Structured directives: paradigms
Picture Quality Words >>
Generally more fixed: masterpiece, masterpiece, best quality, highly detailed, official art, Tyndall effect, fine CG quality, 8K, large wallpaper, etc.. Usually you start with masterpiece, best quality, and then add HDR, UHD, etc. to improve the picture quality.
Artistic style and medium >>
例如彩绘、铅笔画或艺术摄影,或强调艺术家风格。某些个性模型有固定的”触发指令”。
Photographic modifiers (Polaroid, Tilt-shift, Portrait)
Style modifiers (Product Shot Product Shot, Color-Splash Color Splicing, Monochrome Monochrome)
Lens modifiers (Telephoto telephoto, Fish-eye fisheye, CCTV monitor image, 800mm, Gopro)
Art Modifiers (Pencil Drawing, Graffiti, Water Colors, Oil Painting)
Pictures of the subject >>
The subject of the composition, complex composition, the existence of the subject exists in a certain medium or environment, for example: hanging on the wall in the frame of the old man's photo, we have to confirm the only subject for refinement, or distribution generation.
–描述主体造型>>
Note that both the whole and the details are described from top to bottom; here is an example of describing a character:
Hairstyles (dorky hair, hair behind the ears, bangs that cover the eyes, low double ponytails, big wavy curls), the
Hair color (blonde at the top, colored picks at the ends).
Clothing (long dresses, cheongsams, half-length sleeves, knee-high socks, indoor shoes).
Head (cat ears, red eyes).
Neck (necklace).
arms (strapless).
Abdomen (belly button visible).
Legs (long legs).
Footsteps (barefoot)
–描述主体细节
For example, expressive expressions for characters are supported using emoji, which are quite expressive. Expressions can be achieved by adding emoji or face characters, such as using 😍 to describe an expression or 🖐 to correct a hand drawing.:-)
smiles :-(
displeased ;-)
give sb. a meaningful look :-D
make fun of sb :-P
blab :-C
It's sad. :-O
Surprise. Open your mouth. :-/
skeptical
Describe the composition of the subject
In the case of characters, it's mostly action:
Basic movements (standing, sitting, running, walking, squatting, lying down, kneeling), the
head movements (head tilt, head up, head down).
Hand motions (hands in hair, on chest, hands up).
back movements (stooping, straddling, duck sitting, bowing).
Leg movements (cross-legged stance, second leg, M-shaped open leg, cross-legged, kneeling), the
Compound Moves (Combat Stance, JOJO Stand, Back to Back Stand, Side Flip)
–主体的大背景
Indoors, outdoors, woods, beach, under the stars, in the sun, what's the weather?
–杂项>>
For example, the eyes are depicted in detail
Structured instructions: examples
(masterpiece:1.331), best quality,
illustration.
(1girl).
(deep pink hair:1.331), (wavy hair:1.21),(disheveled hair:1.331), messy hair, long bangs, hairs between eyes,(white hair:1.331), multicolored hair, (white bloomers:1.46), (open clothes), (open clothes), (open clothes), (open clothes), (open clothes), (open clothes)
beautiful detailed eyes,purple|red eyes).
expressionless.
Sitting.
dark background, moonlight, ,flower_petals,city,full_moon.
Structured directives: simplification
Subject: East African canyons, mountains, deserts, weeds, an antelope
Modifier: Best Picture Quality
Style: Studio Ghibli style
Midjourney V6 Command Paradigm
Midjourney V6 Alpha 测试已经启动。要启用 Midjourney V6,它不会根据默认设置自动更新。因此,你需要向 Midjourney Discord 服务器或机器人发送信息,输入 /settings 命令,并从下拉菜单中选择 V6。另外,你也可以在提示语的末尾输入 –v 6 来激活该功能。
Midjourney 的创始人 David Holz 提出了一种新的提示方法。V6 对提示的敏感度大大提高。建议避免使用“获奖、超现实、4k、8k”等表述,而应确保请求明确。如果你希望得到更直接、更中立且更现实的输出结果,最好将 –style raw 设置为默认。(原文中直言不讳地将 4k、8k 等描述称为“无用”提示。”V6 对你的提示非常敏感。避免使用‘无用’的表述,如‘获奖、超现实、4k、8k’”)
key constituent | descriptive | typical example |
---|---|---|
hairstyle | The style of the image. | Photo of |
thematic | Describes the subject or object of the image. | An astronaut in a white space suit, helmet visor reflecting stars. |
contexts | Indicates the background setting of the image, i.e. the location or environment. | Standing on a moon with Earth visible in the starry sky. |
(art) composition | Explain the arrangement of elements within an image. | Astronaut centered, Earth in the background. |
(stage) lighting | Describes illumination or light reflection in an image. | Bright sunlight with soft moonlight reflections. |
Additional information | Provides additional details or information contained in the image. | Moon rocks and small craters nearby. |
In turn, following the above cue structure leads to optimal results. In addition, the most notable features of the system include the ability to render text accurately without blurring, and the ability to finely tune textures and lighting.
Midjourney New Tip Structural Analysis
Midjourney v6 is coming soon and it has a better understanding of prompts. This means no more random phrases and words. We need to adopt a whole new way of creating prompts.
I personally plan to test this simple structure, which reflects the typical language framework:
- Style:
- Purpose: To provide a specific aesthetic or artistic style to an image.
- Details to include: preferred style or era.
- Topics:
- Purpose: To clarify the core focus of the image.
- Details to include: characteristics of the central theme (e.g., people, objects, animals), including appearance, color, and special features.
- Background:
- Purpose: To set the environment or context in which the subject is located.
- Details to include: location (indoor, outdoor, fictional), environmental elements (natural, urban), time of day, and weather conditions.
- Composition:
- Purpose: To determine how the subject and its elements will be presented in the picture.
- Details to include: perspective (close-up, wide angle, overhead), angle, and specific compositional preferences.
- Lights:
- Purpose: To set the mood and visual tone of the image.
- Details to include: lighting type (bright, dim, natural), mood (pleasant, mysterious), and atmospheric effects.
- Additional Information:
- Purpose: To add more layers and depth to the image.
- Details to include: secondary objects, people, animals, and their interrelationships and placement in relation to the subject.
I look forward to testing this structure soon and sharing the results.