Midjourney 6 — The Biggest Update You Should Try!

AI Master
26 Dec 202317:43

TLDRMidjourney V6 introduces significant updates to its AI image generation capabilities. The new model, available since December 21st, enhances text generation and image accuracy, with improved prompt following and support for longer prompts. Users can access V6 by specifying 'D-V6' in their prompts or changing the model settings. The update also includes creative upscaling options and a new text generation feature, though it still requires refinement. The community has discovered increased prompt length and ability to specify colors and placements. Despite some limitations, the update showcases potential for future improvements and comparisons with other AI image generators.

Takeaways

  • 🚀 Midjourney V6 is released with significant improvements over previous versions.
  • 🔍 The updated model can be accessed by adding 'D-V6' to the prompt or changing the model in settings.
  • ✍️ V6 introduces the capability to generate text, with specific formatting requirements for prompts.
  • 🖼️ The image generation shows better accuracy, but still has room for improvement in details and style consistency.
  • 📈 The aspect ratio can be specified using 'AR' followed by the desired ratio, like 'AR 16x9'.
  • 📝 Prompt following is more accurate and V6 can handle longer prompts of over 350 words.
  • 🎨 New arguments like 'chaos' and 'weird' have been introduced, though their effectiveness varies.
  • 🔧 The 'stylized' argument is number-based and can range from 0 to 1000, potentially improving aesthetics over prompt understanding.
  • 🎭 V6 has the ability to analyze and transform existing images into different artistic styles.
  • 📚 An official guide on writing prompts has been released by the developers to optimize results.
  • 🔜 V6 is currently in its alpha stage, with regular updates expected to further enhance its capabilities.

Q & A

  • What is the significance of Midjourney V6 update?

    -Midjourney V6 is considered the biggest update so far, with improvements in every part of the system and some previously known tricks no longer working.

  • When did the Alpha version of Midjourney V6 become available?

    -The Alpha version of Midjourney V6 became available on December 21st to all users.

  • How to correctly access the updated Midjourney V6 model?

    -There are two ways to access the updated model: by adding 'D-V6' to your prompt with a space between the 'D' and 'V6', or by changing the model to V6 Alpha in the settings menu.

  • What are the notable improvements in Midjourney V6?

    -The notable improvements include the ability to generate text, more accurate prompt following, accepting longer prompts, and being sensitive to the user's input.

  • How does the text generation feature in Midjourney V6 work?

    -The text generation feature works by using a specific prompt format, where all text should be in quotes and 'AR 16x9' is used to specify the aspect ratio of the generated images.

  • What is the recommended way to upscale images in Midjourney V6?

    -The recommended way to upscale images in Midjourney V6 is to use the 'creative upscaling' option, which provides up to two times higher resolution results.

  • What are the new capabilities of Midjourney V6 regarding prompt length and specificity?

    -The new capabilities include a prompt length of over 350 words, the ability to specify colors and other details, place things where desired, prompt for more than one subject, and interact with MJ like a chatbot.

  • How does Midjourney V6 handle the 'chaos' and 'weird' arguments?

    -The 'chaos' argument randomizes the output based on the number provided, while the 'weird' argument, when set to 100, results in a more chaotic and strange output, but not necessarily weird in the expected sense.

  • What is the role of the 'stylized' argument in Midjourney V6?

    -The 'stylized' argument is number-based and can range from 0 to 1,000. Lower values may have better prompt understanding, while higher values may have better aesthetics.

  • How can Midjourney V6 analyze and transform images into different styles?

    -Midjourney V6 can analyze an image and transform it into different styles like oil paintings or da Vinci sketch styles, although the results may not be direct copies and could vary in similarity to the original image.

  • What is the developers' recommended structure for a 'perfect prompt' in Midjourney V6?

    -The developers recommend starting with the style of the image, defining the main focus, establishing the environment or context, providing composition details, setting the lighting and mood, describing additional details, and finally, specifying secondary objects or interactions relative to the main subject.

