Software

ƒlatcam

FLATCAM-2026-11 · In Testing
ƒlatcam, Software

Overview

Phone cameras bake heavy processing into every photo. ƒlatcam goes the other way: it shoots flat, with natural colour, gentle contrast and room in the shadows and highlights, then applies a recipe, a film-style colour look rendered live in the viewfinder and either baked into the photo or shown only while shooting. Recipes come built in, import from .cube LUT files, or are made from any photo you like: the app reads that photo's tone, its colour in the shadows, midtones and highlights, and its saturation, and builds a look from it. Manual control runs through ruler dials that coast to a stop with a haptic click on every step, and the mode (auto, shutter priority, ISO priority or manual) follows whichever dials you turn. On Android it is built on Camera2 with an OpenGL ES 3 rendering pipeline; on iPhone on AVFoundation, Core Image and Metal. Night mode stacks several frames for cleaner low-light shots, and video records with the look baked in or as LOG.

Stack

Kotlin + Camera2 (Android)OpenGL ES 3 LUT rendererSwift, AVFoundation, Core Image + Metal (iOS)SwiftUI.cube 3D LUT parserMulti-frame night stackingadb device test harness (Python)

Features

Key Technical Challenge

Rendering a 3D LUT look live on the camera preview and applying exactly the same look to full-resolution photos and video, while keeping the viewfinder at the right aspect ratio through every fold state of a Galaxy Z Fold (cover screen, inner screen, tabletop, tent) without restarting the camera.

Lessons Learned

A device test harness pays for itself: scripted checks over wireless adb (cold launch, warm resume and every fold transition, with pixel statistics on screenshots and a crash-buffer check) caught layout and orientation bugs that would never show up on an emulator. Keeping the app offline by design, with no internet permission at all, made the privacy policy one sentence long.

Links

← Back to portfolio