Midjourney Panning + Dreamshaper Version 7

In this video, Team Nostromo introduces the new panning function in Midjourney and the features of Dreamshaper Version 7. The panning function allows for the creation of expansive cyberpunk cityscapes, with the ability to extend the image in any direction. The Dreamshaper model is showcased for its versatility in producing photorealistic, digital art, and anime images. The video provides tips on using the model effectively, including the use of negative prompts to improve image quality. It also demonstrates how to enhance the sharpness and detail of the generated images by adjusting prompts and using tools like the Detail Tweaker. The presenter shares examples of various styles created with the model and invites viewers to suggest the next model to explore.

Takeaways

• 🎨 **Midjourney Panning Feature**: The new panning function allows users to extend their image as wide as desired, adding parts to a scene in a coherent way.
• 🖼️ **Image Upscaling**: To use panning, first create an image and upscale it using the 'U' buttons, then use the arrow keys to extend the image in the chosen direction.
• ➡️ **Directional Extension**: After choosing a direction for extension, the same choice will be presented until the user decides to change it.
• 🔧 **Remix Mode Activation**: If the panning option doesn't show up, users should go to mid-journey, type '/settings', and activate the remix mode.
• 🔄 **Dream Shaper Version 7**: This version of the model can create photorealistic, digital art, and anime images, offering a wide range of styles.
• 📈 **Quality Improvements**: Dream Shaper 7 has seen improvements in image quality, particularly with hair and armor details.
• 🚫 **Negative Prompts**: The use of negative prompts like 'bad dream' and 'unrealistic dream' can help refine the output and address issues in the model's output.
• 📁 **File Organization**: Model files should be downloaded into the 'models' folder, while negative embeddings should go into the 'embeddings' folder.
• 🔄 **CLIP Skip Usage**: The model utilizes CLIP Skip 2, which is crucial for achieving the correct output and should be selected in the settings.
• 🔍 **Image Sharpness**: The model sometimes struggles with sharpness in upscaled images, but specific prompts and techniques can improve the quality.
• 🌟 **Detail Tweaker**: Tools like the 'Laura funic stress version 2' and 'one button prompt extension' can enhance details and suggest prompts for better image results.

Q & A

What is the new feature of mid-journey panning in Dream Shaper Version 7?

The new feature of mid-journey panning in Dream Shaper Version 7 allows users to extend their image in any direction as wide as they want, adding parts to the scene that build up the overall composition.

How does the panning function work in Dream Shaper Version 7?

The panning function works by first creating an image and then upscaling it using one of the 'U' buttons. After upscaling, arrows appear that point in different directions, allowing the user to extend the image in the chosen direction.

What should you do if the panning option does not show up?

If the panning option does not show up, you should go to mid-journey, type '/settings', and activate the remix mode in the settings window.

What are the capabilities of the Dream Shaper 7 model?

The Dream Shaper 7 model is capable of producing photorealistic images, digital art, and even anime-style images, offering a wide range of styles in high quality.

How can you improve the image quality using negative embeddings?

Negative embeddings can be used to improve image quality by specifically addressing and reducing the impact of certain undesirable features in the model's output. They are applied by downloading them into the 'embeddings' folder and then selecting them in the 'negative prompt' field within the application.

What is the significance of the CLIP Skip setting in the model?

The CLIP Skip setting is important as it determines the version of the skip mechanism used by the model. Using the correct version, such as CLIP Skip 2, ensures that the model functions optimally.

How can you add negative embeddings to your negative prompt?

To add negative embeddings to your negative prompt, click inside the negative prompt field and then select the negative embedding you wish to use from the available options.

What is the recommended resolution for images created with Dream Shaper Version 7?

The recommended resolution for images created with Dream Shaper Version 7 is 1024 by 1536, as suggested by the model's page.

How can you achieve better quality in upscaled images?

To achieve better quality in upscaled images, you can use the 'highly detailed' and 'sharp' prompts with increased weights, and also utilize tools like the 'Laura at Detail' tweaker to enhance the details in the image.

What is the purpose of the 'one button prompt extension' mentioned in the script?

The 'one button prompt extension' is a tool that suggests prompts for image creation, making it easier for users to generate creative and detailed images without manually crafting the prompts themselves.

How can you find additional settings like CLIP Skip and SDVAE Quick Chooser in the application?

You can find additional settings like CLIP Skip and SDVAE Quick Chooser by going to the settings, selecting the user interface on the left side, and scrolling down to the quick setting list where you can select and save these options.

What is the trick mentioned for getting a sharper and more detailed upscaled image?

The trick for getting a sharper and more detailed upscaled image involves using the 'sharp' prompt with double round brackets to increase its impact, and possibly adding 'highly detailed' to the prompt for even more detail.

Outlines

00:00

🖼️ Mid-Journey Panning and Dream Shaper 7 Overview

The first paragraph introduces the new panning function in Mid-Journey, a tool used for creating and editing images. The speaker, from team Nostromo, demonstrates the ability to pan across a wide, cyberpunk cityscape, emphasizing the ease of use and the impressive results. They explain the process of upscaling an image and extending it in any direction using arrows that appear after upscaling. The user is guided to activate the remix mode in the settings for full functionality. The paragraph concludes with an introduction to Dream Shaper version 7, a versatile model capable of creating photorealistic, digital art, and anime images. The speaker shares their experience with the model, focusing on its realistic style, and provides tips on how to use the model effectively, including the importance of using the correct CLIP skip version.