Outlines

00:00

🚀 Mid Journey V6 Update Overview

The video script introduces the latest update to Mid Journey, version 6, which is considered a significant upgrade. The update has improved every aspect of the software, rendering some old tricks obsolete. The Alpha version was released on December 21st and, while the list of improvements may seem short, the changes are substantial. The most notable addition is the ability to generate text, which the presenter intends to test extensively. The updated model can be accessed by adding 'D-V6' to the prompt, with a specific requirement to include a space between 'D-' and 'V6'. The video demonstrates the generation of images of a happy puppy with a snack, highlighting that while the images are not perfect, they are accurate enough. The presenter also discusses the new feature of more accurate prompt following and acceptance of longer prompts, testing this with a complex prompt about a medieval marketplace. The video also touches on the sensitivity of the new version to prompts and the avoidance of nonsensical terms like 'award-winning photo realistic 4K 8K', emphasizing that previous versions were not accurate or convenient, and setting expectations for the new version.

05:01

🎨 Testing New Features and Prompts in Mid Journey V6

The script continues with the presenter exploring new features and longer prompts in Mid Journey V6. The community has discovered that the new prompt length has been extended to over 350 words, allowing for more detailed and specific requests. The presenter tests this by asking Mid Journey to generate an image of three best friends of different ethnicities and in specific clothing. Out of the four images generated, only one fully complies with the requirements. The presenter then generates variations of the most accurate image, resulting in four visually pleasing images, though not quite the quality of a real photo. The video also discusses the platform 'poo', a sponsor of the video, which is a tool for painless video production, allowing users to create high-quality videos easily without the need for actors or studios. The presenter then proceeds to test various arguments supported by Mid Journey V6, such as aspect ratio, chaos, and weirdness, noting that some parameters do not fully meet the expectations set by the initial prompt.

10:02

🖌️ Exploring Stylized and Artistic Options in Mid Journey V6

The video script delves into the exploration of stylized and artistic options in Mid Journey V6. The presenter discusses the stylized argument, which is number-based and can range from 0 to 1,000, with developers noting that lower values may improve prompt understanding while higher values enhance aesthetics. The presenter's testing reveals that without specifying a particular style, the generated images are either realistic or well-drawn. The script also covers the 'style raw' argument, which results in more photographic and literal images. The presenter highlights the ability to transform a found photo into different artistic styles, such as oil painting or da Vinci sketch style, although the results are not direct copies and vary in similarity to the original. The video also touches on the capability to analyze images, which is a new feature of Mid Journey V6. Despite its inaccuracies, the presenter is excited about the potential of the software and its future updates.

15:03

📚 Text Generation and Comic Creation in Mid Journey V6

The script concludes with the presenter testing the text generation capabilities of Mid Journey V6. The presenter asks the software to imagine a man in a dystopian city with a neon ad, but the results are not as expected, with the first image being identical to the Cyberpunk 2077 poster. After adjusting the prompt and quotes, the results improve, but the presenter notes that Mid Journey V6 still struggles with numbers and the exactness of the request. The video also explores the possibility of creating comics with the software, with the presenter providing a detailed prompt for a four-panel comic page. Although the images generated are not perfect and require manual adjustment for text, the artwork is praised. The presenter acknowledges that Mid Journey V6 needs further polishing and suggests that the official guide on writing prompts released by the developers may help achieve better results in the future. The video ends with the presenter expressing excitement for the full release of Mid Journey V6 and the potential to compare it with other AI image generators like Dolly 3 and Adobe Firefly.

Mindmap

Keywords

💡Midjourney V6

Midjourney V6 refers to the latest version of the AI image generation software, which has been updated with significant improvements over its previous versions. The update includes enhanced text generation capabilities, better prompt following, and more accurate depictions in the generated images. It is mentioned in the video as a major update that users should try out, and it is the central theme around which the video's content revolves.

💡Text Generation

Text generation is the process by which the AI software creates written content based on the input prompts given by the user. In the context of Midjourney V6, this feature has been notably improved, allowing the software to produce text within the generated images more accurately. The video demonstrates this by showing examples of images with correctly spelled text, which was not always the case in previous versions of the software.

💡Prompts

Prompts are the input text provided by users to guide the AI in generating specific images. The video discusses how Midjourney V6 has become more sensitive to the wording of prompts, requiring users to relearn how to effectively communicate their desired outputs. The video also mentions the ability to accept longer prompts and specify more details, such as colors and subject positioning, which can result in more accurate and detailed images.

