一些用于角色设计的提示词。"Last Ride, 2042"作品。使用 Midjourney 8.2 和 Chat GPT Sol ASTRA 制作。 PROMPT 1（在 MJ 8.2 上进...

Midjourney yagi_dsn Sat Sep 12 16:13:35 +0000 2026

Some prompts for character design. "Last Ride, 2042" piece. Made in Midjourney 8.2 and Chat GPT Sol ASTRA.

PROMPT 1 (character creation on MJ 8.2): ''Full length luxury fashion editorial photograph. Slender adult male, early fifties, long neck, elongated face, defined cheekbones, pale skin, near black dark brown irises. Quiet androgynous elegance. Long opaque white latex trench coat: minimalist tailoring, structured shoulders, narrow collar, concealed fastening, straight silhouette hemming just above ankles; realistic latex thickness, restrained glossy highlights, natural folds. Tonal white tailored trousers. Fluorescent acid green velvet avant garde loafers, elongated square toes, low sculptural soles, soft directional velvet shading; fabric does not emit light. Matching acid green sculptural braided hair: cropped sides, two thick interwoven braids forming a compact low crest, one small polished silver clasp. Subtle 18th century aristocratic makeup: translucent pale powder, muted rose blush, fine black eyeliner, natural lips. Slight three quarter turn, relaxed arms, empty hands. Neutral gray seamless studio background, soft white lighting. Photorealistic skin, latex, and velvet.''

SPECS: --ar 3:4 --raw --profile jcvanqu

MOODBOARDS: highend fashion, f4shi0n777.

PROMPT 2 (character sheet creation on ASTRA): ''This is the final character. There may be slight differences from the initial description, but this is the chosen aesthetic and exact design. Your task: create one (or more) character sheets on a white background featuring various poses and views (front, side, profile, back, walking, neutral, seated, detail). Do not include any typography or text. For the seated pose, you may use a hyper-minimalist black chair or—preferably—no additional elements at all. Keep the design, clothing, textures, lighting, etc., identical to the attached image.''

PROMPT 3 (add the character inside the taxi - 9 attached images): ''Perfect. Now, analyze all the images attached to this message and create the final images for the following shots: "02 · No one at the wheel," "03 · The return," "04 · Coming home," "05 · The place we occupy," and "06 · Rest." Use the "street" and "garage" reference images to maintain consistent lighting and color across all elements of the images you create. Preserve the exact aesthetic and design of the character, taxi, etc. For the images depicting the taxi interior, use the attached "taxi interior" images as a reference for lighting and color (aiming for warmer, orange tones) and adjust all image elements accordingly, keeping the design and aesthetic identical—only modifying colors and lighting as necessary.''

The key to the entire image generation process (given my specific workflow) lies in having a clear, structured script broken down into shots. ASTRA acts as an assistant for scriptwriting, prompt optimization, and image editing. Regardless of the specific task, it must have a deep understanding of your project; its aesthetics, script, shots, emotional tone, and so on.

This means that if (say, midway through production or post-production) you realize you are missing a specific type of image or shot, you can simply ask ASTRA. Drawing on the material already created (character sheets, product sheets, environments, the script, etc.), it can generate a multitude of images that fit your project perfectly from just a single prompt.

For my next project, I'm generating a wide range of product sheets for the car I’ll be using (the Koenigsegg Jesko Attack), as well as for the characters, environments, and more. Once you understand your own workflow, everything becomes easier; the tool itself takes a back seat.

一些用于角色设计的提示词。"Last Ride, 2042"作品。使用 Midjourney 8.2 和 Chat GPT Sol ASTRA 制作。 PROMPT 1（在 MJ 8.2 上进行角色创建）：''全长奢侈时尚编辑级照片。身材修长的成年男性，五十岁出头，长颈，瘦长脸型，颧骨分明，肤色苍白，虹膜近黑的深棕色。安静的雌雄同体式优雅。长款不透明白色乳胶风衣：极简剪裁，结构化肩部，窄领，隐藏式系合，直筒廓形下摆刚好在脚踝以上；真实的乳胶厚度，克制的光泽高光，自然褶皱。同色系白色定制长裤。荧光酸绿色天鹅绒先锋乐福鞋，纤长方头，低雕塑感鞋底，柔和定向的天鹅绒阴影；面料本身不发光。同色系酸绿色雕塑感编织发型：两侧铲青，两条粗辫相互交织形成紧凑的低脊状造型，一个小巧的抛光银色搭扣。微妙的 18 世纪贵族妆容：半透明...

