Midjourney v6 Tips and Tricks for Beginners and Advanced Prompters. Part 1

Thaeyne
26 Dec 202308:22

TLDRThis video offers insights into using Midjourney V6, a new AI model for image generation. It covers basic settings, aspect ratio control, and the menu options for image manipulation in Discord. The tutorial explains the upscaling process, variations, and the remix feature, highlighting current limitations like the absence of pan and zoom. Tips for avoiding unwanted close-ups and the impact of style on image detail are also shared, promising more advanced features in the next video.

Takeaways

  • 😀 Midjourney V6 is different from earlier models and requires new prompting techniques.
  • 🔧 Check your settings with /settings to see which model is active or specify V6 in your prompts.
  • 🔄 Use the D-AR (aspect ratio) argument to control the shape of your image; the default is 1:1.
  • 🔍 Aspect ratios have been improved in V6, allowing for better control without previous limitations.
  • 📊 Generated images in Discord come with nine buttons for upscaling, rerolling, and variations.
  • 🚀 Upscaling options include subtle and creative, with no 4X upscale yet for V6.
  • 💡 The remix feature allows for prompt changes during rerolling and variations if enabled.
  • 📏 Controlling image details can be challenging; use negative prompting (e.g., --no closeup) to adjust.
  • 🖼️ Missing features like pan and zoom make it harder to control image framing in V6.
  • 🧩 Specific details in prompts, like shoe type, can influence the depiction of entire persons.

Q & A

  • What is the main topic of the video transcript?

    -The main topic of the video transcript is discussing the features and usage tips of Midjourney V6, an AI technology for image generation.

  • How can you check which model is active in Midjourney or specify V6 for a prompt?

    -You can check the active model by using the command '/ Slash settings' or you can specify V6 by adding '--V6' or '--v6.0' at the end of your prompt.

  • What does the 'D-AR' argument stand for in Midjourney V6, and what does it control?

    -The 'D-AR' argument stands for 'Dynamic Aspect Ratio'. It controls the shape of the generated image, allowing you to make the image wider or taller.

  • What is the default aspect ratio for images generated in Midjourney V6, and what are its pixel dimensions?

    -The default aspect ratio is 1:1, which produces an image that is 1,024 by 1,024 pixels.

  • What happens when you change the aspect ratio of an image in Midjourney V6?

    -Changing the aspect ratio will elongate the image in one direction while reducing it in the other, but it does not increase the total number of pixels.

  • What are the U1, U2, U3, and U4 buttons for in the Midjourney V6 Discord menu?

    -The U1, U2, U3, and U4 buttons are for upscaling an image from the grid, but they do not actually increase the image size; they provide the selected image separately with additional buttons.

  • What does the refresh button do when used with a generated image in Midjourney V6?

    -The refresh button rerolls the same prompt, generating the image again. However, if the seed is set in the prompt, it will produce the exact same image, which may waste generation hours.

  • What are the V1, V2, V3, and V4 buttons used for in Midjourney V6?

    -The V1, V2, V3, and V4 buttons are used for creating variations of an image you like from the grid, providing images that look similar to the original.

  • What are the differences between the 'subtle' and 'creative' upscaling options in Midjourney V6?

    -The 'subtle' option tries to keep the upscaled image more like the original, while the 'creative' option may change some details in the image.

  • Why might specifying the kind of shoes a person is wearing in the prompt affect the style of the generated image?

    -Specifying details like shoes can influence the style of the generated image, possibly due to the way the AI interprets the prompt, although the exact reason may vary and could be related to the seed used for that particular generation.

  • What features from version 5.2 of Midjourney are currently missing in V6?

    -Features such as pan, zoom, and VAR region options from version 5.2 are currently missing in V6, but Midjourney has promised to add these features gradually.

  • What is a suggested method to avoid close-ups in generated images using Midjourney V6?

    -Using negative prompting by specifying '--no closeup' in the prompt can help to avoid close-ups, as it instructs the AI to zoom out in the image.

Outlines

00:00

🤖 Introduction to Mid Journey V6 AI Prompting

The video script introduces the viewer to the new features and changes in Mid Journey V6, an AI technology. It emphasizes the importance of checking the active model settings and the option to specify V6 in prompts. The script explains the use of the aspect ratio (D-AR) argument to control image dimensions, with a default of 1:1, and how altering this affects image size and shape. It also covers the menu options available under the generated image in Discord, including upscaling (U1-U4), rerolling prompts, and variations (V1-V4). The script highlights the ability to edit prompts with the remix feature and the differences in image handling between V5.2 and V6, noting the absence of certain features like pan and zoom in the new version.

05:01

🔍 Exploring Advanced Features and Challenges in Mid Journey V6

This paragraph delves into the advanced features and current limitations of Mid Journey V6. It discusses the absence of pan, zoom, and VAR region options, which are crucial for controlling image composition. The script shares personal experiences and strategies for overcoming these limitations, such as using negative prompting to avoid close-ups and adjusting the aspect ratio to include more of a subject in the frame. It also mentions the impact of specifying details like shoes on the style of generated images and the inconsistent results when using photography-style language in prompts. The video promises to explore more advanced features in an upcoming video and encourages viewers to share their own tips and findings in the comments.