💡Aspect Ratio

Aspect ratio is the proportion of an image's width to its height, which is important for ensuring that the generated images fit the desired format. In the video, the aspect ratio is specified using the 'AR 16x9' notation, which is an example of a common format used for widescreen displays. The software uses this information to generate images that maintain the correct proportions, as requested by the user.

💡Upscaling

Upscaling refers to the process of increasing the resolution of an image, typically to make it appear clearer and more detailed. The video discusses two types of upscaling: subtle and creative. Subtle upscaling aims to improve clarity without altering the style, while creative upscaling may introduce artistic changes to enhance the image's appearance. The video shows examples of upscaling in action, demonstrating how it can improve the quality of the generated images.

💡Chaos Parameter

The chaos parameter is a feature in Midjourney V6 that introduces an element of randomness to the image generation process. By specifying a number for the chaos parameter, users can influence the variability and unpredictability of the output. The video demonstrates this by using a chaos value of five, resulting in images with different styles and levels of detail, showcasing the parameter's impact on the final product.

💡Stylized Argument

The stylized argument is a numerical parameter in Midjourney V6 that influences the aesthetic style of the generated images. It ranges from 0 to 1000, with lower values potentially offering better prompt understanding and higher values resulting in more aesthetically pleasing images. The video shows that using a maximum stylized value can produce images that are either realistic or highly stylized, depending on the context of the prompt.

💡Style Raw

Style raw is a parameter in Midjourney V6 that, when used, aims to generate images that are more photographic and less stylized. It is intended to produce more literal interpretations of the user's prompts. The video demonstrates this by using the style raw parameter and obtaining images that closely resemble photographs, albeit with some imperfections.

💡Comic Creation

Comic creation is the process of generating a series of images that tell a story or convey a message in a comic book format. The video discusses testing Midjourney V6's capabilities in creating a four-panel comic, with each panel described in detail. While the results are not perfect and require manual adjustment of text, it shows the potential for using AI to assist in comic book creation, offering a new avenue for artists and creators.

💡Dystopian City

A dystopian city is a fictional urban environment characterized by oppressive societal control, suffering, and a lack of human freedom. In the video, the term is used in a prompt to generate an image of a man standing in front of a dystopian city landscape with a neon ad. This concept is often used in cyberpunk-themed content, which is a subgenre of science fiction that typically features advanced technology and futuristic settings.

💡Cyberpunk Style

Cyberpunk style is a visual and thematic style often associated with science fiction media, characterized by a blend of futuristic technology, dystopian settings, and a dark, gritty atmosphere. In the video, the cyberpunk style is requested to generate an image of a man in a dystopian city with a neon ad. The term is used to guide the AI in creating visuals that align with the distinctive aesthetic of the cyberpunk genre.

Highlights

Midjourney V6 is now available, marking the biggest update yet with improvements in every part of the system.

The Alpha version of the V6 model was released on December 21st to all users.

To access the updated model, add 'D-V6' to your prompt; remember to include a space between 'D-' and 'V6'.

Midjourney V6 introduces the ability to generate text, a significant feature for users.

The new model is more sensitive to prompts and can handle longer, more complex prompts effectively.

Users should avoid using nonsensical terms like 'award-winning photo realistic 4K 8K' as the model generates images in roughly the same resolution.

The update allows for creative upscaling of images, improving the quality and detail of the generated content.

The community has discovered that the new prompt length is now over 350 words, enabling more detailed and specific requests.

Midjourney V6 now supports specifying colors and placing objects precisely in the generated images.

The update includes the ability to prompt for more than one subject and interact with Midjourney like a chatbot.

The initial prompt structure is crucial for achieving desired results in text generation.

Midjourney V6 can analyze images and transform them into different artistic styles, like oil paintings or da Vinci sketch styles.

Creating comics with Midjourney V6 is possible, but it still requires manual input for text and may need further refinement.

Developers have released an official guide on writing prompts for better results with the V6 update.

Midjourney V6 is currently in its alpha stage and will receive regular updates to improve its functionality further.

The update has improved the accuracy of text spelling in generated images, a notable enhancement from previous versions.

The style of generated images varies, offering users a range of aesthetic options to choose from.

Midjourney V6's aspect ratio parameter 'AR 16x9' ensures the correct proportions in the generated images.