I wanted a Korean wedding invitation that felt like a Wong Kar-wai film. Deep reds, desaturated tones, that melancholic warmth of In the Mood for Love. The photos were generic stock images. They needed to feel like 1960s Hong Kong.

The solution wasn’t Photoshop. It was Cloudinary URL parameters.

The Prompt

Apply Wong Kar-wai cinematic color grading to photos using Cloudinary transforms.

With a code example:

<CldImage
  src="wedding/parallax/hk-skyline"
  effects={[
    { colorize: '20,rgb:8B0000' },  // deep red overlay
    { saturation: -10 },             // desaturated
    { contrast: 15 }                 // higher contrast
  ]}
/>

Why This Worked

1. Reference Point Was Specific

“Wong Kar-wai” isn’t a vague aesthetic request. It’s a specific visual language:

• Saturated reds and greens
• High contrast shadows
• Warm, amber-tinted highlights
• Desaturated midtones
• Film grain texture

Anyone who’s seen Chungking Express or 2046 knows this look. The reference did the specification work.

2. Implementation Path Was Clear

Cloudinary effects map directly to color grading:

• colorize → color overlay
• saturation → saturation adjustment
• contrast → contrast curve

The prompt connected the aesthetic goal to the implementation tool.

3. Code Example Seeded the Pattern

Starting with a working example, even partial, gives the AI something to build on. The example showed:

• Cloudinary’s effect syntax
• The general approach (layer multiple effects)
• That code-level integration was expected

The Implementation

Different layers got different treatments to create depth:

const wkwGrade = {
  // Background: heavy red, dreamy
  background: [
    { colorize: '20,rgb:8B0000' },  // 20% deep red overlay
    { saturation: -10 },
    { contrast: 15 }
  ],
  
  // Midground: moody green, cinematic
  midground: [
    { colorize: '15,rgb:2D4A3E' },  // forest green
    { saturation: 10 },
    { contrast: 20 }
  ],
  
  // Foreground: warm amber, intimate
  foreground: [
    { colorize: '10,rgb:D2691E' },  // amber/chocolate
    { saturation: 5 },
    { contrast: 25 }
  ]
};

Each layer gets progressively less colorization and more contrast, creating a focal depth effect.

Why CDN Processing Matters

This approach has advantages over local image editing:

The color grading is defined in code. Change the parameters, and every image updates instantly.

The Color Choices

Wong Kar-wai’s cinematographer Christopher Doyle used specific color palettes:

Deep Red (rgb:8B0000):

• Dark, romantic, intimate
• Think: In the Mood for Love stairwell scenes
• Used for backgrounds to create warmth

Moody Green (rgb:2D4A3E):

• Urban, lonely, neon-adjacent
• Think: Chungking Express convenience store
• Used for midground depth

Warm Amber (rgb:D2691E):

• Golden hour, nostalgic
• Think: 2046 hotel corridors
• Used for foreground intimacy

Usage Pattern

For any image that needs the treatment:

<CldImage
  src={imageId}
  effects={wkwGrade.background}
  width={1920}
  height={1080}
  alt="Hong Kong skyline with WKW grading"
/>

The effect array is applied server-side at Cloudinary’s CDN. The client receives a pre-graded image at the requested dimensions.

When This Approach Works

CDN-side color grading makes sense when:

• You have many images needing consistent treatment
• Responsive sizing is required
• The effect is purely color/contrast (not masking or compositing)
• You want to change the look without re-processing files

It doesn’t work for:

• Complex retouching (skin, blemishes)
• Layer masking or selective adjustments
• Effects requiring image-specific tuning

The Takeaway

The prompt worked because it connected a specific aesthetic reference (Wong Kar-wai) to a specific implementation tool (Cloudinary effects) with a working code example.

Generic stock photos now feel like moody Hong Kong cinema. No local image editing, no asset pipeline complexity. Just URL parameters that turn colorize: '20,rgb:8B0000' into nostalgia.

Sometimes the best Photoshop alternative is not Photoshop at all.