Getting help and adjusting settings - Midjourney Tutorial

- [Instructor] In this video, we're going to get familiar with how we can ask Midjourney for help. Not to mention, discover the base settings that we're working with. Just to recap, in the last few movies I've been using a direct messaging system with the Midjourney Bot 'cause I'm on a paid subscription. This also allows us to see the images that are being produced without the hassle of a channel. So this is how we can ask Midjourney for help. All we have to do, like we've done with Imagine, is type in forward /help followed by the return key, and that's going to produce a bunch of different resources of where we can go for that help. One of my favorite areas is the Quick Start Guide. If we click on this, it's going to lead us to a webpage of Midjourney documentation, where we'll click on the Quick Start Guide. Here, you'll discover a recap of things that we've already done in this training, such as the imagine command, its power, not to mention the options that we have after it's created its job, including upscaling that image. Not to mention creating variations. One of my favorite areas of this Quick Start Guide happens to be next steps. Let's first of all go to Prompts, and I'm going to scroll a little bit down where you're going to find a really interesting memo under Prompting Notes in the Grammar section, which is how you should prompt Midjourney based on the model that you're using. So version 4 and version 5 are much better at natural language than earlier versions. I would say version 5 takes the cake. With earlier versions, you might want to slightly change your sentence structure to get a better end result. So something to keep in mind. And the other thing to note is exploring prompting, which are different styles in Midjourney, by simply asking it. For instance, if I type in the imagine prompt, followed by folk art as the style in the style of a cat, it might produce something similar to this. Take a note at some of these styles listed here. We'll try one of these out here in a second. And how we could even get specific about different types of line styles, travel at different generations, and creating emotion. Definitely something to look at in terms of exploring prompting. Let's head back to Discord and now take a look at our settings. I'll type in that forward slash followed by typing settings. We'll hit return, and this is what I'm currently using. At the time of this recording, the base Midjourney version is 4. Although I did use version 5 in the last two exercises. It produces a base quality image, but if I wanted something higher quality because of my subscription, I could pay a little bit more in terms of time in order to get those files. Style is something important to keep in mind, which is essentially how tightly Midjourney is going to follow your text prompt. A Low style is going to follow your text prompt, literally, versus a very High style might even ignore most of your text prompts in certain situations. You'll notice I'm also in Stealth mode, which keeps my prompts private, as well as I've enabled Remix mode, which I'm going to get to in a second. This, in combination with Fast mode, which generates the images more quickly at the expense of more time spent on the server, and potentially a higher subscription price. I'm just going to change two more of these settings. I want to set the style at the very high while still keeping this at Midjourney version 4. You'll see that my settings are going to end at the end of my prompt, this stylization setting of 750. Okay, let's see this in action. I'm going to type in imagine. I'm going to type in "A blacklight painting of sad elephants "from the early 1900s." I'm not going to add anything to the end of this prompt, but I'm just going to press the return key. Keep in mind it's using Midjourney version 4, but it has added the suffix style 750, because I changed my settings to a very high style. So let's see how much it follows my text prompt. Okay, so this is what it's created. So if you recall the blacklight image that was shown to us, this looks slightly different. They didn't follow it as closely. I'm not sure if these elephants are sad either. I'm going to go back up to my settings and make some changes. I'm going to set the style to low. Can see that it's now set to 50. I'll continue using Midjourney version 4. And again, we're going to get to Remix mode in a second. But let's copy this part of the prompt, and I'll come down here to the message, and add that /imagine. Command V or control V to paste, followed by the return key, where it appends the s50 to the end of it. Let's see what Midjourney comes up with. You can see here with the images that it generated, this is mimicking a blacklight painting style much more than the higher style value that was given in the previous prompt. In these images right now I like the third version down here, so I'm going to create a new variation of that. And because I've enabled Remix, I can actually add something to this before it creates its new variations. For instance, rather than sad elephants, I'm going to choose sad cats, and actually say from the 1800s instead. We'll keep the aspect ratio and the style at the same. So let's press Submit, and now Remix has rewritten that prompt for us, while keeping the majority of it, and is going to produce an image based on that remix. So you might be able to see that Remix could probably be a huge time saver, rather than you having to copy and paste or rewrite a prompt altogether. We can simply enable Remix in our settings, then vary the image that we like, but with slight changes to the actual text prompt. So there's how we can ask for help, not to mention change our settings, here inside of Midjourney.