05:00

📈 Enhancing Image Quality with Negative Embeddings and Settings

The second paragraph delves into the specifics of improving image quality using negative embeddings in the Dream Shaper 7 model. The speaker explains how to add negative prompts to refine the output and avoid unwanted elements like 'bad dream' and 'unrealistic dream'. They guide users on where to find and how to input CLIP skip and SDVAE quick choices in the settings. The paragraph showcases examples of image outputs, highlighting the model's proficiency with photorealistic details like hair and chainmail. The speaker also discusses common upscaling issues and shares a trick to achieve better image quality. They emphasize the model's effectiveness with detailed prompts and the use of the 'Laura funic stress' version 2 for additional detail enhancement.

10:02

🎨 Exploring Artistic Styles and Image Enhancement Techniques

The third paragraph discusses various artistic styles that can be achieved with the Dream Shaper 7 model, showcasing a range of visually stunning images. The speaker provides insights into how to enhance image details and resolution, suggesting the use of specific terms like 'full sharp' and 'highly detailed' within the prompt. They also mention the 'Laura at detail' tweaker as a useful tool for improving image quality. The paragraph includes a direct comparison to demonstrate the effectiveness of these techniques. The speaker concludes by inviting viewers to share their feedback and preferences for future video topics and signs off with a prompt for engagement, such as liking the video.

Keywords

Midjourney Panning

Midjourney Panning refers to a new feature in the video that allows for the expansion of an image in various directions, such as left, right, up, or down. This feature is significant as it enables the creation of a wider view of a scene, enhancing the visual storytelling. In the context of the video, it is used to extend a cyberpunk cityscape, adding to the depth and detail of the image.

Dreamshaper Version 7

Dreamshaper Version 7 is a new version of a model discussed in the video that is capable of generating photorealistic, digital art, and even anime images. It is highlighted for its versatility and the high quality of its outputs across different styles. The video demonstrates its use with various prompts and settings to achieve a range of artistic effects.

Upscaling

Upscaling is the process of increasing the resolution of an image, which is a common technique in digital art and photography to improve the clarity and detail of visuals. In the video, upscaling is mentioned in the context of enhancing the quality of the cyberpunk city image and other examples, with the use of specific buttons to perform the action.

Negative Prompt

A Negative Prompt in the context of the video refers to specific instructions given to the AI model to avoid generating certain unwanted elements or styles in the output image. The video discusses 'bad dream' and 'unrealistic dream' as negative prompts, which are used to refine the image quality and direct the AI towards more desirable results.

Clip Skip

Clip Skip is a parameter used in the AI model that affects how the model processes the input to generate an image. The video emphasizes the importance of using the correct version of Clip Skip (in this case, version 2) for the best results with the Dreamshaper model. It is part of the technical settings that users need to configure to achieve optimal image generation.

Photorealism

Photorealism is a style of art where the subject is depicted with a high degree of realism, resembling a photograph. The video discusses the Dreamshaper model's ability to generate photorealistic images, particularly noting the quality of hair and armor in the samples provided, which is a testament to the model's capability to handle fine details.

CFG Scale

CFG Scale is a setting within the AI model that controls the level of detail or 'creativity' of the generated image. The video does not provide explicit details about CFG Scale, but it is implied to be one of the parameters that users can adjust to fine-tune the output of the model.

Stable Diffusion

Stable Diffusion is likely referring to the AI model's approach to generating stable and coherent images from textual prompts. While the term is not explicitly defined in the video, it is implied to be a part of the technology that enables the creation of high-quality images, as it is mentioned alongside the model file and negative embeddings.

Textual Inversion

Textual Inversion is a feature within the AI model that allows users to include specific negative embeddings in their prompts to guide the image generation process. The video demonstrates how to activate and use bad dream and unrealistic dream negative embeddings through the Textual Inversion tab to improve the output quality.

Detail Tweaker

A Detail Tweaker is a tool or setting that enhances the detail in the generated images. The video mentions a 'Laura at Detail' tweaker, which is used to add more detail to the dress in one of the examples, making the image sharper and more visually appealing.

One Button Prompt Extension

The One Button Prompt Extension is a tool that suggests prompts for image generation, simplifying the process for users. The video showcases how this extension can be used to create a variety of images with minimal input, highlighting its utility in generating creative and detailed visuals.

Highlights

Midjourney Panning and Dreamshaper Version 7 introduces a new panning function for creating expansive cyberpunk city scenes.

The panning feature allows users to extend their images as wide as desired, enhancing the scene with added details.

To upscale an image, click on one of the 'U' buttons, then use the directional arrows to extend the image in the chosen direction.

If the panning option doesn't show up, activate the remix mode by typing '/settings' in the mid-journey interface.

Dreamshaper Version 7 is capable of producing photorealistic, digital art, and anime images.

The model has been tested extensively with realistic styles and can handle hair and shining armor with high quality.

Negative prompts such as 'bad dream' and 'unrealistic dream' are used to improve the model's output.

Negative embeddings are downloaded into the 'embeddings' folder and can be activated through the 'textual inversion' tab.

Clip Skip 2 is a crucial setting for this model, which can be found and selected in the user interface settings.

Highly detailed skin with a weight of 1.2 can be added to prompts to enhance facial details.

The model excels at rendering photorealistic hair, showing each strand with clarity.

Upscaling can sometimes result in a lack of sharpness; a trick will be shown later to improve this.

Chainmail reflections, including cityscapes, can be rendered with impressive realism.

The 'Laura funic stress version 2' and 'add detail detail tweaker' are tools used to enhance image details.

One-button prompt extension is a useful tool for generating creative and detailed prompts.

To improve image sharpness, use terms like 'full sharp' with double round brackets and 'highly detailed' in the prompt.

The model is effective for artistic styles, producing images that are crisp and visually appealing when viewed at typical sizes.

The video provides a direct comparison showcasing the impact of using detailed prompts on image quality.

Different artistic styles can be achieved with this model, as demonstrated by the variety of example images.