloons, which is used for characters spoken dialogue, and it's typically around with a tail pointing to the speaker. In manga, the tail can vary in style depending on the mood or emphasis. If your character is yelling or screaming, then this is the balloon you are looking for. You will find more of these balloons in Clip Studio library where other members share their work. Three thought balloons for characters and tonal thoughts. It used to be cloud like or a series of small bubbles leading to the character's head, but with more modern manga. Cloud balloons are similar to speech balloons in shape, but without tail. This new approach simplifies the visual presentation and can help integrate thoughts more seamlessly into the dialogue flow. But suppose your protagonist is facing a dangerous situation where he must think very quickly as every second counts. In that case, you want to use spike balloons for f internal thoughts. You also can find this in Clip Studio library as well for there is also sound effect balloons. But let's go back to our comic panel. This kid is looking at the girl running. He wants to stop her. We can use speech balloon, either rounded ellipse or curved balloon for his speech after you create the balloon. It's time for the tail. Click on balloon tail and while the balloon layer is selected, draw the tail pointing to the kid mouth. There are three types of balloon tiles. You can experiment with them and see which is better for each comment. But actually, I prefer the boy yelling. Maybe the girl is far away and he will speak very loudly so she can hear him. In this case, I will use this imported spike balloon and that's it. Now after we import the focus lines and balloons, you can import them in separate layers for flexibility inside Photoshop. Now we finish refining the panel, and that's it. We have created a Level one comic panel using P assets and we have learned how to use Clip Studio in creating focus lines and balloons. In the next lesson, we will start creating Level two comic panel. See you there. 21. The Pro assets Creation Level 2: Moving to level two comic panels, let's create comic panels for our character with a background in one mid journey prompt. See this man walking on a sidewalk in city will somehow cloudy dusty mood in the background. What if I want my character in this place? Here you can see my character reference. So there we go. See this image link and generate a text prompt with describe command. Look at number three, a handsome eneme style character with green hair and black hoodie is standing on the street. His left hand is holding the sleeve of his jacket in front of him while he looks at the camera. We can ignore all that because it describes the character that we want to replace. Let's continue. The background features building behind the fence in an anime inspired style with a side view angle and cinematic steel shot. Let's copy all that and the aspect ratio as well. So mid journey understood our image very well, especially the background. Is building behind the fence, but I'm going to refine it even more. My text prompt will be a close up side view of a boy standing near our mesh fence with city buildings behind him. Now we can add any more detail about our character, expression, mood or pose. We can ignore our character appearance because we are going to use character reference images. I will add the scene captures a moment where he appears to be deep in thought. Here is the result. Our character face is correct, but not the clothing because we lowered character weight and the image prompt had caused some unwanted clothing and color from the old character. To fix that, we can use the editor and remove the image prompt and face only character reference, but only keep his full body side view and set the character weight back at 100 works every time. But if you really want to prevent the clothing problem from the start, then you should use the default character weight like this prompt. But if you want your character in more dynamic pose expression, then you might want to lower character weight and fix the clothing problem with the editor, whichever fits your goal. Now, as you see, I can even extend the image further with the editor. I remove the image prompt and let me join creatively expand our background. Do whatever you like until you get the composition you like. Now that we have put our new character in a desired background with one prompt only, what if we want a background scene with only a text without image prompt? This is a great example, but I only use that image as style reference because that's the mood vibe and style I want my character to be in. I want him taking a photo of someone. He said in the style of Makutushin kai and the result is phenomenal. This works great if your character is in a random location in your story. We can use this method by relying on a text prompt and style reference to generate our character on a background scene. Look at this example. Creating comic panels in Level two works great if your background is only a ground shot like this. This is a high angle shot of the boy and we only see the ground behind him. This also works for a sky background or any simple textual background scene. I will leave here more examples with all the images I used so that you see more ideas of creating an image with our character and a background in Makutushin. Here is a twist. What if we don't want a similar background? Rather, we want a similar character expression. You can use an image prompt for achieving an extreme character expression as well. If text alone was not enough. See, I wanted my character in deep shock like this man does, and as always, copy the image link and let me journey describe it for you. I will use number one and refine the text prompt, and I will use these two images for my character. And now I have my character in that extreme expression. Now let's move to level three comic panels where we will create the character and background separately for more control on our background, then blend them inside Photoshop. 22. The Pro assets Creation Level 3: Suppose our story chapter takes place in a setting like this, a misty first with towering trees next to our wooden house. The atmosphere is mysterious. To generate a scene like this, let's copy the image link and create our text prompts. Then copy any of the prompts to our creation tab. For this example, I will use the image not only as an image prompt, but also as a style and character reference. You might be wondering why would a background image also serve as a character reference? I discovered that using a character reference can also help with the composition. Take a look at these great examples created by this guy named Alan. He got the same building structure using this image as only a character reference. Using a character reference clearly helps with the composition. This is why if you remember when we choose a character image as character reference, we get a similar composition and we need to lower the character weight to free our characters composition, allowing them to match the desired pose guided by either a text prompt or an image prompt. Now back to our creation tab, I will use image weight three and the white screen aspect ratio and then submit. See how we got the same mood and composition as our initial image. If you want to create more scene images that will occur near this one, you might want to select the image as only a style reference. However, in that case, make sure your text describes the scene very well. Next, let's open Photoshop. Here is the image I have chosen as my background. Now it's time to add our hero, Nishida, remember him. I already have a great full body shot of him at an angle that fits this background scene. If you don't have one, you can generate your character in a different