Meem #63

Documento original

GPT Image 2 · en

{
  "master_prompt": {
    "global_settings": {
      "resolution": "8K ultra-high-definition",
      "aspect_ratio": "3:4 vertical",
      "style": "hyper-realistic AI-edited commercial product photography",
      "sharpness": "extreme clarity, micro-detail visibility",
      "lighting_quality": "cinematic studio lighting with controlled highlights and shadows",
      "motion_freeze": "high-speed capture, frozen splashes and particles",
      "noise": "none",
      "artifacts": "none"
    },

"module_1_image_1_style": {
      "subject": {
        "type": "orange-flavored soda can",
        "color": "bright orange",
        "surface_details": "condensation droplets across can",
        "branding_text_visible": [
          "Sesla",
          "cola"
        ]
      },
      "pose_and_orientation": {
        "position": "upright, slightly embedded in splash",
        "angle": "centered frontal view"
      },
      "liquid_and_motion": {
        "liquid_color": "transparent orange-tinted liquid",
        "motion": "high-energy splash bursting upward and outward",
        "splash_detail": "clear liquid arcs with suspended droplets"
      },
      "floating_elements": {
        "oranges": [
          "whole oranges",
          "halved oranges with visible pulp"
        ],
        "ice_cubes": "clear sharp-edged cubes floating mid-air",
        "droplets": "numerous spherical liquid droplets"
      },
      "background": {
        "color": "solid bright orange",
        "lighting": "high-key, evenly lit"
      },
      "surface_and_reflection": {
        "base": "shallow liquid pool",
        "reflection_quality": "strong mirror-like reflection with ripples"
      }
    }
  }
}

Cómo usar

1. Copia el prompt original completo, incluidos los parámetros.
2. Adapta el tema, el texto y la composición a tus necesidades.
3. Usa una herramienta externa compatible; comprueba el acceso y el resultado.

Antes de empezar

• Comprueba si necesita una imagen de referencia. La compatibilidad no demuestra qué modelo generó el ejemplo.
• Revisa nombres, cifras y textos; el resultado puede variar entre versiones del modelo.