How to generate the most REALISTIC images possible with Midjourney v6

WesGPT
26 Dec 202315:03

TLDRMidjourney v6, the latest version of the AI image generator, offers significant improvements over its predecessor. It provides more accurate prompt following, allowing for longer and more detailed prompts to influence the generated images. The new version also introduces a remix mode for transforming photorealistic images into illustrations, and a minor text drawing ability. Users can now include text within images by placing it in quotations. Midjourney v6 also enhances upscaling, which enlarges images while maintaining high resolution. The platform operates through Discord and offers various subscription plans, starting from $10 per month for basic to $30 for unlimited generations. To use the generator, users connect to their Discord profile, subscribe to a plan, and then generate images using simple, straightforward prompts. The video demonstrates the process of generating images, including using variations and remix mode, and highlights the impressive photorealism of the generated images, especially human faces with their imperfections.

Takeaways

  • πŸŽ‰ Midjourney has released a new model, version 6, which offers more accurate prompt following and longer prompts.
  • πŸ“ˆ Improved coherence and model knowledge are part of the enhancements, along with better image prompting and a new remix mode.
  • πŸ–ŒοΈ Version 6 introduces minor text drawing ability, making it easier to include text in generated images.
  • πŸ” Upscaling has been improved, allowing for higher resolution images when increasing the size of the output.
  • πŸš€ New features like pan, zoom, tune, and describe will be rolled out later, enhancing the capabilities of the model.
  • ⚑ Prompting with V6 is significantly different from V5, requiring a relearning of how to create prompts for better results.
  • πŸ“ Short, simple, and direct prompts are preferred in version 6 due to the model's improved understanding of user intent.
  • πŸ’‘ Built-in features like '--style Raw' can be used for photographic images, and adjusting 'stylize' values can affect prompt understanding and aesthetics.
  • πŸ”— Midjourney works through Discord, which is a chat app and a platform for creating bots and APIs.
  • πŸ’» Users need a subscription to generate images, with different plans offering varying amounts of generations and fast generation hours.
  • 🏠 The video demonstrates how to set up and use the Midjourney bot on Discord, including changing the model version and generating images with prompts.
  • πŸ“œ The script also covers how to upscale and create variations of images, as well as the potential for text inclusion in images.

Q & A

  • What is the latest version of Midjourney's model discussed in the video?

    -The latest version discussed is Midjourney version 6.

  • What improvements does Midjourney version 6 offer in terms of prompt following?

    -Version 6 provides more accurate prompt following and allows for longer prompts, giving a better chance for words later in the prompt to appear in the generated image.

  • How does the remix mode in Midjourney version 6 work?

    -The remix mode allows users to take a photorealistic image and remix it into an illustration by adding specific instructions at the start of the prompt.

  • What new text drawing ability does Midjourney version 6 introduce?

    -Midjourney version 6 introduces minor text drawing ability, allowing users to include text in their outputs by placing the text within quotations.

  • How has upscaling been improved in Midjourney version 6?

    -Upscaling in version 6 is improved by taking a small image and making it larger while maintaining a higher resolution.

  • What are the built-in features suggested for use with Midjourney version 6?

    -The built-in features suggested for use include adjusting the style with '--style' followed by a stylized value (e.g., '--style raw' for a photographic image) and using the '--coherence' flag for better prompt understanding.

  • How does the subscription process work for Midjourney?

    -To subscribe, users go to Midjourney.com, sign in through Discord, and choose a subscription plan. After subscribing, they gain access to the Midjourney Discord server and the image generator.

  • What are the different subscription plans offered by Midjourney?

    -Midjourney offers a basic plan at $10 per month for about 200 generations, a standard plan at $30 per month for unlimited generations, and other plans that provide more hours of fast generations.

  • How can users generate images using the Midjourney bot?

    -Users can generate images by using the command '/imagine' followed by their prompt in the Discord server where the Midjourney bot is added.

  • What does the 'U' button do when generating images with Midjourney?

    -The 'U' button (U1, U2, U3, U4) is used to upscale a selected image (image 1, 2, 3, or 4), allowing users to save a higher resolution version of the generated image.

  • How can users provide feedback or request additional features for Midjourney?

    -Users can provide feedback or request features by commenting on the video or engaging with the Midjourney community on Discord.

  • What is the aspect ratio feature in Midjourney version 6 used for?

    -The aspect ratio feature allows users to specify the shape of the generated image, such as square, landscape, or vertical, by using a command like 'ar16x9' for a 16:9 aspect ratio.

Outlines

00:00

πŸš€ Introduction to Mid Journey Version 6

The video introduces the latest model from Mid Journey, version 6, and outlines the new features it offers compared to the previous model, version 5.2. These enhancements include more accurate prompt following, longer prompts, improved coherence, and model knowledge. The video also discusses the new remix mode, which allows for variations in image style, and the minor text drawing ability. The host explains that the new model is more sensitive to the prompts and suggests using short, simple prompts. The setup process for using Mid Journey through Discord is also covered, including authorizing the connection to the user's Discord profile and browsing community-generated images.

