SREF RANDOM released for Midjourney. 25 example styles.

Thaeyne
29 Apr 202417:51

TLDRThe video script discusses the 'Midjourney' style reference feature, with a focus on the newly released 'sref random' option. The host, Th, shares their enthusiasm for discovering new styles and explains how to use the feature by adding 'D-sref random' to a prompt, resulting in a different style each time. They also explore different style versions (SV1-SV4), noting that SV4 is the current default. Th encountered an issue with seeds not working as expected with 'style random' and discovered that a style seed number locks both the seed and style, leading to the same image. They suggest using a number for 'dh-S ref' for control over random testing. The video presents 25 unique styles found using 'sref random,' each with a brief description of its visual characteristics, such as color schemes, lighting, and artistic techniques. Th concludes by expressing the preference for exploring the impact of words on prompts rather than using 'srf' for sharing styles due to its tedious nature.

Takeaways

  • 🎨 The introduction of the 'sref random' feature by Midjourney allows for the discovery of new and varied styles through randomization in prompts.
  • 🔗 To apply a preferred style to new prompts, use the image link with a dash-sref after it, offering control over the style used.
  • 📈 There are four different style reference versions (sv1 to sv4), each producing a slightly different look for the generated images.
  • 🚫 An issue encountered with 'sref random' is that using the same seed number does not change the style with each prompt.
  • 🔢 The unofficial way to reference styles involves using a style seed number, which locks in both the seed and style seed numbers for an image.
  • 🔍 The documentation on Midjourney's website reveals that style seed numbers are directly correlated with seed numbers, offering a vast array of styles to explore.
  • 🌟 The speaker has enjoyed most of the styles discovered using 'sref random' and is considering testing more or sharing personal favorites.
  • 🌈 The video showcases 25 distinct styles discovered with the random feature, each with its unique characteristics and color schemes.
  • ✨ The styles range from vibrant and energetic to monochromatic and gothic, reflecting a wide spectrum of artistic possibilities.
  • 📚 The speaker plans to look more into the differences between the style versions and how they affect the final output.
  • ✍️ The process of sharing styles using image links is considered tedious, leading to a preference for exploring the impacts of words on prompts instead.

Q & A

  • What is the latest feature discussed in the video script?

    -The latest feature discussed is the 'sref random' feature in Midjourney, which allows for the discovery of new styles by adding 'D-sref random' to the prompt.

  • How can you apply a specific style to new prompts?

    -To apply a specific style, you can specify the link to the image after the dash-sref in your prompt.

  • What are the four options for style reference versions?

    -The four options are sv1 (original style reference algorithm), sv2 (the version that came after sv1), sv3 (more 'viby' than sv2), and sv4 (the current default version with improved performance).

  • What issue was encountered when using the style random feature with seeds?

    -The issue was that the seeds did not work with style random as expected. Using the same seed number with style random resulted in images that looked basically the same, indicating that the style did not change with each prompt.

  • How does the unofficial way of referencing styles work?

    -The unofficial way involves specifying a seed number for 'dh-sref random', which locks in both the seed number and style seed number, resulting in the same image each time with the same values for both.

  • What is the current number of styles to discover for version 6.0?

    -There are 4,294 million,967, 25 styles to discover for version 6.0.

  • What is the main focus of the video script?

    -The main focus is on exploring and sharing different styles discovered with the 'sref random' feature in Midjourney, and discussing the process and findings.

  • What is the significance of the 'sref random' feature for AI art enthusiasts?

    -The 'sref random' feature is significant as it allows AI art enthusiasts to find new styles that they might like and share with others, enhancing the creative process and discovery of unique artistic expressions.

  • How does the video script describe the process of testing out the style random feature?

    -The process involves using a specific seed (like 777) and observing the changes in style with each prompt. The script also discusses the discovery of using a style seed number instead of a traditional seed number for more control over the style randomization.

  • What are some of the visual characteristics mentioned for the styles discovered?

    -The visual characteristics include vibrant colors, fantastical landscapes, celestial wonders, rich and intense hues, ethereal quality, modern interpretation, texture and depth, neon hues, hyperrealism, intricate details, monochromatic palette, and strategic use of light and shadow.

  • What is the creator's approach to sharing styles with the audience?

    -The creator shares styles by generating example images and discussing the process of finding and applying these styles. They also consider the audience's preference on whether to test out as many styles as possible or to share personal favorites.

Outlines

00:00

🎨 Mid Journey Style Reference and Discovery

The speaker introduces themselves as 'th' and dives into discussing the Mid Journey style reference feature. They mention recent updates, highlighting the 'sref random' feature as a favorite for discovering new styles. The process of using 'D- sref random' in prompts to generate varied styles is explained, along with the method of applying preferred styles to new prompts using image links. Four different style versions are introduced, each offering a distinct visual outcome. The speaker discusses an issue with using seeds with 'style random' and shares an unofficial workaround involving style seed numbers. They also mention the vast number of styles available in version 6.0 and express their excitement about exploring and sharing these styles with the audience.

05:01

🌈 Vibrant and Ethereal Styles in Mid Journey

This paragraph showcases a variety of styles discovered using the Mid Journey feature. Each style is characterized by its use of color, light, and shadow to create unique atmospheres and moods. The styles range from rich and intense color palettes that evoke fantasy and surrealism to monochromatic palettes with strategic lighting that adds depth and complexity. The techniques described include a blend of digital and traditional methods, hyperrealism, and classical aesthetics. The speaker's intent to explore and share more styles is evident, as they invite audience interaction regarding their preferences for future content.

