It is my opinion that for any serious analysis one needs to use black frames and disable subtraction of the recommended black level. There is also a factor of any hard-wired non-uniformity control, including vignetting (tests are better done with the lenses the camera doesn't know and can't read the aperture setting)

it is not a given that black level is uniform across the image serious cameras include 'black level flat field' data (usually it needs interpolation as it is sparse), or separate black levels for top and bottom of the frame (in assumption that a gradient ca get one close enough) - Sony are not one of those cameramakers to include such data, or we haven't found it yet. the recommended level is not always maintained in practice the thing that is has an unfortunate naming of PostAEGain - it is the amount of digital scaling applied to raw data to calibrate the sensor response In reality there are some other things skew the distribution: The way the black level is formed may involve clipping of some data, and it is quite often that the black level includes an artificial pedestal. If RawDigger / LibRaw is set to auto black level, that is what is subtracted. That is the recommended black level for consumer-oriented converters to hold to. Tag 0x7310 in makernotes reflects the programmed black level. Also, I just checked and Libraw's 4channels.exe uses subtraction by default, in case others use that for their raw file splitting like I do - you can turn off the subtraction via -B. Iliah gave me some additional details behind that, which I'll post below. I kind of expected some measure of imbalance since it's chopping off half the data but I didn't realize it be to this extent. Profile Edition allows applying white balance, normalization, equalization (Flat Field), and filtering of outlying values before data export.I manually calculated the subtracted the biased data and arrived at the same result as Libraw, so operating as expected. Profile Edition adds selection grids for faster processing of step wedges and color targets, as well as for calculating non-uniformity maps. Profile Edition is for those who need to create sensor calibration data, linearization data, or device data for color profiling. Research Edition adds TIFF export, multiple selected areas via samples, tables of sampled data statistics and histograms over multiple samples, exporting sampled data as CSV and CGATS. It is suitable for a wide range of tasks, starting from personal use and preparing reviews and all the way to programming RAW data processing.

Research Edition is meant for camera/sensor evaluation and statistical analysis of raw data. RawDigger is available in several editions:Įxposure Edition is for everyday use and is intended to help those who are serious about extracting the maximum quality from the camera to get precise exposures.Įxposure Edition displays RAW data in all modes (RGB, RAW Composite, and RAW per channel), indicates over- and underexposed areas, displays RAW histogram, image statistics, as well as statistics and histogram for selected area. It helps prepare data for calculations of color transforms and camera profiles. It helps to determine how exposure meters are calibrated, what is the headroom in highlights, which camera setting affect RAW data and how, how a raw converter interprets RAW data and which 'hidden' compensation it applies, what tone curve is used for in-camera JPEGs and in a converter. RawDigger can be used to solve a multitude of tasks, including the diagnosis of different problems, such as those with studio lights, lenses, flash, shutter, and aperture accuracy and repeatability, camera drifts, exposure, noise. The primary goals of RawDigger are to help you get better shots and increase the number of 'keepers' through better knowledge of the 'digital film' you are using, that is, RAW data. In essence, RawDigger is a microscope of sorts that lets you drill down into the RAW by means of visualizations, histograms, and statistics over selections and samples.

Libraw rawdigger