05:01

πŸ“ˆ Subscribing and Generating Images with Mid Journey

This paragraph explains how to subscribe to a plan on Mid Journey and start generating images. It details the different subscription plans available, from the basic plan offering around 200 generations per month to the unlimited plan. The host also discusses the speed of image generation and the option to upscale images for higher resolution. The process of adding the Mid Journey bot to a personal Discord server for private image generation is described, along with how to use the bot's commands to change the model version to version 6 and generate images using specific prompts. The importance of prompt length in version 6 is highlighted, and the host demonstrates using the bot with a sample prompt.

10:02

🎨 Exploring Mid Journey's Advanced Features

The host delves into Mid Journey's advanced features, including the ability to generate variations of an image, the remix mode for transforming photorealistic images into illustrations, and the minor text drawing capabilities. The video shows how to upscale and save individual images, create variations for fine-tuning, and use the remix mode to achieve different styles. The host also tests the text drawing feature with different prompts and aspect ratios, noting the improvements in text accuracy. The video concludes with a call to action for viewers to share their interests in learning more about specific features and how they plan to use the new version of Mid Journey.

Mindmap

Keywords

Midjourney v6

Midjourney v6 refers to the latest model version 6 of the AI image generator, Midjourney. It is significant because it introduces new features and improvements over the previous version 5.2. In the video, it is the central focus as the presenter discusses its capabilities and how to utilize it for generating realistic images.

Prompt following

Prompt following is the ability of the AI to accurately interpret and generate images based on the textual description provided by the user, known as the 'prompt'. Midjourney v6 has improved prompt following, allowing for longer and more complex prompts to be effectively translated into images, which is a core aspect discussed in the video.

Remix mode

Remix mode is a feature in Midjourney v6 that allows users to take an existing image and transform it into a different style, such as turning a photorealistic image into an illustration. It is showcased in the video as an innovative way to create variations of images, demonstrating the versatility of the AI.

Text drawing ability

Text drawing ability refers to the AI's capacity to include text within the generated images. In Midjourney v6, this feature has been enhanced to allow for 'minor text' to be incorporated more effectively. The video demonstrates how to use this feature by placing text within quotation marks in the prompt.

Upscaling

Upscaling is the process of enlarging a smaller image while maintaining or enhancing its resolution. Midjourney v6 has improved upscaling capabilities, allowing for higher quality larger images to be produced from smaller ones. This feature is highlighted in the context of improving image quality in the video.

Discord

Discord is a chat application that is used as the platform for interacting with the Midjourney AI. It is mentioned in the video as the medium through which users can subscribe to plans, generate images, and engage with the community. The integration of Midjourney with Discord facilitates a social and interactive user experience.

Subscription plans

Subscription plans are the various pricing options offered by Midjourney for accessing and using their AI image generation services. The video outlines different plans, such as the basic and standard plans, which offer different numbers of image generations and access speeds to the AI's capabilities.

Aspect ratio

Aspect ratio determines the proportional relationship between the width and height of an image. Midjourney v6 allows users to specify the aspect ratio of their generated images, enabling the creation of square, landscape, or vertical images. The video demonstrates how to use this feature to control the shape of the output images.

Variations

Variations refer to the different versions of an image that can be generated by the AI based on a single prompt. In the video, the presenter uses the 'V' buttons to generate alternate versions of an image, offering the user more options and flexibility in the final output.

Photorealistic

Photorealistic describes images that closely resemble real-life photographs in terms of detail and quality. The video emphasizes the enhanced photorealistic capabilities of Midjourney v6, noting the impressive level of detail and realism in the generated human faces and scenes.

Coherence and model knowledge

Coherence and model knowledge pertain to the AI's ability to understand and generate images that are not only visually consistent but also contextually relevant to the prompt. The video discusses how Midjourney v6 has improved in this area, leading to more meaningful and contextually aware image generation.

Highlights

Midjourney has released its latest model, version 6, offering more accurate prompt following and longer prompts.

With version 6, the influence of words in the prompt has been adjusted, with earlier words having more impact on the generated image.

Improved coherence and model knowledge are part of the enhancements in version 6.

Remix mode allows for variations in image generation, such as transforming a photorealistic image into an illustration.

Minor text drawing ability has been introduced in version 6, making it easier to include text in outputs.

Upscaling features have been improved to maintain higher resolution when enlarging images.

Features like pan, zoom, tune, and describe will be released later this month.

Prompting with V6 is significantly different from version 5, requiring a relearning of how to prompt for better results.

Short, simple, and straight-to-the-point prompts are recommended for version 6.

Built-in features like '--style Raw' can be used for photographic images.

The new model is in its alpha stage and available for testing to paid subscribers.

Midjourney works through Discord, a chat app that also supports APIs and bots.

Subscription plans range from basic to unlimited generations, with varying costs and features.

Users can generate images in public rooms for community interaction or in private servers for more personal use.

The '/settings' command allows users to switch to Midjourney version 6 within the Discord server.

The '/imagine' command is used to generate images with a given prompt.

Upscaling and variations can be done on generated images for better results.

The text in images should be enclosed in quotation marks for better accuracy.

Remix mode can transform photorealistic images into different styles, such as 2D illustrations.

The video demonstrates the creation of beachfront mansion images with sailboats using the new model.

The human faces generated by Midjourney version 6 show impressive realism and imperfections.