I'm pretty sure this isn't a Libraw problem, but a problem with my code! I use unpack() to extract the raw image data and then convert the data into my application's internal image format using code based on code I found in the Libraw/dcraw source code.
The problem is that my images are somewhat too red in colour as compared to how (e.g.) Windows "Icon view" or IrfanView display them.
Is there anyone who would be prepared to look at my code and work with me to help work out what I am doing wrong.
I just downloaded the latest Snapshot 202502 from Github.
Nice that it has added 14-bit support for Olympus cameras.
However, I'm seeing 2 big issues:
1. Extremely slow decoding speed. Decoding a compress Fuji RAF file is now at least 10x slower.
2. Yellow tint on every decoded image (RAF, CR3, ORF, etc.) Almost seems like WB and tint are off.
I have the following comprehension question to development of LibRaw:
Till yesterday I thought that it is a more or less a linear development
- so snapshot 2025-02 is fully based on version 0.21.3 and contains all features of 0.21.3
- same for version 0.21.4: based on snapshot 2025-02 and contains all features of it
I'm currently using LibRaw to extract a completely raw image without any interpolation or processing, similar to dcraw's document mode (dcraw -D -4). I'm attempting to achieve this by setting the following parameters:
In a world where smartphone camera matrices have too many megapixels, there is a need for an algorithm that best demosaic an oversampled image.
Remosaic + debayer interpolate - give a very synthetic and ugly result.
4 RGGB subpixels in 1 RGB pixel must be simple.
Add it support, please.
It appears LibRaw now supports the Canon R5 Mark II camera in beta. Any idea how long on average it takes to move to production? More importantly the amount of time before darktable incorporates this version?
I've been able to debayer RAW images and "import" them into OpenCV, but I was reading through the website and now I'm wondering if I'm using the best process. Currently, I do the following:
Recent comments