color.maximum and camera white level

Is the value in color.maximum the camera white level ("full well" or saturation value)? If so I am a bit confused as the data I'm testing with appears to contain pixel values that are greater than color.maximum (after subtraction of dark level).

To be more specific I'm using a CR2 file from a Canon EOS 60Da where color.maximum is showing up as 12279 but the largest pixel value in raw_image (ignoring the frame) is 13259 (after subtraction of dark value of 2047).

If color.maximum isn't the white level, is the white level for the camera available in libraw?

Why am I asking? Because when I apply my own linear image stretch to the raw data in raw_image, I wish to scale it so that a white level pixel comes out as USHORT_MAX so that files with difference exposures are represented correctly.



For Canon cameras, imgdata

For Canon cameras, imgdata.color.maximum is set by metadata provided by vendor in CR2 file.

Not doing so results in 'pink highlights' problem.

-- Alex Tutubalin

It's great that you get it

It's great that you get it from the CR2 file, but you didn't answer my question - does it represent the "white level" or not? And as an aside is it normal to see pixel values > that level?

David Partridge

AFAIK, this is vendor

AFAIK, this is vendor-specified value for white point.

-- Alex Tutubalin