The Darkroom

The Philosophy

“Human Eye, Machine Hand.”

We do not rely on build-time plugins to guess how an image should look. Art direction (color, crop, tone) is a human task performed in professional tools (Lightroom). Optimization (compression, formatting, resizing) is a machine task performed by scripts. The two never overlap.

The Pipeline

1. Ingest (The Darkroom)

Raw camera files (CR2, NEF) are ingested into a local workspace. We use Adobe Lightroom Classic for culling and development.

2. Master Export

We export high-resolution TIFFs (4000px, sRGB) to a “Master” directory. This is the Source of Truth.

• Naming Convention: {slug}-{view}-{sequence}.tif
• Example: xbox-detail-01.tif

3. The Machine (Python)

Our custom processing engine (scripts/process_images.py) takes over:

• Resizing: Lanczos resampling to xl (1920px), lg (1280px), md (800px), and sm (500px).
• Format Conversion: Generates next-gen AVIF (Primary) and WebP (Fallback).
• Animation Stitching: Detects image sequences and compiles them into high-quality Animated WebP, replacing legacy GIFs.

Case Study: Project “Base”

We used the base project as a testbed for this pipeline.

The Challenge

The original project relied on heavy unoptimized GIFs and PNGs, resulting in slow load times and poor LCP (Largest Contentful Paint).

The Transformation

We purged the legacy assets and ran the Image Pipeline.

Comparison

Metric	Before (Legacy)	After (System)
Format	GIF / PNG	AVIF / WebP
Payload	~15 MB	1.2 MB (92% Reduction)
Responsive	No (Single file)	Yes (XL, LG, MD, SM)
Animation	Dithered GIF (256 colors)	Animated WebP (Full RGB)

Future Roadmap

• HDR Support: Leveraging AVIF’s HDR capabilities for “True Black” OLED displays.
• 3D Pipeline: Automating .glb compression and texture optimization.