一键生图查看来源

中文说明

一些用于角色设计的提示词。"Last Ride, 2042"作品。使用 Midjourney 8.2 和 Chat GPT Sol ASTRA 制作。 PROMPT 1（在 MJ 8.2 上进行角色创建）：''全长奢侈时尚编辑级照片。身材修长的成年男性，五十岁出头，长颈，瘦长脸型，颧骨分明，肤色苍白，虹膜近黑的深棕色。安静的雌雄同体式优雅。长款不透明白色乳胶风衣：极简剪裁，结构化肩部，窄领，隐藏式系合，直筒廓形下摆刚好在脚踝以上；真实的乳胶厚度，克制的光泽高光，自然褶皱。同色系白色定制长裤。荧光酸绿色天鹅绒先锋乐福鞋，纤长方头，低雕塑感鞋底，柔和定向的天鹅绒阴影；面料本身不发光。同色系酸绿色雕塑感编织发型：两侧铲青，两条粗辫相互交织形成紧凑的低脊状造型，一个小巧的抛光银色搭扣。微妙的 18 世纪贵族妆容：半透明...

原始 Prompt

Some prompts for character design. "Last Ride, 2042" piece. Made in Midjourney 8.2 and Chat GPT Sol ASTRA.

PROMPT 1 (character creation on MJ 8.2): ''Full length luxury fashion editorial photograph. Slender adult male, early fifties, long neck, elongated face, defined cheekbones, pale skin, near black dark brown irises. Quiet androgynous elegance. Long opaque white latex trench coat: minimalist tailoring, structured shoulders, narrow collar, concealed fastening, straight silhouette hemming just above ankles; realistic latex thickness, restrained glossy highlights, natural folds. Tonal white tailored trousers. Fluorescent acid green velvet avant garde loafers, elongated square toes, low sculptural soles, soft directional velvet shading; fabric does not emit light. Matching acid green sculptural braided hair: cropped sides, two thick interwoven braids forming a compact low crest, one small polished silver clasp. Subtle 18th century aristocratic makeup: translucent pale powder, muted rose blush, fine black eyeliner, natural lips. Slight three quarter turn, relaxed arms, empty hands. Neutral gray seamless studio background, soft white lighting. Photorealistic skin, latex, and velvet.''

SPECS: --ar 3:4 --raw --profile jcvanqu

MOODBOARDS: highend fashion, f4shi0n777.

PROMPT 2 (character sheet creation on ASTRA): ''This is the final character. There may be slight differences from the initial description, but this is the chosen aesthetic and exact design. Your task: create one (or more) character sheets on a white background featuring various poses and views (front, side, profile, back, walking, neutral, seated, detail). Do not include any typography or text. For the seated pose, you may use a hyper-minimalist black chair or—preferably—no additional elements at all. Keep the design, clothing, textures, lighting, etc., identical to the attached image.''

PROMPT 3 (add the character inside the taxi - 9 attached images): ''Perfect. Now, analyze all the images attached to this message and create the final images for the following shots: "02 · No one at the wheel," "03 · The return," "04 · Coming home," "05 · The place we occupy," and "06 · Rest." Use the "street" and "garage" reference images to maintain consistent lighting and color across all elements of the images you create. Preserve the exact aesthetic and design of the character, taxi, etc. For the images depicting the taxi interior, use the attached "taxi interior" images as a reference for lighting and color (aiming for warmer, orange tones) and adjust all image elements accordingly, keeping the design and aesthetic identical—only modifying colors and lighting as necessary.''

The key to the entire image generation process (given my specific workflow) lies in having a clear, structured script broken down into shots. ASTRA acts as an assistant for scriptwriting, prompt optimization, and image editing. Regardless of the specific task, it must have a deep understanding of your project; its aesthetics, script, shots, emotional tone, and so on.

This means that if (say, midway through production or post-production) you realize you are missing a specific type of image or shot, you can simply ask ASTRA. Drawing on the material already created (character sheets, product sheets, environments, the script, etc.), it can generate a multitude of images that fit your project perfectly from just a single prompt.

For my next project, I'm generating a wide range of product sheets for the car I’ll be using (the Koenigsegg Jesko Attack), as well as for the characters, environments, and more. Once you understand your own workflow, everything becomes easier; the tool itself takes a back seat.

全身构图多分屏长发妆容影棚角色时尚人物人物肖像