I’ve been building an open-source RAW photo editor called Pu-erh Lab as a personal project before graduating from college.
I started it just because I’m an amateur photographer and know something about programming, and I kept feeling that today’s photo workflows and video workflows are still oddly disconnected. Many parts of the imaging pipeline remain black boxes, and most of the choices available to users are constrained by whatever a given piece of software decides to expose. I wanted to explore whether a RAW editor could offer stronger performance, a more transparent pipeline, and more meaningful control over the image itself.
So far, I’ve focused on a few areas:
- a CUDA-based processing pipeline for responsive high-resolution preview
- a 32-bit float imaging pipeline with ACES 2.0/OpenDRT display rendering transform
- LUT support and non-destructive editing
- a Git-like edit history/branching model
- a custom asset-management backend using DuckDB
One result I’m reasonably happy with is that the editor can preview large RAW images in real time, and library browsing stays fairly memory-efficient even with hundreds of high-resolution files.
It’s still early, still rough in many places, and currently Windows-CUDA-only. I’m not trying to out-market Lightroom or replace existing open-source tools outright. This project started more as an attempt to test my software engineering skills and understanding of image processing.
yurunzi•2h ago
I’ve been building an open-source RAW photo editor called Pu-erh Lab as a personal project before graduating from college.
I started it just because I’m an amateur photographer and know something about programming, and I kept feeling that today’s photo workflows and video workflows are still oddly disconnected. Many parts of the imaging pipeline remain black boxes, and most of the choices available to users are constrained by whatever a given piece of software decides to expose. I wanted to explore whether a RAW editor could offer stronger performance, a more transparent pipeline, and more meaningful control over the image itself.
So far, I’ve focused on a few areas:
- a CUDA-based processing pipeline for responsive high-resolution preview - a 32-bit float imaging pipeline with ACES 2.0/OpenDRT display rendering transform - LUT support and non-destructive editing - a Git-like edit history/branching model - a custom asset-management backend using DuckDB
One result I’m reasonably happy with is that the editor can preview large RAW images in real time, and library browsing stays fairly memory-efficient even with hundreds of high-resolution files.
It’s still early, still rough in many places, and currently Windows-CUDA-only. I’m not trying to out-market Lightroom or replace existing open-source tools outright. This project started more as an attempt to test my software engineering skills and understanding of image processing.