Extracting Bayer array from DNG file results in black image

CFA
1

I’m trying to get the Color Filter Array (CFA) from a DNG file using MATLAB. But I’m running into a problem. When I use the code to read the CFA, I just get a matrix full of zeros. This shows up as a completely black image when I plot it.

Here’s the code I’m using:

imageInfo = getImageDetails('myPhoto.dng');
tiffObj = openTiffFile('myPhoto.dng', 'read');
subDirOffsets = tiffObj.fetchTag('SubIFD');
tiffObj.moveToSubDir(subDirOffsets(1));
rawCFA = double(tiffObj.readData());

I expected to see the Bayer array, but instead, I’m getting this all-black image. Does anyone know what might be going wrong here? How can I correctly extract the Bayer pattern from my DNG file?

I’d really appreciate any help or suggestions. Thanks!

2

I’ve encountered a similar issue when working with DNG files. One potential solution is to check if the raw data is stored in a separate IFD (Image File Directory) within the DNG. Some cameras store the raw data in a different location than the main image.

Try modifying your code to iterate through all SubIFDs:

for i = 1:length(subDirOffsets)
    tiffObj.setDirectory(subDirOffsets(i));
    rawCFA = double(tiffObj.read());
    if any(rawCFA(:))
        break;
    end
end

This should help locate the correct IFD containing the raw data. Additionally, ensure you’re applying any necessary bit shifts or white balance coefficients specified in the DNG metadata. These steps are crucial for properly interpreting the raw sensor data.

If you’re still having trouble, consider using a dedicated DNG parsing library like LibRaw, which can handle various DNG formats and quirks more robustly.

3

hey ava_books, have u tried checkin the compression settings? some dngs use lossless compression which can mess with direct readin. also, make sure ur bit depth is set right - might need to scale up the values to see em properly. if all else fails, maybe try a different dng reader library? sometimes they handle weird formats better. good luck!

4

Hey there, Ava_Books! :wave:

I’m really intrigued by your question about extracting the Bayer array from a DNG file. It’s such a cool topic! :camera_flash:

Have you considered that the DNG file might be compressed? Sometimes, compressed DNG files can throw a wrench in the works when trying to read the raw data directly. Maybe we could try decompressing it first?

Also, I’m curious about the bit depth of your DNG. Could it be that the values are there, but they’re just really small compared to the full range? Maybe we need to scale the values up to see them properly?

Oh, and here’s a thought - what if we try reading the metadata first to check the CFA pattern? That might give us a clue if something’s off with the raw data reading.

What do you think about these ideas? Have you tried any of them already? I’d love to hear more about your experiments and what you’ve discovered so far!