Ultra-Vivid Strawberry Splash: The Exact AI Prompt Revealed

A breakdown of a structured prompt for four color-coded strawberries in a pink liquid crown splash, explaining how explicit lighting ratios, repeated seed-color tokens, and Midjourney-specific parameters control the generated composition.

Image prompt

Hyper-realistic food photography of four strawberries in explosive liquid splash composition, pure black background, one strawberry natural bright red with yellow achenes and fresh green calyx leaves, one strawberry vivid magenta pink with yellow achenes, one strawberry turquoise cyan with yellow achenes, one strawberry deep crimson with yellow achenes, all strawberries suspended in dynamic pink liquid splash with crystalline water droplets, aerated bubbles, and crown splash formation frozen in mid-air, dramatic studio lighting with 45-degree key light from above-left creating specular highlights on wet surfaces, fill light from front-right at 2:1 ratio, sharp macro detail revealing seed texture and water refractions, pink liquid pooling at base with mirror-perfect reflections, high-speed photography freeze-frame at 1/8000s, Hasselblad H6D-100c aesthetic, commercial advertising quality, ultra-detailed 8K render --ar 9:16 --style raw --s 250 --q 2

The primary failure mode in AI-generated splash photography is material collapse: the liquid reads as solid plastic, or the subject's surface texture merges with the surrounding fluid. This prompt addresses that problem by treating color assignment, lighting geometry, and motion as separate clauses rather than stacking descriptive adjectives.

Achene Isolation Across Color Variants

The prompt assigns four distinct body colors to the strawberries: natural bright red, vivid magenta pink, turquoise cyan, and deep crimson. If you only specify the body color, diffusion models tend to apply that pigment uniformly across the entire mesh, including the seeds. The phrase yellow achenes is repeated after every single color assignment to prevent this bleed. Achenes are the small, seed-like structures embedded in the strawberry's receptacle. By naming them explicitly and assigning a contrasting color each time, the prompt forces the model to maintain textural separation between the fruit's skin and its seeds.

If your output shows seeds matching the cyan or magenta body, the model has ignored the token repetition. You may need to increase the weight of the achene clause or move it earlier in the prompt string, depending on the platform's attention weighting.

Lighting Geometry and the 2:1 Ratio

The lighting instructions specify a 45-degree key light from above-left and a fill light from front-right at a 2:1 ratio. In physical studio photography, a 2:1 ratio means the key light delivers twice the illuminance of the fill, producing a shadow on the unlit side that retains visible detail without flattening the subject's volume. Diffusion models do not calculate photon paths or measure actual lux values. Instead, they associate the textual pattern "2:1 ratio" and "45-degree key light" with training images exhibiting that specific contrast falloff.

The result should be a bright specular highlight on the upper-left curve of the wet strawberries and the liquid droplets, with a softer, dimmer reflection on the right side. If the highlights appear large and diffuse rather than tight and sharp, the model is defaulting to softbox lighting. Removing terms like "dramatic" and relying strictly on the angle and ratio numbers sometimes corrects this, though behavior varies by model version.

Liquid Physics and the Black Grounding Plane

The pure black background serves a functional purpose beyond aesthetics. It removes ambient environmental data from the generation, forcing all visible illumination to come from the described studio lights. This maximizes the contrast of the pink liquid splash.

The prompt requests a "crown splash formation" and "pink liquid pooling at base with mirror-perfect reflections." Crown splashes occur when a falling object displaces liquid upward into a hollow cylinder with a rim of droplets. Specifying "frozen in mid-air" alongside "high-speed photography freeze-frame at 1/8000s" targets the visual signature of ultra-fast strobe photography, where individual droplets appear as sharp spheres rather than motion-blurred streaks. The pooling liquid at the base provides a reflective floor plane, anchoring the floating elements so they do not look pasted into a void.

This is a simulation, not a physical photograph. The model does not run fluid dynamics calculations. Expect the detail to soften at the edges of the frame, particularly where the complex intersection of multiple strawberries, bubbles, and refracting droplets occurs. Water refractions inside the droplets—where the black background and pink liquid would theoretically invert through a sphere—are frequently approximated as simple gradients rather than optically accurate distortions.

Platform-Specific Parameters

The prompt ends with four flags: --ar 9:16, --style raw, --s 250, and --q 2. These are specific to Midjourney. They will not function in Stable Diffusion, DALL-E, or other generators, and applying them elsewhere will either cause errors or be interpreted as part of the text prompt.

• --ar 9:16 sets the aspect ratio to vertical, suitable for mobile displays or poster formats.
• --style raw reduces Midjourney's default aesthetic stylization, pushing the output closer to the literal photographic description.
• --s 250 sets the stylize parameter to 250 (the default is typically 100). This introduces slightly more artistic interpretation while remaining relatively constrained.
• --q 2 increases the quality parameter. In older Midjourney versions, this allocated more GPU time per generation. In current versions, its effect is limited or deprecated, but it remains in the prompt structure.

The exact Midjourney version used to generate the reference image is unknown. Parameter behaviors change between releases; --style raw functions differently in v5 than in v6.

Camera Shorthand and Limits

The phrase "Hasselblad H6D-100c aesthetic" acts as a style token. The H6D-100c is a 100-megapixel medium format camera known for high color depth and sharpness. Naming it signals the model to draw from training data associated with high-end commercial product photography. Similarly, "commercial advertising quality" and "ultra-detailed 8K render" act as general polish directives.

Swapping the pink liquid for a different viscosity—such as dark chocolate ganache or honey—is an untested suggestion. Thicker fluids behave differently in real physics, but prompting a diffusion model to simulate higher viscosity requires changing the shape descriptors (e.g., "thick, slow-moving ribbons" instead of "crystalline water droplets") rather than just swapping the material noun. Keeping the 1/8000s shutter speed token while requesting thick honey may produce contradictory visual cues, as high-speed flash photography of viscous fluids yields different geometries than water splashes.