BIGGEST Midjourney Update EVER - Consistent Style!

The video introduces a significant update to the AI image generation tool, Midjourney, focusing on the new feature of consistent styles. Users can now input a reference image to guide the style of all subsequent image generations, which can be uploaded via Discord. The process involves typing 'D-sref' followed by the image link in the prompt box. The video demonstrates how this feature allows for the creation of images in a specific style, such as a dog, cat, and stallion, with the same color palette and artistic nuances. The narrator, Nolan, also discusses the new 'style weight' parameter, which adjusts the influence of the reference image on the generated images, and the importance of providing context in the prompt for better results. Additionally, the video shows how to combine multiple style references and assign different weights to them for a blended style. The update is presented as a game-changer for AI image generation, simplifying the process and allowing for more creative control.

Takeaways

• 🎨 **Consistent Style Feature**: Midjourney introduces a new feature that allows for consistent styles across image generations using a reference image.
• 📄 **Uploading Reference Image**: Users can upload a reference image via Discord's upload feature to set the style for all generations.
• 🔗 **Linking the Reference**: The new parameter '-d sref' is used in the prompt box, followed by a space and the link to the reference image.
• 🐕 **Style Application**: The style of the reference image is applied to any subject matter, as demonstrated with a dog, cat, and stallion.
• 🎭 **Beyond Color**: The style reference captures not just color but also intricate details and nuances of the chosen style.
• 🔢 **Image Weight Parameter**: A new 'style weight' parameter can be adjusted between 0 and 1000, with 100 being the default setting.
• 🦇 **Batman Example**: Style weight differences are demonstrated using Batman images, showing the impact of varying the weight from 1000 to 0.
• 📝 **Context in Prompt**: Providing more context in the prompt helps in generating images that are more consistent with the reference style.
• 🔥 **Stylize Value**: A high stylize value (e.g., s1000) is recommended for better results with style references, enhancing the generation's style consistency.
• 🤹 **Combining Styles**: Multiple style references can be combined in a single generation, allowing for a blend of different styles.
• ⚖️ **Weighting Individual Styles**: Individual style references can be weighted to control the prominence of each style in the final generation.

Q & A

What is the new feature called in Midjourney that allows for consistent styles across images?

The new feature is called 'Style reference' or 'srf', which allows users to provide a reference image for Midjourney to use as a style guide for all subsequent generations.

How do you upload a reference image to Midjourney through Discord?

You can upload the reference image by clicking on the little plus button beside the prompt box, then selecting 'upload file' and choosing the image from your computer.

What is the parameter needed to trigger the consistent style feature in Midjourney?

To trigger the consistent style feature, you need to type in the parameter 'D- sref' followed by a space and then drag and drop the reference image into the prompt box.

Can the reference image for the consistent style feature only be an image created by Midjourney?

No, the reference image can be any image, it does not have to be made by Midjourney. The user can select any image they prefer for the style reference.

What is the minimum style weight that should be used for the consistent style feature to have an effect?

The minimum style weight that should be used is 20. Below this value, the reference image has no effect on the generations.

What is the default style weight in Midjourney's consistent style feature?

The default style weight in Midjourney's consistent style feature is 100.

How does providing more context in the prompt help with the consistent style feature?

Providing more context in the prompt helps the AI to better understand the desired outcome and maintain consistency with the style reference. It aids in generating images that are not just similar in color but also in style and details.

What is the recommended stylize value to use when working with style references in Midjourney?

The recommended stylize value to use with style references is s1000, as it plays a significant role in the quality of the generations.

Can you combine more than one style reference in Midjourney's consistent style feature?

Yes, you can combine multiple style references by listing them after the 'D- sref' parameter with each reference link followed by a space and then dragging the images into the prompt box.

How do you add a weight to individual style references when combining them?

To add a weight to individual style references, after each image link, you add a space, then two colons, and then the number representing the weight. For example, 'D- sref <image link> ::2'.

What is the significance of the chaos value when using style references in Midjourney?

The chaos value, when increased, can result in more creative and varied generations while still maintaining the consistency of the style reference. It is recommended to use higher chaos values like 50 for interesting results.

What does the presenter, Nolan, aim to achieve with his videos about AI?

Nolan aims to make learning about AI as straightforward as possible, helping viewers understand and utilize features like Midjourney's consistent style update.

Outlines

00:00

🖼️ Introducing Consistent Styles Feature

The video begins with an introduction to the new feature called 'Consistent Styles' in MidJourney, which allows users to maintain a consistent style across all image generations by using a reference image. The process involves uploading the reference image to Discord, selecting subjects, and using the 'D-sref' parameter followed by the image link to trigger the consistent style. The video demonstrates how this feature works with various subjects, such as a dog, a cat, a stallion, and even poetry, showcasing how each subject takes on the style of the reference image. The presenter, Nolan, emphasizes the ease of use and the powerful results that can be achieved without the need for complex style tuning or remixing.

