BIG Midjourney Update ~ IN-PAINTING Full Breakdown

Future Tech Pilot
22 Aug 202316:27

TLDRThe video provides an in-depth look at the new 'In-Painting' feature on the Midjourney platform. The host demonstrates how to use the feature to edit specific areas of an image, such as changing a man's portrait to show him laughing or angry, and even transforming a woman into a male character or an Android. Tips include using the lasso tool for precise selection, writing detailed prompts, and adjusting the 'chaos' and 'stylize' parameters for more control over the output. The feature is shown to work on mobile and can be applied to old pictures by copying the job ID. However, the video also highlights the feature's limitations, such as unpredictable results and instances where it doesn't work as expected. The host suggests that seeds may not influence the outcome and that the feature is still in its early stages, with potential for future improvements.

Takeaways

  • ๐ŸŽจ **In-Painting Feature**: The new mid-journey feature 'In-Painting' allows users to edit specific parts of an image without affecting the rest.
  • ๐Ÿ–ผ๏ธ **Remix Mode**: Having 'Remix mode' turned on is crucial for the 'In-Painting' feature to work effectively.
  • โœ‚๏ธ **Selection Tools**: The 'lasso tool' provides better control over the area to be edited, although it might be a bit finicky to use.
  • ๐Ÿ“ **Prompt Editing**: Users can add or modify words in the prompt while using 'In-Painting' to achieve different results.
  • ๐Ÿ˜€ **Emotion in Faces**: The feature can change facial expressions, but it might not always connect the emotion to the eyes correctly.
  • ๐Ÿ‘“ **Creative Edits**: 'In-Painting' can be used to add accessories like sunglasses, but the results may vary and require specific prompts.
  • ๐Ÿ” **High Variation Mode**: For users seeking diverse outcomes, the 'high variation mode' is recommended during the 'In-Painting' process.
  • ๐Ÿ‘ค **Character Transformation**: The feature can seamlessly insert a new character into an image, changing the original subject entirely.
  • ๐Ÿ“ฑ **Mobile Compatibility**: 'In-Painting' is accessible and functional on the mobile Discord app.
  • ๐Ÿšซ **Seeds Irrelevance**: The script suggests that the use of seeds may not influence the 'In-Painting' results as expected.
  • ๐Ÿค– **Creative Potential**: The feature shows promise for creative editing, such as turning a character into an android with the right prompts.

Q & A

  • What is the new feature introduced in the mid-journey update?

    -The new feature introduced is called 'In-Painting', which allows users to edit specific areas of an image, such as adding elements or changing features.

  • What is the importance of having 'Remix mode' turned on for In-Painting?

    -With Remix mode turned on, users can edit the prompt while using the In-Painting feature, allowing for more control and customization over the generated image.

  • How does the 'Variation options' affect the In-Painting results?

    -The 'Variation options', which include low and high variation modes, influence the diversity of the generated images. High variation mode produces more distinct and varied results, whereas low variation mode generates images that are more similar to each other.

  • Why is it suggested to write a longer and more specific prompt when using In-Painting?

    -Writing a longer and more specific prompt helps the In-Painting feature to generate more accurate and desired results, as it provides clearer instructions for the AI to follow.

  • How can the 'Chaos' and 'Stylize' parameters impact the In-Painting results?

    -The 'Chaos' parameter introduces more variation into the generated images, while a higher 'Stylize' value makes the AI's interpretation more literal. Adjusting these parameters can significantly affect the style and appearance of the final image.

  • What is the significance of the 'Undo' button in the In-Painting tool?

    -The 'Undo' button allows users to revert changes made during the In-Painting process, which is useful for correcting mistakes or trying different editing approaches.

  • Can the In-Painting feature be used on old pictures?

    -Yes, the feature can be used on old pictures by reacting with the envelope emoji on the image and then using the job ID to bring up the image for In-Painting.

  • What is the role of the 'Seed' in the In-Painting process?

    -The 'Seed' is used to generate variations of the image. However, the script suggests that seeds may not have a significant impact on the In-Painting feature, as the same seed can produce different results.

  • How does the In-Painting feature handle adding multiple characters to an image?

    -The In-Painting feature can add multiple characters by selecting a larger area than expected. It may not always perfectly represent the characters but can still generate acceptable results.

  • What are some limitations or challenges of using the In-Painting feature?

    -The In-Painting feature may not always work as expected, and there can be inconsistencies in the results. It may struggle with certain prompts or fail to make significant changes to the image.

  • What additional insights can the audience gain from the video script regarding the In-Painting feature?

    -The audience can learn about various tips and tricks for using the In-Painting feature effectively, such as the importance of a detailed prompt, using the lasso tool for better control, and understanding the impact of different parameters on the final image.

Outlines

00:00

๐ŸŽจ Introduction to Mid-Journey's In-Painting Feature

The video introduces a new feature called 'In-Painting by Region' from Mid-Journey, which allows users to edit specific areas of an image. The host demonstrates the feature using a portrait and explains the importance of enabling 'remix mode' for more control. Various tools like the lasso and rectangle tools are showcased for selecting areas to edit. The video also covers how to adjust prompts for different outcomes and the potential unpredictability of the feature's results.

05:02

๐Ÿ–Œ๏ธ Exploring In-Painting with Different Prompts and Settings

The host shares different examples of using the In-Painting feature, such as changing eye color and adding sunglasses to a portrait. The importance of writing detailed prompts is emphasized, as is the impact of the 'variation' settings on the output's diversity. The video also discusses the ability to transform characters and the influence of 'stylize' and 'chaos' parameters on the image generation process.

