OOC jpeg color in raw


#1

I have used rawtherapee since its inception until 2 years ago when I switched to darktable. But, I still keep track of its new features in the new releases. The recent release 5.3 has this feature added
"New histogram matching tool, to have RawTherapee automatically adjust the image for you to match the out-of-camera look "
So, I stalled it, and amazingly, the color match of raw to ooc jpeg is really close. But overall, I find darktable has better and easier features that fit my need.
Question for developers, any chance that we can have this ooc jpeg color automatically in raw without having to make a whole lot of fine tuning?


(Alan Gibson) #2

Isn’t that what darktable-chart does? See https://www.darktable.org/usermanual/en/darktable_chart.html


(Ingo Weyrich) #3

No, RT can do the histogram matching for each image automatically now, at the cost of some processing time of course. That feature builds a luminance tonecurve to match the contrast (iirc not the colours) of the jpeg which is embedded in the raw file.


(Andreas Schneider) #4

Yes, I agree but creating camera profiles is the more accurate way to go.


(Alberto) #5

hi,

histogram matching is not meant to be a substitute for camera profiles – the two are quite complementary in fact.


#6

I thought darktable already has camera profile.


(Glenn Butcher) #7

Yes, ‘camera profile’ is a somewhat overloaded term. Let’s add some adjectives:

‘Calibrated camera profile’ contains the color information needed to describe the particular camera’s color gamut. You shoot an image of a color target and make the profile, either ICC or DNG, and give it to the raw processor to use as the basis for all further work with the image.

‘Camera look profiles’ (okay, I made up a term, hope it doesn’t generate the trouble ‘unbounded’ did in another thread) are profiles that make images from a particular camera take on a ‘look’, such as ‘Vivid’, etc. They may also incorporate the calibrated information for the particular camera, so the image transform can be done in one operation. Cameras themselves usually apply such profiles in producing their OOC JPEGs.

‘histogram matching’ is IMHO a really neat way for a raw processor to mine the OOC JPEG that’s embedded in the raw file for the things the ‘camera look profile’ did to make the OOC JPEG, and apply them to the raw image. So now, from the raw file you can start with the image looking like the JPEG, but with 16-bit or floating point precision, so you can do more work on the image without collecting noticeable ill effects such as posterization. So, with histogram matching, the ‘camera look profile’ can provide the from the JPEG the basis for raw editing that formerly required ferreting out the look profiles from other software.

I hope this helps a bit; I’ve stumbled on such overloaded terminology while re-learning photography, and sometimes the nature of forum posting doesn’t help to make clear…