angle that fits the background scene you have created as we showed in the previous lessons, but for now, I will just do a simple retouch to blend the character well into the scene. Take note, I'm not a pro at blending in Photoshop. If you really want to master this skill, you can look up tutorials on how to blend images and create composites in Photoshop. But let me quickly show you what I did. First, I did some cleaning work. Now we want to add his shadow. With the background scene selected, we can select the shadow area and lower its brightness as shown. A better method is to duplicate the background scene, lower the brightness of the full layer, then create an invert mask. While the mask is selected, brush the areas where we want the shadow to appear. This method is more flexible if you change your character's position and allows you to quickly fix the shadow as well. Now I will do the same for Nishida. I will duplicate the character layer, lower the brightness, and use a mask. This is a low key method to blend our character with the background scene. To spice things up, let's imagine Nishida is fighting a monster. Here is our creepy monster seen from behind. At a low angle, I forgot to separate our monster layer from the background. To fix that, we can do this in Photoshop or faster. We can use Adobe's free AI background removal. Now let's add our monster back into the scene. This is already an epic scene. You can see how flexible Level three comic panel are. We have more control over our characters placement, angle, pose, and type of shot, something that's much harder to achieve with Level two panels. Now let's move to our final lesson. Let's create a full AI Comic page inside Photoshop using the assets that we learned how to create. See you there. 23. Create A professional looking Comic Page: Let's create Comic page in Photoshop. Four dimensions, a common Japanese manga page is around 7.42 by 10.37 " and US standard comic page is slightly different. It's 6.875 times 10.5 ". These dimensions include the bleeding area. So what is the bleeding in comic pages? The bleeding area is when you make your page a little bigger than the paper, so that when you cut it in printing, the color still goes all the way to the edge with no white showing. So how much bleed? Well, the standard bleeding thickness is about 0.125 times 10.25 times 10.25 times 10.25 times 10.25 times 1010. Let me show you real quick. It's actually simpler than it sounds. In our example, I will create a page with US standard dimension. So to do that, create a new page. Now, make sure the color mood selected is CMYK and the resolution is 300 and let's make the dimension 6.875 by 10.5. Now we need to know our bleeding area to avoid adding any drawings or words in it because it's not a live area. Click view the new guide layout. You don't need to worry about columns or rows. We can uncheck those. We are only interested in the margin here. Typically, we want to add the bleeding thickness in all the edges to 0.125. Click this. But don't forget in comic book, there are left and right pages and there is in between place called bending Area Margin or Gutter Margin. The last thing you want is your comic page panels going to that bending area. We need to increase the margin in the bending area. The right page will have bending margin in left side and this page will have the bending margin on the right side, and I'm creating a right page. I will add the bending area margin in left side. And this margin, also known as the Gutter margin is usually 0.5 to 0.75 ". I will choose 0.75 as a standard bending area margin. Now you can see my page with the margin that is outside the live area. Now we create a new layer and with the guide showing, you will see Photoshop automatically snap to the selected location for you. As you see, I'm selecting all the outer margins, including the gutter margin and will create a solid color. I will change the color now to something like bluish or greenish. This is just to visualize your trimmed area. Eventually, after done creating the beach, we will fill the margins back to white, place another layer and fill it with any color. Now it's very simple. Select the brush. Make sure your brush is white and are set to the highest hardness and simply brush on any point, click Shift and then brush on another point and Photoshop will automatically create a straight line. Now with the panel lines layer selected, choose magic one tool to select our first panel area. Then we can simply create a new layer. Let's call it panel one and fill that layer with any color. Do the same for other panels. Now we have fill layers for all our panels that we can use as clipping mask. Let's put our margin layer at the top and fill it back to white. Let's say you want to make the first panel, simply paste the panel image and put it over the desired panel layer we just created. In our case, it's panel one layer. Hold Alt or Option key and click in this little gap between the two layers. In this way, you create a clipping mask, meaning your image will only appear in the fill area of panel one. But with all that said, I actually prefer creating my comic pages. In another way, let me show you what I mean. Let's delete all that for now and only keep the margin layer and fill it with back. Now here you can see in my folder, these are the assets created in my journey that I will use to make my comic panels in this page that I will use to create my comic page. Let's grab Image one. Sometimes I do this trick, create a mask and fade my picture edge. Or we can just extend our picture inside mid journey. Then I assemble all other images. In this way, I can visualize where the story boarding or the panel composition and lines will be. Now, I create the panel lines. Let's create a new layer. Again, let's make the straight lines with white brush. Also, make sure to select hard round brush to avoid round bubbly edges of your line. Let's adjust the brush size. Click here, hold Shift and click here, and we instantly get straight lines. However, in this page in a specific, there is a vertical angle line. So let's fix that. Usually, you won't need to do that if you are just creating straight panel lines. Now finally, the remaining is the speech and balloons. Again, before going to Clip Studio and create the speech balloons. First, I create the text in the desired areas that I want to put my balloons. Make sure the text is in the least interesting part of your image because the balloons were covered up. I forgot to paste the boy close shot. So let's put him in this comic panel. And once I finish align all my text, I exported this image to Clip Studio. To know where I will create the balloons, as I showed you earlier, then I only export the balloons in a separate layer. And here I pasted the balloon layer. But you can see I can easily resize each balloon to put it in the right place. And see here, actually, this balloon edge is not showing because of our outer margin, so we know that we shouldn't put the balloon this far. Let's move it slightly to the right and do the same for the text. Uh, Now we are finally there. In less than 20 minutes, we created a professional looking comic panel in Photoshop. With the help of mid journey, as we reach the conclusion of our course, I want to say you have unlocked a secret formula, one that allows you to blend AI with your vision and creativity in ways that were once unimaginable, opening doors to endless storytelling possibilities. Remember, if the image you create, every character you bring to life is a piece of your unique vision. Keep experimenting, keep pushing boundaries, and most importantly, keep telling your story. Thank you for joining me on this adventure. Now it's your turn to create something extraordinary, happy creating, and as always, see you in the next one.