10:02

🖌️ Artistic Techniques and Color Palettes in Mid Journey

The paragraph delves into the artistic techniques and color palettes used in Mid Journey, emphasizing the dramatic effects created by the interplay of light and shadow. It describes styles that range from Gothic ambiance with a restrained color palette to vibrant and eclectic color fusions that stimulate the senses. The techniques mentioned include chiaroscuro, impressionism, and a blend of digital painting and photo manipulation. Each style is designed to evoke a sense of wonder, depth, and narrative, with a focus on creating immersive and evocative visual experiences.

15:02

📚 Style Exploration and Audience Engagement

The final paragraph reflects on the process of exploring styles and the challenges of sharing them via image links. The speaker contrasts their experience with Mid Journey version 5.2, where they recall many dark and less useful styles, with version 6, which has yielded many cool styles. They express gratitude to the audience for watching and encourage continued exploration and prompting, indicating an ongoing dialogue and shared enthusiasm for the creative process.

Mindmap

Keywords

💡Midjourney

Midjourney refers to a software or tool used for creating AI-generated art. In the context of the video, it is the platform where the artist is exploring and experimenting with different style features to generate unique visual outputs. The artist discusses updates and new features, such as 'sref random,' which are integral to the creative process within Midjourney.

💡Style Reference (sref)

Style Reference, abbreviated as 'sref,' is a feature within Midjourney that allows users to apply specific styles to their art prompts. It is a core concept in the video as the artist demonstrates how to use 'sref' to create a variety of artistic styles, emphasizing its importance in personalizing the creative output.

💡Random Feature

The 'random feature' is a function within the Midjourney tool that generates different styles each time a prompt is entered, leading to a diverse range of artistic outcomes. It is highlighted in the video as one of the artist's favorite new features for discovering novel styles and enhancing the creative process.

💡Seed Number

A 'seed number' is a specific value used in conjunction with the 'random feature' to produce a particular style or outcome. The artist discusses an issue with the seed numbers not working as expected with the 'style random' feature and explores alternative methods to achieve desired styles.

💡Style Seed Number

A 'style seed number' is similar to a seed number but specifically for the style reference feature. The video explains that the artist discovered a way to control the randomness of styles by using a style seed number, which locks in both the seed number and the style, leading to consistent results.

💡SV (Style Version)

SV refers to different versions of the style reference algorithm within Midjourney. The artist outlines four versions (SV1 to SV4), each producing a slightly different look for the generated images. The discussion of SV is important for understanding the nuances in the style variations that can be achieved.

💡Viby

The term 'viby' is used informally to describe a particular style that is more vibrant or lively. In the context of the video, SV3 is described as being 'more viby' than SV2, indicating a more dynamic and energetic visual style.

💡Ethereal

Ethereal is a term used to describe a style or quality that is dreamlike, otherworldly, or not of this world. It is a recurring theme in the video, with many of the styles described as having an ethereal quality, emphasizing soft, muted colors, and a sense of wonder.

💡Futurism

Futurism is an artistic and social movement that emphasizes themes of technology, speed, youth, and objects such as cars, airplanes, and industrial cities of the future. In the video, the artist mentions styles that incorporate elements reminiscent of futurism, suggesting a forward-looking, innovative approach to art.

💡Expressionism

Expressionism is an artistic movement characterized by the use of bold, distorted imagery to express emotional intensity. The artist in the video refers to a style that is reminiscent of expressionism, where the focus is on conveying emotion through aggressive brushwork and dynamic lines.

💡Impressionism

Impressionism is an art movement that emphasizes the sensory effect of natural light and color. The video discusses a style that is reminiscent of impressionism with a modern twist, highlighting the use of visible brush strokes, fluidity, and abstraction to create a dreamlike quality.

Highlights

Midjourney has released a new feature called 'sref random' which allows for random style generation in prompts.

The 'sref random' feature is a favorite among updates as it facilitates the discovery of new styles.

By adding 'D-sref random' to a prompt, each generated image will have a completely different style.

If a preferred style is found, it can be officially copied by specifying the image link after the '-sref'.

There are four style reference versions (SV1 to SV4), each offering a slightly different visual outcome.

SV1 is the original style reference algorithm, while SV2, SV3, and SV4 represent subsequent updates.

SV3 is described as 'more viby' than SV2, although the exact meaning is unclear.

SV4 is the current default version with improved performance.

An unofficial way to reference styles involves using a dash seed in conjunction with 'sref random'.

The style seed number seems to be directly correlated with the actual seed numbers, expanding the potential styles.

Combining style numbers directly (e.g., '--sr12') is currently not possible in the system.

Image links can be mashed together to combine styles, offering more control over random testing.

The video showcases 25 unique styles discovered using the 'sref random' feature.

Different 'sref' versions were tested with the same prompt to observe variations in style.

The speaker contemplates testing all possible styles due to the vast number available in version 6.0.

The speaker finds the styles from version 6.0 to be more appealing and diverse compared to previous versions.

The video concludes with a discussion on the practicality of using image links versus numbers for style referencing.