Mindmap

Keywords

💡Midjourney V6

Midjourney V6 refers to the sixth iteration of the AI model developed by the company Midjourney, which is designed to generate images from textual prompts. In the video, the host discusses the differences between this version and earlier ones, highlighting new features and changes in the way users should prompt the AI for desired outcomes. The script mentions checking settings with 'Slash settings' to ensure V6 is active, indicating the importance of this version in the context of the video.

💡Aspect Ratio (D-AR)

The aspect ratio, denoted as D-AR in the script, is a critical parameter that determines the shape and dimensions of the generated image. It controls whether the image is wider or taller, with a default ratio of 1:1 producing a square image. The script explains how changing the aspect ratio affects the image size, with wider or taller images resulting in a reduction in pixel count in one dimension. This concept is central to understanding how to control the visual output of the AI in Midjourney V6.

💡Upscaling

Upscaling in the context of the video refers to the process of increasing the resolution of an image. The script mentions U1, U2, U3, and U4 buttons, which are used to upscale images from the grid, although it notes that currently, there is no upscaling happening with these buttons in V6. The concept is important as it relates to the quality and size of the final image output, and the script discusses the limitations and expectations regarding upscaling in the new version of Midjourney.

💡Rerolling

Rerolling is the action of generating a new image using the exact same prompt, essentially creating a variant of the original image. The script describes a refresh button that is used for rerolling, but cautions that it may not be useful if the seed is set in the prompt, as it would result in the same image, thus wasting generation hours. This term is significant as it relates to the experimentation and variation aspect of using AI image generation tools.

💡Variations (V1, V2, V3, V4)

Variations in the script refer to the process of generating images that are similar to, but not identical to, a selected image from the grid. The V1 to V4 buttons are used to create these variations, with the script noting that they allow for more variation if the 'strong' button is used. This feature is important for exploring different interpretations of a prompt and finding the most appealing visual outcome.

💡Remix Feature

The remix feature allows users to completely change their prompt, offering flexibility in the creative process. The script mentions that this feature can be enabled in settings and, when active, presents a popup with the prompt text for editing. This feature is significant as it provides users with the ability to iterate and experiment with different textual descriptions to achieve desired results in image generation.

💡Negative Prompting

Negative prompting is a technique used in AI image generation where the user specifies what they do not want to see in the generated image. In the script, the host shares an example of using 'dash dash no closeup' to avoid close-up shots, illustrating how this technique can influence the AI's output. This concept is important for users looking to refine and control the details of their generated images.

💡Photography Language

Photography language in the script refers to the descriptive terms used to communicate specific visual effects or compositions, such as 'closeup' or 'full shot'. The host notes that Midjourney V6 seems less responsive to some of this language compared to previous versions, indicating a change in how users need to communicate their intentions to the AI. Understanding this concept is crucial for effectively prompting the AI to generate images with desired compositions.

💡Seed

In the context of AI image generation, the seed is a value that helps determine the randomness in the generation process, allowing for the reproduction of the same image if needed. The script mentions that rerolling with an explicitly set seed will result in the same image, which is important for users who want to maintain consistency in their image outputs.

💡Pan and Zoom

Pan and zoom are features that allow users to control the field of view in an image, either by panning across a scene or zooming in for a closer view. The script notes the absence of these features in Midjourney V6, which the host finds challenging as it affects the ability to control the composition of generated images. This concept is significant for users who require precise control over the visual focus of their AI-generated images.

💡Style

Style in the script refers to the artistic or visual approach applied to the generated images, such as 'photography style' or 'painted style'. The host observes that adding certain details to prompts can influence the style of the resulting images, as seen when specifying shoes led to images in a painted style. Understanding the impact of style on image generation is important for users aiming to achieve a particular aesthetic.

Highlights

Midjourney V6 is distinct from earlier models and offers new ways to prompt for AI-generated images.

Check settings with /settings to see which model is active or add --v6 to specify the model in your prompt.

The aspect ratio (D-AR) argument allows control over the shape of the image, with a default ratio of 1:1.

Changing the aspect ratio does not increase the image size; it only elongates one direction while reducing the other.

A 16x9 aspect ratio results in an image of 1456x816 pixels, maintaining the pixel count.

Upscaling options (U1-U4) are available under the generated image in Discord, but do not change the image size.

The refresh button rerolls the same prompt, which may consume generation hours without significant changes.

V1-V4 buttons offer variations of a liked image from the grid, with options for subtle or strong variations.

The remix feature allows for complete prompt changes when enabled in settings.

Upscaled images in V6 result in a 248x248 pixel image, with options for subtle or creative changes.

The heart button sets the highest rating for an image on Mid Journey's website for better organization.

The web button links to the image on Mid Journey's website, a method for downloading or viewing images.

V6 currently lacks pan, zoom, and VAR region options, making it harder to control certain aspects of image generation.

Negative prompting, such as specifying '--no closeup', can help avoid unwanted close-ups in images.

Adjusting the aspect ratio can provide more room for body representation in images.

Specifying details like the type of shoes can influence the style and composition of the generated image.

The speaker will explore more advanced features of V6 in a follow-up video and encourages sharing of tips.

The video concludes with an invitation for viewers to subscribe for more content on Mid Journey and V6.