10:04

๐Ÿ”„ Advanced Techniques and Mobile App Compatibility

Advanced techniques for using the In-Painting feature are explored, including panning for more image area and using custom zoom to refresh the image options. The video demonstrates how to fix specific parts of an image, like hands, and change hair color. It also addresses the mobile app's compatibility with the feature and provides a trick for using In-Painting on older pictures without the button.

15:05

๐Ÿš€ Conclusion and Testing the Limits of In-Painting

The host concludes with additional examples, such as adding characters to a scene and attempting to generate text with mixed results. They discuss the limitations and the potential for future improvements. There's also a mention of the feature being subject to change and an invitation for viewers to share their findings in the comments section.

Mindmap

Keywords

๐Ÿ’กIn-Painting

In-Painting is a feature in image editing software that allows users to fill in or remove selected parts of an image in a way that appears seamless and natural. In the context of the video, it is used to modify and enhance images, such as changing facial expressions or adding elements like sunglasses or hair color to a portrait. It's a core part of the video's demonstration, showcasing its capabilities and limitations.

๐Ÿ’กRemix Mode

Remix mode is a setting that, when enabled, allows users to edit the original prompt while using the in-painting feature. This is significant in the video as it provides more control and customization options for the in-painting process. The presenter uses this mode to change the expression of a portrait from neutral to 'laughing' and 'angry,' highlighting the feature's flexibility.

๐Ÿ’กLasso Tool

The Lasso tool is a selection tool used in image editing that allows for freehand drawing to make selections. In the video, the presenter uses the Lasso tool to make precise selections for the in-painting feature to act upon. It's depicted as a bit finicky but a powerful tool for detailed in-painting tasks.

๐Ÿ’กVariation Options

Variation options refer to settings that control the degree of change or diversity in the output images generated by the in-painting feature. The video mentions 'low variation mode' and 'high variation mode,' which affect how similar or different the resulting images are. This is important for users looking for a wide range of options or a more consistent set of edits.

๐Ÿ’กChaos and Stylize Parameters

Chaos and Stylize are parameters within the in-painting feature that influence the level of randomness and style in the generated images. A higher chaos value results in more varied outputs, while adjusting the stylize parameter can make the image more literal or abstract. These parameters are showcased in the video as ways to fine-tune the creative output of the in-painting tool.

๐Ÿ’กJob ID

A Job ID is a unique identifier assigned to a specific task or operation within a software system. In the context of the video, the presenter demonstrates how to use a Job ID to retrieve and manipulate older images using the in-painting feature. It's a crucial detail for users looking to work with existing images that don't initially display the in-painting option.

๐Ÿ’กSeeds

Seeds are values used in generative systems to produce a specific outcome or set of outcomes. The video discusses how seeds do not seem to affect the results of the in-painting feature, indicating a limitation or a feature that may be bypassed in the current implementation. This is an important consideration for users relying on predictability in their image generation.

๐Ÿ’กPanning Feature

The Panning feature allows users to expand the frame of an image in a specific direction, which can be useful for adding more context or content to an image. In the video, the presenter uses the panning feature to add more of the background to an image of 'Marcia the Martian,' demonstrating how it can be combined with in-painting for more extensive image manipulation.

๐Ÿ’กCustom Zoom

Custom Zoom is an advanced function that enables users to adjust the zoom level of an image for further editing. The video shows a trick involving Custom Zoom to refresh the image and regain access to other editing options that were not available initially. This technique is particularly useful for users who need more control over the editing process.

๐Ÿ’กImage Prompt

An Image Prompt is a description or another image used to guide the in-painting feature to generate specific content. The video explores the concept of using an image prompt to influence the in-painting process, such as adding a car to a country road scene. Although it doesn't always work as expected, it demonstrates the potential for more complex and targeted image manipulation.

๐Ÿ’กGraffiti

Graffiti refers to the images or text that are added to a surface, such as a wall, in a form of street art. In the video, the presenter tests the in-painting feature's ability to add graffiti to a brick wall image. The successful addition of graffiti-like elements to the wall illustrates the feature's capability to understand and apply complex visual textures.

Highlights

Midjourney has introduced a new feature called 'In-Painting', which allows for editing specific areas of an image.

In-Painting can be activated by entering '/remix' in the prompt box to turn on Remix mode.

The feature offers three variation options: very strong, very subtle, and very region.

Using the lasso tool provides more precise control over the area to be edited.

The transcript demonstrates how to transform a portrait of a man into a laughing man using In-Painting.

In-Painting can make significant changes to a subject's facial expression, though sometimes the result may not be perfect.

The feature can also be used to change a character's gender, as shown by turning 'Marsha' into a man.

Writing a longer and more specific prompt can yield better results with In-Painting.

The settings include options for low and high variation modes, affecting the diversity of the results.

In-Painting can introduce new characters into an existing scene, maintaining the original background.

The feature works on mobile platforms, including the Discord app.

Seeds do not appear to influence the outcome of In-Painting results.

In-Painting can be used to edit old pictures by utilizing the job ID and the 'show' command.

The feature can sometimes fail to make expected changes, indicating it is not foolproof.

In-Painting can fix specific parts of an image, such as hands, without needing a detailed prompt.

The panning feature of Midjourney can be used in conjunction with In-Painting for expanded editing.

In-Painting allows for creative transformations, such as turning a character into an Android or adding blonde hair.

The transcript provides various examples of how In-Painting can be used to add elements like graffiti to a wall or characters to a scene.

The feature is in its early stages and subject to change, with potential for more capabilities in the future.