LUMA TECHNICAL OVERVIEW

LUMA is an iPhone RAW camera and development system for creators who want maximum image quality without leaving mobile. A modular multi-target design keeps capture, editing, and export consistent across the app, lock-screen camera, and quick-launch surfaces.

LUMA system architecture

LUMA is RAW-first end-to-end and built as a layered system of local feature and core modules. More than twenty modules compose a three-target product architecture, so the same imaging behavior carries from in-app camera to secure lock-screen capture and final export.

RAW-first modular core

RAW-capable lens discovery, Bayer capture, and non-destructive development preserve dynamic range while an inward dependency model keeps the imaging core stable as features expand.

Unified destination routing

Creators can choose deliberate Gallery curation or instant Photos delivery, including captures that begin from lock screen, without changing tools, presets, or processing behavior.

HDR/EDR-aware non-destructive processing

The imaging engine uses a shared plugin-style pipeline for preview, capture, editor rendering, gallery updates, and export. Every image follows the same deterministic stage order, while edits stay metadata-based and source RAW remains untouched.

Deterministic non-destructive stages

Linear, tonal, grading, and curve stages run in a fixed order for predictable behavior across live preview, Lab editing, and re-processing workflows.

HDR/SDR output strategy

EDR 10-bit, EDR Safe 10-bit, SDR 8-bit, and original RAW outputs provide quality, compatibility, and source-handoff options per destination.

How LUMA builds one 48MP photo from a RAW burst

In 48MP mode, LUMA captures a rapid same-exposure RAW burst, then fuses it into one high-resolution frame with an 8064-pixel long edge. Burst size scales with each iPhone's capture headroom, and flash is disabled to keep frame-to-frame consistency.

1. Burst capture

LUMA buffers the full RAW burst before processing, so fusion runs on a complete stack instead of partial frames.

2. Coarse-to-fine alignment

Each frame is aligned against a reference through a two-stage motion search, then refined to sub-pixel precision for cleaner edge placement.

3. Robust weighted merge

Frames with weak alignment confidence are dropped. Accepted frames are merged with adaptive weights that reduce moving outliers and suppress noise while preserving detail.

4. Final development

After fusion, LUMA applies its standard non-destructive look pipeline and prepares output for Gallery curation or direct Photos delivery.

  • Output target: 8064-pixel long edge (approximately 48MP)
  • Exposure strategy: same-bias RAW burst for consistency, not mixed-exposure HDR bracketing
  • Quality gating: alignment score, residual checks, and confidence thresholds before merge
  • Post-merge cleanup: adaptive denoise tuned by accepted frame count
  • Access model: Club unlock with safe fallback to high-quality 12MP capture

Production-ready workflow across app and extensions

Unified on-device runtime

Live preview, still capture, and preset-aware rendering run locally with no cloud dependency in the core imaging flow.

Camera access everywhere

LUMA is available from the app, Lock Screen, Control Center, and Action button paths, with pro controls mapped to each device's camera capabilities.

Shared preset intelligence

Preset recipes and camera context are shared across launch surfaces, so looks, compatibility behavior, and destination intent stay consistent.

Club-aware access control

Club membership unlocks expanded capture and preset paths, while non-Club states safely fall back to high-quality defaults without breaking workflow continuity.

  • Platform: iOS (iPhone), tuned for modern system camera entry points
  • Product surfaces: Main app, secure lock-screen capture extension, and control launcher extension
  • Architecture: 24 local Swift packages (feature + core) with inward dependency layering
  • Imaging model: One shared RAW-first pipeline for preview, capture, editing, and export
  • Camera stack: AVFoundation with lock-screen capture support
  • Rendering and processing: Core Image + Metal
  • Media paths: Curated Gallery workflow or direct Photos workflow
  • Scalable operations: RAW import, multi-image export, and batch RAW processing

LUMA is designed as one coherent system: modular architecture, one imaging pipeline, and secure multi-surface capture flowing into the same professional mobile workflow.