AI Image Prompt for Photorealistic Smartphone Selfies: Mastering 24–30mm Wide-Angle Distortion and Skin Perspiration

While modern AI image generation models effortlessly produce high-resolution portraits, recreating the authentic, candid snapshot texture of an everyday smartphone photo remains an elusive challenge. Most AI-generated portraits suffer from telltale digital artifacts: unnaturally airbrushed porcelain skin, idealized facial symmetry, and sterile studio illumination that immediately betray their synthetic origin. On September 10, 2026, AI creator 硅基废话 (@DDJCXX) on X published a sophisticated, structured prompt system designed to systematically eliminate these artificial cues and generate hyper-realistic bedroom selfie portraits that genuinely feel like unedited mobile camera snapshots.

Image source: 硅基废话 (@DDJCXX) on X

Rather than relying on empty quality buzzwords like 'photorealistic' or 'masterpiece', this structured prompt approaches image generation with the precision of a camera director. By meticulously specifying the physical optical characteristics of smartphone lenses, authentic human body proportions under perspective foreshortening, and the delicate specular highlights created by natural perspiration and ambient morning lighting, this technique offers a reusable framework for anyone seeking organic realism in modern diffusion models.

Wide-Angle Perspective Distortion and Anatomical Proportions

A primary giveaway in synthetic portraits is the fundamental mismatch between the prompt premise and the lens perspective chosen by the AI model. When instructed to generate a portrait, most text-to-image engines default to the compressed, flattering perspective of a 50mm to 85mm portrait telephoto lens. However, a genuine handheld selfie taken from an arm's length on a bed is captured exclusively through the wide-angle field of view characteristic of mobile front cameras, typically between 24mm and 30mm equivalent focal length.

The prompt from @DDJCXX solves this discrepancy by defining the camera optics and spatial geometry with exacting precision. It mandates a 24–30mm equivalent wide-angle front lens operating at an ultra-close distance of 35–60cm, positioned directly above the subject's face in a downward high-angle orientation.

• Dramatic Fore-to-Aft Perspective: Objects closer to the wide-angle lens appear significantly enlarged. The subject's face, neck, clavicle, and upper chest occupy the vast majority of the foreground frame, while the lower torso, hips, and legs recede rapidly toward the bottom-left corner and remain partially cropped out.
• Explicit Framing Constraints: The prompt strictly forbids full-body shots or pulled-back camera framing. This enforces an intimate physical proximity that anchors the viewer inside the personal space of the bedroom snapshot.
• Organic Anatomical Proportions: While the optical foreshortening naturally accentuates the upper torso volume, the prompt carefully balances this distortion by requiring a narrow shoulder frame and a realistic hourglass silhouette. This prevents the model from generating distorted spherical exaggerations or synthetic paper-doll proportions.

Crucially, the prompt specifies that only a portion of the subject's outstretched forearm enters the lower-right edge of the frame to imply the act of holding a phone, while explicitly instructing that the smartphone device itself need not be rendered. This avoids the awkward geometric deformities and floating-phone anomalies common in AI generation while maintaining complete narrative credibility.

Perspiration Highlights and Authentic Organic Skin Texture

Another major obstacle in photorealistic AI portraits is the lifeless smoothness produced by diffusion denoising algorithms. In standard generation pipelines, high-frequency details such as skin pores, fine peach fuzz, and natural micro-wrinkles are routinely smoothed out into an artificial plastic sheen.

To counter this plastic effect, the author injects genuine biological texture into the scene through physical warmth and perspiration.

• Micro-Droplets and Specular Highlights: The prompt specifies visible beads of fine perspiration and natural moisture across the forehead, bridge of the nose, neck, clavicle hollows, and upper chest. These moisture droplets act as microscopic prisms, creating authentic specular highlights that scatter light irregularly across the skin surface and restore visceral depth.
• Natural Ambient Illumination: Instead of artificial studio rim lighting, the scene relies on soft morning daylight diffused through a bedroom window, complemented by subtle reflections bouncing off the rumpled bedding. A faint cool catchlight emanating from the smartphone screen glints in the subject's pupils and across the bridge of the nose.
• Candid Facial Expression and Wet Hair Strands: The prompt rejects staged, smiling expressions in favor of half-lidded, languid eyes looking into the lens with subtle morning fatigue, accompanied by softly parted lips showing just a hint of upper teeth. Dark straight hair cascades naturally across the pillow and bedsheets, with individual damp strands clinging realistically to the skin along the neck and collarbone.
• Sensor Texture and Depth of Field: Incorporating natural mobile sensor noise and a soft falloff in focus outside the primary focal plane ensures that the image mimics the hardware limitations of a real smartphone camera sensor rather than an immaculate 3D render.

