Underused Midjourney v5 Prompt Commands :: How to use Text Weight and Image Weight
TLDRIn this informative video, the host delves into the underutilized yet powerful techniques of Image Weight and Text Weight in mid-journey prompts. The video explains how these weights can be used to control the output of an image by adjusting the emphasis on certain keywords within the prompt. The host provides a clear explanation of how prompts work, with an emphasis on the importance of the order of keywords and the distribution of tokens. Practical examples are given to illustrate the use of text weights, such as adjusting the ratio of tokens to emphasize specific elements in an image. The video also covers the limitations of weights and offers solutions, such as using natural language prompts for more accurate results. Additionally, the host discusses the use of reference images and negative prompts to refine the image generation process. The video concludes with a demonstration of how to achieve desired results through a combination of text and image weights, and by using photobashing techniques to overcome challenges in image generation.
Takeaways
- 📝 **Understanding Prompts**: Midjourney scans prompts for keywords and assigns tokens to assemble images based on its database.
- 🔍 **Token Allocation**: Midjourney assigns about 75 tokens per prompt, with more emphasis on keywords at the beginning of the prompt.
- ⚖️ **Text Weighting**: Use `::` followed by a number to increase the importance of a keyword, with no space before the `::` but a space after the number.
- 🔢 **Weighting Format**: Incorrect formatting, such as adding a space before `::` or not using a comma after the weighted number, can cause Midjourney to ignore the weight.
- 🍰 **Example of Text Weight**: Weighting 'cupcake' with `::1` and 'coffee' with `::2` results in an image favoring coffee over the cupcake.
- 🖼️ **Image Weighting**: Use `--image_weight` or `--IW` followed by a number between 0.5 and 2 to control the influence of a reference image on the output.
- 🎨 **Artistic Style**: Image weights can be adjusted to incorporate elements of an artist's style, such as Jaylee's, into the generated image.
- 🔄 **Negative Prompting**: Adding a negative number after `::` can theoretically remove elements from the image, but it can be tricky and may not always work as expected.
- 🤔 **Challenges with Negative Prompts**: Midjourney might ignore negative prompts if they conflict with the reference image or the model's understanding.
- 📸 **Photobashing**: When direct text prompts fail, photobashing techniques can be used to manually adjust the generated image to meet the desired outcome.
- 🎥 **Project Example**: The speaker used Midjourney's features to create images for a fictional documentary about a 'Dark Tower' movie, demonstrating the practical application of text and image weights.
- 📈 **Iterative Process**: The process of generating images with Midjourney often requires trial and error, adjusting weights, and using reference images to achieve the desired composition and style.
Q & A
What are the two powerful techniques discussed in the video?
-The two powerful techniques discussed in the video are Image Weight and Text Weight, which are used in mid-journey prompting to control the output of generated images.
How does the mid-journey system assign tokens to keywords in a prompt?
-Mid-journey scans the prompt for keywords and assigns tokens to those keywords. The tokens are then used to assemble the image based on what mid-journey knows in its database. It is suggested that about 75 tokens are assigned per prompt, with more emphasis on the words at the start of the prompt.
How can text weights be used to influence the importance of a keyword in a prompt?
-Text weights are used by replacing a comma with a double colon followed by a number. This number indicates the weight or importance of the keyword. For example, using '::1' after a keyword would give it a weight of one, while '::2' would double its importance compared to a keyword with a weight of one.
What is the recommended range for text weights?
-The recommended range for text weights is between one and ten to maintain simplicity and avoid excessive calculations.
How does the format of text weights affect the mid-journey output?
-The format of text weights is crucial. There should be no space between the keyword and the double colon, but there should be a space after the double colon and the weighted number. If the format is incorrect, mid-journey may ignore the weight.
What is the role of image weights in mid-journey?
-Image weights are used with reference images in mid-journey. By using '--IW' followed by a number between 0.5 and 2, you can indicate how much reliance is placed on the reference image for the final output. This helps in generating images inspired by the reference image.
How does the image weight affect the style of the generated image?
-The image weight determines how much influence the reference image has on the final output. A higher image weight makes the generated image closer to the style of the reference image, while a lower weight results in a more independent style inspired by the reference.