05:02

🎨 Adjusting Style Weight and Context in Prompts

The second paragraph delves into adjusting the 'style weight' parameter, which ranges from 0 to 1000, to control the influence of the reference image on the generated images. Nolan explains that the default setting is 100, but it can be fine-tuned according to the user's needs. He provides examples of how different style weights affect the output, from a weight of 1000, which closely matches the reference image, to a weight of 0, which results in no discernible influence from the reference image. Additionally, Nolan discusses the importance of providing context in the prompt to enhance the generation results. He also touches on the use of a high 'stylize' value, which he recommends for better results when using style references. Finally, he explores the possibility of combining multiple style references and how to apply different weights to them for a more nuanced outcome.

Keywords

💡Consistent Style

Consistent Style refers to the ability of an AI image generation system to produce images that maintain a uniform visual theme or aesthetic across multiple generations. In the context of the video, it is a new feature of the Midjourney system that allows users to input a reference image to guide the style of subsequent images generated by the system. This feature is significant as it simplifies the process of achieving a desired artistic style without the need for complex tuning or remixing.

💡Midjourney

Midjourney is an AI-driven image generation platform that allows users to create unique visuals by providing prompts and reference images. The system uses advanced algorithms to interpret these inputs and generate corresponding images. In the video, Midjourney is highlighted for its update that introduces consistent styling across image generations, which is a significant advancement for users seeking a cohesive artistic theme.

💡Reference Image

A Reference Image is a specific example or sample image that users can provide to the AI system to guide the style and appearance of the generated images. It serves as a visual template for the AI to follow. In the video, the reference image is crucial for the consistent style feature, as it dictates the color palette, artistic details, and overall look that the AI should replicate across different subjects.

💡Discord

Discord is a popular communication platform that supports text, voice, and video conversations. In the video, it is mentioned as a medium through which users can upload their reference images to Midjourney. The process involves using the platform's file upload feature to share the image that will influence the style of the AI-generated content.

💡Prompt Box

The Prompt Box is an input field within the Midjourney interface where users type in their requests or instructions for the AI to generate specific images. It is a key component of the user interaction with the system. In the video, the prompt box is where users type in the subject of the image they want to generate, along with the new 'D-sref' parameter to trigger the consistent style feature.

💡D-sref

D-sref stands for 'Style reference' and is a new parameter introduced in the Midjourney update. It is used in the prompt box to tell the AI to apply the consistent style from a provided reference image to the generated images. The script demonstrates how to use this parameter by typing 'D-sref' followed by uploading the reference image to ensure that all generated images match the desired style.

💡Style Weight

Style Weight is a numerical value between zero and 1,000 that users can set to determine the influence of the reference image's style on the generated images. A higher style weight means the generated images will closely resemble the style of the reference image. In the video, it is demonstrated that the default style weight is 100, and the minimum effective weight is 20, below which the style reference has no discernible effect on the image generation.

💡Stylize Value

The Stylize Value is a parameter in the Midjourney system that controls the level of stylization applied to the generated images. A higher stylize value increases the artistic interpretation of the input prompt. In the video, it is suggested that using a high stylize value, such as 1,000, in conjunction with the consistent style feature can produce particularly striking and coherent artistic results.

💡Chaos Value

Chaos Value is another parameter in the Midjourney system that introduces randomness or variability into the image generation process. It is used to create more diverse and less predictable outputs. The video highlights that when using style references, a higher chaos value can still yield coherent and interesting images, which would not be the case with a regular prompt.

💡Combining Style References

Combining Style References is the ability to use more than one reference image to influence the style of the generated images. This allows for a blend of styles from different sources. In the video, it is shown how to include multiple 'D-sref' parameters in the prompt, and optionally assign different weights to each reference image to control their relative influence on the final style.

💡PDFs

PDFs, or Portable Document Format files, are a type of document that can be shared and viewed on various devices without losing formatting. In the video, the host mentions that there are free PDFs available in the description below, presumably related to the Midjourney system or AI image generation. These documents likely provide additional information or tutorials for users to enhance their understanding and use of the platform.

Highlights

Midjourney introduces consistent style feature for image generation.

Users can provide a reference image for all generations to match the style.

The reference image can be uploaded to Discord through the prompt box.

The new parameter 'D-sref' is used to trigger the consistent style.

Any subject can be translated to the style of the reference image.

Style tuning and remix features are no longer needed with this update.

The style reference captures color palette and fine details.

An image weight parameter is introduced, with a default of 100.

The minimum style weight that should be used is 20.

Providing more context in the prompt helps achieve consistent style.

High stylize value (s1000) is recommended for better results with style references.

Chaos value can be increased for better integration with style references.

Multiple style references can be combined in a single generation.

Individual weights can be assigned to each style reference for better control.

The feature allows for fine-tuning the influence of each style reference.

Nolan emphasizes the significance of this update for AI image generation.

Free PDFs are available for more information based on version 6.

Subscribers will be updated on new features and tutorials.