Attempting to learn how to use the RAW Black Points tool. Fuji X-T2 & other X-trans imbed a black level in EXIF like many cams. For the X-T2, the level varies in the range of 1024. This appears to result in black clipping - perhaps as a noise reduction strategy. But for astro image pre-processing, it drops out faint shadow detail.
I’d like to lower the black point used during RAW conversion so blacks are not clipped. First question: is this what happens if I used negative slider values in the RAW Black Points tool? Second question: Are the slider units arbitrary or related to pixel values? It would be useful if the slider units corresponded to EXIF black level values.
Interesting… I just procured a Nikon Z6, which has a non-zero black point in the metadata. When I open the raw image with no processing, the histogram doesn’t start until that black point. I have to subtract it from the image to get blacks down to zero…
I doubt you’ll find a camera w/ a zero black point. Pretty sure all digital RAW has an offset so the lowest analog pixel values will not suffer from quantization errors when ADC is applied. Moving the black level up is done to remove the off-set in the digital levels.
I have a collection of images from many Canon, Sony, Fuji, Oly, Panasonic cameras & none have a zero black level in EXIF. I have no Nikon images.
Idealy, the black level is close to the off-set value so blacks are not clipped during RAW conversion. That’s the problem in the XT2.
My D7000 has a metadata-reported blackpoint of zero…
Now, to be fair, for the Z6 I haven’t inspected the data mins to see if they line up with the metadata number. In my software, the up-coming 0.9 will have a subtract tool with three options: 1) user-supplied number, 2) a “flat” file, and 3) the metadata number. I may add a fourth option: 4) min data value…
Interesting. Some Nikons have been critiqued in the astro community for clipping blacks. Maybe Nikon is employing the same strategy as Fuji, setting a black level above the noise floor to reduce noise but also clipping blacks.
What software do you use to inspect data minimums?
And FWIW, the Fuji X-trans cams use a variable black level depending on some unknown characteristic of an image. CORRECTION: the XT2, XPro2 & XT3 use variable BL. The XT10 does not nor doe the bayer XT100.
One of the options is to open the file without any processing, e.g., dcraw -D. The histogram is then of the raw image data as read from the file, discounting the conversion from unsigned short to float…
I’m on a Mac & not command line literate so have not jumped into DCraw.
RT allows viewing images w/o any processing or demosiacing but I’m not sure where the histogram comes from in that case & if its w/o black level adjustment. Thats an important question for RT support folks.
Maybe I’m just doing something stupid but I got slightly confused over the result.
I opened a dark image with some very bright areas ( cars in the night, bright head lights) and moved the black point far left and then added some exposure compensation to brighten the image. And all of a sudden the bright burnt out areas turned black. Almost as if clipped highlights warning was enabled (which it wasn’t).
Is that supposed to happen?
To test it: Neutral profile, black point to max negative, raise exposure compensation. And bright areas turns black. RT 5.6-947-something.