What is the purpose of negative prompting in mid-journey?
-Negative prompting is used to remove or avoid certain elements in the generated image. This is done by adding a keyword followed by a double colon and a negative number, which tells mid-journey to omit those elements from the output.
What is the challenge when using negative prompts with reference images?
-The challenge with negative prompts is that mid-journey tends to be 'tricky' and may not fully comply with the request to remove certain elements, especially if those elements are present in the reference image. It may require alternative methods, such as photobashing, to achieve the desired outcome.
How can photobashing be used to refine mid-journey outputs?
-Photobashing involves manually editing the generated image to achieve a more desired result. For example, if mid-journey struggles to generate an image without a specific element present in the reference image, one can edit the image to remove that element and then use the edited image as a new reference for further prompts.
What is the importance of experimenting with different weights and prompts in mid-journey?
-Experimenting with different weights and prompts is crucial for fine-tuning the output to match the desired image. It allows for greater control over the compositional balance and style of the generated images, leading to more accurate and creative results.
Outlines
🖌️ Understanding Prompt Weighting in Mid-Journey
This paragraph discusses the concept of prompt weighting in Mid-Journey, an AI art generation platform. It explains how the system assigns tokens to keywords in a prompt, with more emphasis on words at the beginning of the prompt. The speaker uses examples to illustrate how adjusting the order of keywords can change the resulting image. They introduce text weights, which allow users to assign more tokens to specific keywords, thus influencing the importance of those elements in the final image. The paragraph also touches on the limitations and workarounds when dealing with prompt weights.
🎨 Utilizing Image Weights and Reference Images
The second paragraph delves into the use of image weights and reference images in Mid-Journey. It explains how users can upload a reference image and assign it an image weight using the '---IW' command, which influences how much the AI relies on the reference image for the final output. The speaker provides examples of how varying the image weight can lead to different interpretations and styles in the generated images. They also discuss the use of negative prompts to remove certain elements from the image, although they note the challenges in doing so effectively.
🌟 Experimenting with Negative Prompts and Photobashing
In the final paragraph, the speaker shares their experiences with negative prompting and photobashing in Mid-Journey. They describe how attempting to remove elements like a hat from an image can be tricky, as the AI tends to retain certain features even when instructed not to. The speaker then explores the use of photobashing, where they combine elements from different images to achieve a desired result. They demonstrate this by overlaying Clint Eastwood's face onto a body from a reference image of Harrison Ford, and discuss the iterative process of refining the prompt to achieve a satisfactory outcome.
Mindmap
Keywords
💡Midjourney
💡Text Weight
💡Image Weight
💡Prompt
💡Tokens
💡Compositional Balance
💡Reference Image
💡Natural Language
💡Negative Prompt
💡Photobashing
💡Illustrator Style
Highlights
Introduction to two powerful techniques in mid-journey prompting: Image Weight and Text Weight.
Explanation of how prompts work in mid-journey by assigning tokens to keywords.
The importance of the order of keywords in a prompt, with more emphasis on earlier words.
Text weights allow users to assign more tokens to specific keywords by using a colon colon and a number.
Image of a cupcake created with text weights to emphasize different aspects of the prompt.
Demonstration of how image weights can affect the compositional balance in a longer prompt.
Using reference images in mid-journey by uploading an image and using it as a style reference with an image weight.
Experimenting with image weights to achieve a closer resemblance to a desired style or pose.
The difference between image referencing and stable diffusion's posed image in mid-journey.
Using negative prompts to remove or avoid certain elements in the generated image.
Challenges and workarounds when trying to remove specific elements like a hat from an image using negative prompts.
Combining photobashing with negative prompts to achieve desired results not possible through text alone.
The effectiveness of photobashing in refining the generated image to better fit the user's vision.
An example of using mid-journey for a fictional documentary project about a Dark Tower movie.
The process of experimenting with different weights and styles to achieve the best image composition.
The importance of playing around with weights to find the right balance for a complex prompt.
Encouragement for viewers to like and subscribe to the channel for more content on mid-journey techniques.
A reminder that fine-tuning prompts and using a combination of techniques can lead to more accurate and desired outputs.