Maybe I am just not understanding something right, so ping @heckflosse or @jdc.
If I look at the overexposed area in the sky in the provided raw image of the boat, and set demosaicing to ‘None’, I readout R = 16383, G1 = 15340 (or slightly different), G2 = 15351 (or slightly different), B = 16383. I now set these levels as white levels in camconst.json. If I then reload the image and set demosaicing to ‘Mono’ I get consistent values of RGB = 1. I interpret this as a confirmation that the white levels are set properly. But when I then switch to any other demosaicing method, the readout shows that the G channel is less than 100%.
I cannot understand why, so I hope someone can enlighten me 
Hi @Thanatomanic thanks for your time…actually I Am busy but I will make new shot soon… Many thanks for now
Hi @Thanatomanic,
I made the shots …I used gofile because filebin got network error 
I hope it’s the same…
Ask me whatever you need 
If I’m correct… Do I have to make a camconst.json file where I have options file ??
Would it be as
/home/gabriele/.config/RawTherapee5-dev ??
I don’t know if it’s of value but if I disable Clip out-of-gamut colors and I decrease the Exposure the magenta is more evident 
1 Like
To update your own camconst.json, just edit the file in the installation folder of RawTherapee. Look for the D500 and just add the D300 entry above it. I am not sure if it would work if you create your own file… Never tried
The updated version of the camconst.json entry looks like this:
{ // Quality B, samples provided by dasafraga (pixls.us)
"make_model": "Nikon D300",
"dcraw_matrix": [ 9000, -1966, -711, -7030, 14976, 2185, -2354, 2959, 7990 ], // DNG_v13 (Neutral v4)
//"dcraw_matrix": [ 9030, -1992, -716, -8465, 16302, 2256, -2689, 3217, 8068 ], // DNG_v13 (Neutral)
"ranges": {
"white": [
{ "iso": [ 200, 250, 320, 400, 500, 640, 800, 1000, 1250, 1600 ], "levels": [ 16383, 15292, 16383, 15317 ] }, // R, G1, B, G2 - R G typically saturate, G1 G2 differ slightly
{ "iso": [ 2000, 2500, 3200 ], "levels": [ 16383, 16323, 16383, 16347 ] }
],
"white_max": 16383,
"aperture_scaling": [ // There seems to be no aperture scaling at ISO2000+, but we cannot deal with that
{ "aperture": 2.8, "scale_factor": 1.0246 },
{ "aperture": 3.2, "scale_factor": 1.0214 },
{ "aperture": 3.5, "scale_factor": 1.0140 },
{ "aperture": 4.0, "scale_factor": 1.0106 },
{ "aperture": 4.5, "scale_factor": 1.0076 },
{ "aperture": 5.0, "scale_factor": 1.0024 }
]
}
},
I’m sorry to say that this does not solve the purple issue, and I am really out of my depth to answer the question “why?”. I expect I am missing some essential information on how these entries actually influence the processing, but currently I have no desire to dive into the code. I hope somebody else can take a look. Maybe it would be good to report this as an issue on the GitHub tracker…
To be continued, I hope.
I wonder why some camconst.json entries have only one white level, while the comment mentions different ones, for example
{ "iso": [ 100, 125, 160, 200, 250, 320, 400, 500 ], "levels": 15520 }, // typical G1,G2 15520-15800 R,B 16383
When I use the lowest WL from your entry for all channels, the purple issue goes away.
1 Like
Hi,
I created a camconst.json file next my options file into .config
“camera_constants”: [
{ // Quality B, samples provided by dasafraga (pixls.us)
“make_model”: “Nikon D300”,
“dcraw_matrix”: [ 9000, -1966, -711, -7030, 14976, 2185, -2354, 2959, 7990 ], // DNG_v13 (Neutral v4)
//“dcraw_matrix”: [ 9030, -1992, -716, -8465, 16302, 2256, -2689, 3217, 8068 ], // DNG_v13 (Neutral)
“ranges”: {
“white”: [
{ “iso”: [ 200, 250, 320, 400, 500, 640, 800, 1000, 1250, 1600 ], “levels”: [ 16383, 15292, 16383, 15317 ] }, // R, G1, B, G2 - R G typically saturate, G1 G2 differ slightly
{ “iso”: [ 2000, 2500, 3200 ], “levels”: [ 16383, 16323, 16383, 16347 ] }
],
“white_max”: 16383,
“aperture_scaling”: [ // There seems to be no aperture scaling at ISO2000+, but we cannot deal with that
{ “aperture”: 2.8, “scale_factor”: 1.0246 },
{ “aperture”: 3.2, “scale_factor”: 1.0214 },
{ “aperture”: 3.5, “scale_factor”: 1.0140 },
{ “aperture”: 4.0, “scale_factor”: 1.0106 },
{ “aperture”: 4.5, “scale_factor”: 1.0076 },
{ “aperture”: 5.0, “scale_factor”: 1.0024 }
]
}
}
]}
and it works 
Hi Ingo,
as you stated a single level of 15292
{“camera_constants”: [
{ // Quality B, samples provided by dasafraga (pixls.us), intermediate aperture scaling not available
“make_model”: “Nikon D300”,
“dcraw_matrix”: [ 9000, -1966, -711, -7030, 14976, 2185, -2354, 2959, 7990 ], // DNG_v13 (Neutral v4)
//“dcraw_matrix”: [ 9030, -1992, -716, -8465, 16302, 2256, -2689, 3217, 8068 ], // DNG_v13 (Neutral)
“ranges”: {
“white”: [
{ “iso”: [ 200, 250, 320, 400, 500, 640, 800, 1000, 1250, 1600 ], “levels”: 15292 }, // R, G1, B, G2 - R G typically saturate, G1 G2 differ slightly
{ “iso”: [ 2000, 2500, 3200 ], “levels”: [ 16383, 16323, 16383, 16347 ] }
],
“aperture_scaling”: [
{ “aperture”: 2.8, “scale_factor”: 1.025 },
{ “aperture”: 4.0, “scale_factor”: 1.011 }
]
}
}
]}
removes the pink/magenta color but when I compress highlights until to have no clipping and apply HR as Color propagation which make more clipping and I apply more highlights compression I have the artefacts.
Neutral with clipping
Highlights compression until no more clipping
Reconstruction with color propagation which gives clipping
More compression to delete clipping, but artefacts
Perhaps I want more that RT can do ??
@dafrasaga The color propagation method has its limits, and I think you found them. There is very little color information to work with in nearly-overexposed areas, so it is also very difficult (if not impossible) to make an educated guess to the reconstruction of these colors.
1 Like