Full Structured Prompt and Practical Implementation Guide

Below is the complete original structured prompt shared by AI creator 硅基废话 (@DDJCXX). It can be directly loaded into modern diffusion models such as FLUX.1, Midjourney v6, or Stable Diffusion, or adapted into English modular tags for fine-tuned workflows.

Prompt：超写实近距离卧室床上自拍摄影，竖版约 4:5，一名年轻成年东亚女性，明确成年，半躺半侧卧在凌乱床铺上，从身体斜上方用手机近距离自拍。采用约 24–30mm 等效前置/手机广角镜头，摄影距离非常近，约 35–60cm，镜头位于人物脸部右上方并向下俯拍，画面带明显近大远小透视：脸部、肩颈、胸口和靠近镜头的上半身占据大面积前景，下半身只保留部分腰胯和大腿。不要全身，不要拉远。

人物为年轻成年东亚女性，中小骨架，肩部偏窄，胸廓相对窄，上半身明显丰满，胸部体积自然偏大，腰腹较纤细，胯部和大腿有柔软真实体积。身材必须呈自然丰满沙漏比例，不要生成高挑纸片人、窄胯细腿，也不要健美肌肉型。近距离广角会让胸肩略微放大，但人体结构必须真实，不要夸张成不自然球形。

人物身体斜躺在床上，头部位于画面右上，骨盆和大腿延伸向画面左下。上半身微微朝镜头抬起，右肩更靠近镜头，左肩略低。躯干有轻微侧转，形成自然胸腰曲线。下半身侧躺，左侧大腿从画面左下进入，另一条腿被身体和床铺部分遮挡。

人物一只手臂从画面右下向前伸出，像正在举手机自拍，手臂只保留部分前臂和上臂，不需要手机进入画面。另一只手自然放在下腹/裙摆附近，手指轻轻抓住或压住睡裙布料，姿态随意，不做比心、剪刀手等动作。

人物留乌黑色中长直发，长度到胸口以下，发量自然偏多。头发因为侧躺而向人物右侧和床面散开，部分发丝贴在脸颊、颈部、锁骨和胸口。额前是长碎刘海/偏分刘海，多束细发垂过额头和眼睛附近，略显凌乱、潮湿，不要整齐齐刘海，不要高马尾或大卷发。

脸型小巧偏鹅蛋形，下颌柔和，肤色白皙偏冷粉。人物双眼略微半睁，眼神看向镜头，眼皮有轻微疲倦感，表情慵懒、发热、像刚睡醒。嘴巴自然微张，露出少量上牙，不笑。嘴唇为浅粉、珊瑚粉色，表面明显湿润有光泽。

眼妆偏自然但有清楚睫毛，眼线细，眼尾轻微上扬；眼周有淡淡粉红色。鼻梁小巧，鼻尖和鼻梁有明显高光。不要浓重欧美妆，不要过度修容。

人物皮肤呈非常明显的潮湿/出汗/微汗质感，额头、鼻尖、人中、颈部、锁骨凹陷处以及胸口皮肤表面有清晰可见的细小汗珠和水光反光，汗水顺着锁骨微弱滑落，形成真实自然的高光漫反射。彻底消除AI平滑塑料感，保留清晰真实的皮肤毛孔、细微绒毛和自然肌理。

身穿一件极简贴身的浅色细吊带真丝睡裙，面料由于汗水和体温轻微贴合肌肤，形成柔软自然的织物褶皱与凹凸阴影。

拍摄环境为清晨室内卧室，背景是略显凌乱的纯棉白色床单与松软枕头。主光源来自侧面窗户透射进来的柔和晨曦自然光，伴随室内空气中的漫反射，手机屏幕微弱的冷光在人物鼻梁与瞳孔中形成点状眼神光。画面整体呈现轻微的高感光度噪点与柔和的前置镜头色散，景深浅而自然，完美还原日常移动端快照的真实胶片感。

The underlying architectural lesson of this workflow is the importance of conceptual layering: establishing camera optics and physical distance first, governing skeletal anatomy and posture second, detailing surface micro-textures and lighting third, and finally anchoring the atmospheric environment. By ordering directives according to physical reality, creators can prevent prompt bleeding and guide AI models to produce exceptionally convincing photorealistic imagery.

Original source

The photorealistic smartphone selfie prompt structure and optical control methodology documented here were originally researched and published by AI creator 硅基废话 (@DDJCXX) on X on September 10, 2026.

• 硅基废话 (@DDJCXX) on X: Ultra-realistic Smartphone Selfie Prompt